Economic analysis, moral philosophy, and public policy [Third reprinted.] 9781107158313, 1107158311, 9781316610886, 1316610888

Citation preview

ECONOMIC ANALYSIS, MORAL PHILOSOPHY, AND PUBLIC POLICY

Third Edition This book shows through argument and numerous policy-related examples how understanding moral philosophy can improve economic analysis, how moral philosophy can benefit from economists’ analytical tools, and how economic analysis and moral philosophy together can inform public policy. Part I explores the idea of rationality and its connections to ethics, arguing that when they defend their formal model of rationality, most economists implicitly espouse contestable moral principles. Part II addresses the nature and measurement of welfare, utilitarianism, and cost–benefit analysis. Part III discusses freedom, rights, equality, and justice – moral notions that are relevant to evaluating policies, but that have played little if any role in conventional welfare economics. Part IV explores work in social choice theory and game theory that is relevant to moral decision making. Each chapter includes recommended reading and discussion questions. Daniel M.  Hausman is the Herbert A.  Simon and Hilldale Professor of Philosophy at the University of Wisconsin-Madison. A founding editor of the journal, Economics and Philosophy (with Michael McPherson), his research has centered on epistemological, metaphysical, and ethical issues at the boundaries between economics and philosophy. He is the author of Capital, Profits, and Prices (1981), The Inexact and Separate Science of Economics (1992), Causal Asymmetries (1998), Preference, Value, Choice, and Welfare (2012), and Valuing Health: Well-Being, Freedom, and Suffering (2015). Michael S.  McPherson is President of the Spencer Foundation and Past President of Macalester College. He cofounded the journal Economics and Philosophy with Daniel Hausman and has worked on problems on the borders of economics and philosophy. He is coauthor of six books on higher education policy and economics, including Lesson Plan:  An Agenda for Change in American Higher Education (2016), Crossing the Finish Line: Completing College in America’s Public Universities (2009), and The Student Aid Game:  Meeting Need and Rewarding Talent in American Higher Education (1998). Debra Satz is the Marta Sutton Weeks Professor of Philosophy and Ethics in Society at Stanford University, where she is also the Senior Associate Dean for Humanities and Arts. Her research interests include the moral limits of the market, the nature of equality, and the public/private boundary. She is the author of Why Some Things Should Not Be for Sale: The Moral Limits of Markets (2010) and the coeditor of Toward a Humanist Justice: The Political Philosophy of Susan Moller Okin (2009) and Occupy the Future (2013) and is the author of numerous articles.

Economic Analysis, Moral Philosophy, and Public Policy Third Edition

DANIEL HAUSMAN University of Wisconsin–Madison

MICHAEL McPHERSON Spencer Foundation, Chicago

DEBRA SATZ Stanford University

One Liberty Plaza, New York NY, 10006 USA Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781316610886 © Daniel Hausman, Michael McPherson, and Debra Satz 2017 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2017 Printed in the United States of America by Sheridan Books A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Hausman, Daniel M., author. | McPherson, Michael S., author. | Satz, Debra, author. Title: Economic analysis, moral philosophy, and public policy / Daniel Hausman, University of Wisconsin–Madison, Michael McPherson, Spencer Foundation, Chicago, Debra Satz, Stanford University. Description: Third edition. | New York, NY: Cambridge University Press, 2017. | Includes bibliographical references and index. Identifiers: LCCN 2016033203| ISBN 9781107158313 (hard back) | ISBN 9781316610886 (paper back) Subjects: LCSH: Economics – Moral and ethical aspects. | Ethics. | Political planning. Classification: LCC HB72.H357 2017 | DDC 174/.4–dc23 LC record available at https://lccn.loc.gov/2016033203 ISBN 978-1-107-15831-3 Hardback ISBN 978-1-316-61088-6 Paperback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party Internet Web sites referred to in this publication and does not guarantee that any content on such Web sites is, or will remain, accurate or appropriate.

To the three major influences on our thinking about economics and ethics Kenneth Arrow John Rawls Amartya Sen

Contents

List of Figures

page xiii

List of Tables

xiv

Preface to the Third Edition

xv

Preface to the Second Edition

xvii

Acknowledgments

xix

1

2

Ethics and Economics?

1

1.1 Economics and Morality in Contemporary Controversies 1.2 What Are Moral Questions and How Can They Be Answered? 1.3 How Is Moral Philosophy Relevant to Economics? 1.4 Organization of the Book Suggestions for Further Reading Questions for Study and Discussion

2 8 12 14 16 16

Ethics in Welfare Economics

19

2.1 2.2 2.3 2.4 2.5

19 20 23 24

A Shocking Memorandum Eight Distinctive Features of Welfare Economics The Economic Benefits of Exporting Pollution to LDCs The Argument of the World Bank Memorandum Should the World Bank Encourage Migration of Dirty Industries to LDCs? 2.6 School Vouchers 2.7 Conclusions Suggestions for Further Reading Questions for Study and Discussion

vii

27 31 36 36 37

Contents

viii 3

Ethics in Positive Economics: Two Examples

39

3.1 Is Unemployment Involuntary? 3.2 Overlapping Generations 3.3 Conclusions Suggestions for Further Reading Questions for Study and Discussion

40 48 51 51 51

PART I 

4

5

6

RATIONALITY, MORALITY, AND MARKETS

53

Rationality and Utility Theory

55

4.1 Certainty and Ordinal Utility Theory 4.2 Expected Utility Theory 4.3 Questions about Utility Theory Suggestions for Further Reading Questions for Study and Discussion

56 60 64 68 69

Rationality and Morality in Positive Economics

70

5.1 5.2 5.3 5.4 5.5

Rationality and Positive Economics Self-​Interest, Rationality, and Morality The Moral Danger of Ignoring Morality The Influence of Moral Norms on Economic Behavior How Do Norms Motivate Human Actions and What Sustains Them? 5.6 Conclusions: Rationality, Morality, and Positive Economics Suggestions for Further Reading Questions for Study and Discussion

70 74 78 81

The Ethical Limits to Markets

92

6.1 6.2. 6.3 6.4 6.5

93 94 95 98

Are There Any Limits to the Reach of the Market? Market Virtues Market Limits Morals and Markets Policy Implications: How Should Policy Respond to Morally Problematic Markets? 6.6 Market Alternatives 6.7 Conclusions Suggestions for Further Reading Questions for Study and Discussion

84 88 89 90

102 103 104 104 104

Contents PART II 

7

WELFARE AND CONSEQUENCES

107

Utilitarianism, Consequentialism, and Justice

109

7.1 Clarifying Utilitarianism 7.2 Interpersonal Comparisons of Well-​Being 7.3 Justifying Utilitarianism 7.4 Contemporary Consequentialism 7.5 Is Utilitarianism Plausible? 7.6 Consequentialism and Deontology 7.7 Conclusion: Should Economists Embrace Utilitarianism? Suggestions for Further Reading Questions for Study and Discussion

110 114 116 117 119 121 123 123 124

8 Welfare

9

ix

126

8.1 Theories of Well-​Being 8.2 Welfare in Economics 8.3 Against the Constitutive View: Well-​Being Is Not Preference Satisfaction 8.4 Conflicting Preferences and Well-​Being 8.5 Assessing Preferences 8.6 The New Hedonist Welfare Economics 8.7 Other Theories of Well-​Being 8.8 Conclusions Suggestions for Further Reading Questions for Study and Discussion

127 127

Welfare Economics

146

9.1 Preference Satisfaction, Pareto Efficiency, and Competitive Equilibrium 9.2 Pareto Efficiency, Pareto Improvements, and Benevolence 9.3 How Welfare Economics Narrows Normative Questions 9.4 Cost–​Benefit Analysis 9.5 Objections to Cost–​Benefit Analysis 9.6 Cost–​Benefit Analysis and Social Welfare Functions 9.7 Putting Cost–​Benefit Analysis to Use 9.8 Welfare Economics Suggestions for Further Reading Questions for Study and Discussion

147 150 154 158 161 165 167 169 170 171

130 134 135 138 141 143 143 144

Contents

x PART III 

LIBERTY, RIGHTS, EQUALITY, AND JUSTICE

10 Liberty, Rights, and Libertarianism

11

173 175

10.1 Freedom 10.2 What Are Rights? 10.3 The Importance of Rights 10.4 The Justification of Rights and Freedoms 10.5 Weighing Rights, Liberties, and Welfare 10.6 Libertarianism 10.7 Libertarian Paternalism Suggestions for Further Reading Questions for Study and Discussion

176 180 181 183 184 184 189 190 191

Equality and Egalitarianism

194

11.1 What Is Wrong with Contemporary Income and Wealth Inequalities? Consequences and Causes 11.2 What Is Wrong with Inequality Itself? “Basic” Egalitarianism and Prioritarianism 11.3 Egalitarian Objectives 11.4 The “Currency” of Egalitarian and Prioritarian Justice 11.5 Relational Equality and Equality of Moral Status 11.6 The Measurement and Importance of Inequality Suggestions for Further Reading Questions for Study and Discussion 12 Justice and Contractualism 12.1 The Social Contract Idea 12.2 Justice as Reciprocity: Rawls’s Theory of Justice 12.3 How Can Rawls’s Principles Guide Policy? 12.4 Other Contract Theories 12.5 Challenges to Contractualism 12.6 Conclusion: Social Contract Reasoning and Economics Suggestions for Further Reading Questions for Study and Discussion PART IV 

MORAL MATHEMATICS

13 Social Choice Theory 13.1 The Social Welfare Function and Arrow’s Theorem 13.2 The Interpretation of Arrow’s Theorem

197 201 206 208 215 217 220 221 223 224 226 233 236 238 240 241 242 245 247 247 251

Contents 13.3 Social Choice Theory and Moral Philosophy 13.4 Social Welfare Functions with Interpersonal Comparisons 13.5 Rights, Freedom, and Fairness in Social Choice Theory 13.6 Other Developments of Social Choice Theory 13.7 Conclusions Suggestions for Further Reading Questions for Study and Discussion 14 Game Theory 14.1 What Is a Game? 14.2 Moral Philosophy and Some Simple Games 14.3 Cooperation and Justice 14.4 Paradoxes and Difficulties 14.5 Bargaining Theory and the Social Contract Suggestions for Further Reading Questions for Study and Discussion PART V 

CONCLUSIONS

15 Putting Economics and Ethics to Work 15.1 Involuntary Unemployment and Moral Baselines 15.2 Overlapping Generations 15.3 Do Pollution Transfers and Vouchers Promote Welfare? 15.4 The Bearing of Capabilities and Primary Goods on Pollution Trades and Voucher Proposals 15.5 A Utilitarian Perspective on Pollution Transfers and Educational Vouchers 15.6 Other Ways of Evaluating Vouchers and Pollution Transfers 15.7 Conclusions Questions for Study and Discussion 16 Economics and Ethics, Hand in Hand 16.1 Health Care 16.2 Uber and the “Gig” Economy 16.3 Environmental Protection and Global Warming 16.4 Conclusions Suggestions for Further Reading Questions for Study and Discussion

xi 254 256 258 263 265 266 267 269 269 274 278 280 286 289 290 293 295 296 299 302 305 307 309 312 312 314 316 320 327 334 335 336

Contents

xii

APPENDIX: HOW COULD ETHICS MATTER TO ECONOMICS?

A.1 Objection 1: Economists as Engineers A.2 Objection 2: Positive Economics Is Value Free: The Standard View A.3 Objection 3: Normative Questions Are Not Subject to Rational Evaluation A.4 Answering the Standard View: How Knowing Ethics Contributes to Positive Economics A.5 Conclusions: Ethics into Economics Questions for Study and Discussion

337 338 340 342 343 349 350

Glossary

353

References

361

Index

385

Figures

  3.1   3.2   9.1   9.2 11.1 14.1 14.2 14.3a 14.3b 14.4 14.5 14.6 14.7 14.8 14.9 14.10 14.11

Supply and demand in the labor market Sticky wages Cash versus in-​kind benefits Interdependence of equity and efficiency The Gini coefficient An extensive form game Simultaneous play A normal form game with sequential play A normal form game with simultaneous play The prisoners’ dilemma A pure coordination game Battle of the sexes An assurance game A game of chicken Divide the good Subgame perfect equilibrium A centipede game

xiii

page 41 43 155 162 218 270 272 273 273 275 277 277 277 278 281 283 291

Tables

  1.1   2.1   4.1   4.2   4.3   9.1 10.1 13.1 13.2 13.3 13.4 13.5 A.1

Exaggerated contrasts between facts and values The moral framework of normative economics The unusual Asian disease The unusual Asian disease (again) Allais’s problem Potential Pareto improvements Concepts of freedom Intransitivity of majority voting Interpretations of social choice theories Fair expectations vs. fair outcomes Lewd’s and Prude’s preferences The discursive dilemma Donating vs. selling blood

xiv

page 13 26 60 60 65 159 180 250 252 259 260 263 347

Preface to the Third Edition

The ideas of economists and political philosophers, both when they are right and when they are wrong are more powerful than is commonly understood. Indeed, the world is ruled by little else. Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually slaves of some defunct economist. John Maynard Keynes

In undertaking a third edition of this text, the authors of the first and second editions, Dan Hausman and Michael McPherson, set themselves a series of tasks:  to update the discussion, especially in the context of the financial crisis of 2007–​2009; to address issues that have become prominent in the decade since the previous edition such as growing social and economic inequality; to fill in some of the gaps in the range of issues covered previously; and to take cognizance of developments within ethics and relevant areas of economics. This is a tall order, for the relevant literature is immense and the conceptual challenges in selecting materials to discuss and synthesizing their insights are daunting. To help with this task, Hausman and McPherson asked Debra Satz to join them as a third author, and this third edition is the result of our three-​way collaboration. It has been extremely rewarding for us three, and we have learned a great deal from one another in writing together. We hope that our efforts will be rewarding to our readers, too. Although some sections of the second edition have made their way into this third edition with few changes, the book has been extensively rewritten with a great deal of new material. One unfortunate consequence is that it has grown longer, and bigger books are often more expensive, more intimidating, and less likely to find readers. But we believe that the new material is as important and timely as it is interesting (to us, at least). We hope that it will stimulate more economists and philosophers to learn from each other. xv

xvi

Preface to the Third Edition

We also hope that it will intrigue and delight our readers and more than compensate them for the greater investment of time and money that this edition demands. Among the most obvious differences between this edition and the second edition are the addition of questions for study and discussion at the end of each chapter and a completely new chapter on the limits of markets. Since the first edition of this book was published in 1996, much has changed. There was an air of naïve optimism in many quarters in the late 1990s: the Soviet Union had dissolved; many saw democracy as inevitably spreading everywhere; and it appeared that economists had largely tamed the business cycle. There was no shortage of difficult policy questions to which both economics and ethics had a great deal to contribute, but the task was more often to build upon the accomplishments of the past than to stave off disaster. Today, the situation is both better and worse. By many measures, the world has been making progress:  life expectancies are on the rise, and the United Nations estimates that the number of people living in extreme poverty around the world has fallen by 50  percent, from 1990 when 1.9 billion people lived on less than $1.25 a day to 836 million in 2015. At the same time, inequality is growing within both poor and rich countries; the risks to the financial system became increasingly evident in 2008; democracy remains only an aspiration to millions around the globe; international security is threatened; and the problems of climate change grow worse every year. The problems we face are complex. This book has no easy answers to the panoply of contemporary problems. But it continues to express our faith in the efficacy of rigorous and open-​minded thinking that draws upon both the tools and models of economists and the conceptual distinctions and theories of moral philosophers. Working at the boundaries between economics and philosophy, policy makers can, we hope, make progress on the problems faced by individual nations and the world as a whole. By making ethics accessible to economists and making economic tools and methods accessible to moral philosophers, we hope to encourage their collaboration on important social problems.

Preface to the Second Edition

This book is a heavily revised and retitled version of Economic Analysis and Moral Philosophy. We added “Public Policy” to the title to emphasize the relevance of this analysis to policy questions. The book is a descendant of a survey essay, “Taking Ethics Seriously: Economics and Contemporary Moral Philosophy,” which we published in the July 1993 issue of the Journal of Economic Literature. Though now dated, that survey essay may still be of use to readers for its extensive references to relevant literature. We would like to thank John Roemer for commissioning that essay and for the detailed criticisms he offered of several drafts. Others who were of tremendous help with the first edition were Richard Arneson, Henry Bruton, Nancy Cartwright, Marc Fleurbaey, John Kautsky, Eric Kramer, Philippe Mongin, Amartya Sen, Julius Sensat, Max Steuer, Hamish Stewart, Alain Trannoy, Gordon Winston, students at Williams College and the London School of Economics, and anonymous referees. Harry Brighouse, Henry Bruton, Lester Hunt, Andrew Levine, Patrick McCartan, Jonathan Riley, David Ruben, Larry Samuelson, and Daniel Wikler read drafts of chapters of the first edition and offered valuable assistance. The research and writing of the first edition were supported by a collaborative research grant from the National Endowment for the Humanities, and Hausman also gratefully acknowledges the support of a Vilas Associate Award from the University of Wisconsin–Madison. Since philosophical reflection on ethics continues apace, as does the development of economic concepts and tools that may be of use to moral philosophers, we thought that a new edition was called for. Although we have preserved the overall structure and many of the specific analyses, distinctions, and arguments of the first edition, we have brought the discussion up to date and added examples that we hope will further illuminate the issues we discuss. We aim to reach a large audience of those interested xvii

xviii

Preface to the Second Edition

in economics and policy analysis, and we have tried to avoid unnecessary jargon and complexities. In preparing this revised edition, we were aided by and would like to thank Elizabeth Anderson, Mavis Biss, Richard Bradley, Harry Brighouse, Michel De Vroey, Jeffrey Friedman, Francesco Guala, David Hausman, Joshua Hausman, Bernd Irlenbusch, William Jaeger, Philippe Mongin, Colin Patrick, David Schmitz, Russ Shafer-​Landau, William Thomson, Peter Vanderschraft, Joel Velasco, and David Zimmerman for detailed criticisms and suggestions for improvement. Marc Fleurbaey read the entire manuscript and offered extensive and insightful comments. Colin Patrick did a wonderful job preparing the Index, and Matt and Vickie Darnell did a superb job copyediting and typesetting this book. Over the years we have received a good deal of correspondence with suggestions for improvements, and we apologize to those whose help we may have forgotten to acknowledge.

Acknowledgments

In preparation for our work, we sought advice from many scholars, both economists and philosophers, and Cambridge University Press contacted a number of reviewers as well. Altogether, we received helpful comments from more than twenty people, whose recommendations have guided us in making significant changes to the manuscript. This mass of comments has brought its own challenges: it has led us to make far more extensive changes in the third edition than we had initially envisioned. Even allowing for the greater length of this edition, we could not follow all of the good advice we received, and we regret if we have disappointed some of our interlocutors through our failure to include their suggestions. In addition, when our critics did not agree, the rules of logic interceded, and we could not follow all of their recommendations. We are grateful to our copy-editor, Susan Thornton, and once again to Colin Patrick, who created the index and helped to correct the page proofs. In addition to seven anonymous reviewers who responded to Cambridge University Press, we would like to thank the following for their extremely generous comments. This edition would have more errors and less useful content but for the criticisms and suggestions of Matthew Adler, Elizabeth Anderson, Jonathan Anomaly, Geoffrey Brennan, Mikael Cozic, Marc Fleurbaey, Peter Hammond, Frances Kamm, Paul Kelleher, Julian Le Grand, Joshua Preiss, John Roemer, Frank Thompson, Philippe van Parijs, Alex Voorhoeve, and Jonathan Wolff. Special thanks are due to Marc Fleurbaey and Paul Kelleher for their comments on some of the new material in this edition. Hausman had the opportunity during the fall of 2015 to teach a draft of this edition to a wonderful group of University of Wisconsin undergraduates:  Colin Barushok, Jesse DeLap, Lindsay Doman, Jonah Jaxen, Ethan Kay, Mikhail Kuczmarski, Timothy McGuire, Andrew Murdoch, Jeong (Peter) Park, Adrian Rice, Jacob Stack, Joseph Stadler, Daniel Stoehr, Jard xix

newgenprepdf

xx

Acknowledgments

van Heerde, Steven Walker, Justin Walters, and Laura Wilk. There was also a senior auditor, Wesley Foell. The class members were very patient and willing guinea pigs, and we are indebted to them for their questions, reactions, and suggestions. Satz also taught a class with Alan Ryan in which chapters in the earlier edition were read and discussed. She is grateful to all of the participants for their feedback. This book traces back to a survey essay Hausman and McPherson published in the Journal of Economic Literature in 1993 in response to an invitation from John Roemer, who is in effect the grandfather of this work. Since then many people have contributed to this project in addition to those named here. No doubt errors and misconceptions remain; we reluctantly claim these as our own.

ONE

Ethics and Economics?

Economic Analysis, Moral Philosophy, and Public Policy is concerned with economics and ethics, but it is not about how to behave ethically when doing business or undertaking economic research. We prescribe no code of conduct and preach no sermons. Rather, in this book we try to show how understanding moral philosophy can help economists to do economics better and how greater knowledge of economics and ethics can help policy analysts to improve their evaluations of alternative policies. We also hope to show how philosophers can make progress in ethics by drawing on insights and analytical tools from economics. While we are writing mainly for those who are interested in economics and aim at helping them to think more deeply about this field, we think that economics has some important concepts and methods to offer to those interested in ethics, too. This focus may seem a big letdown. Surely it is critically important to grapple with the moral problems that arise inside corporations and businesses, as well as in the global economy. You will get no argument from us about that. We are not proposing that people stop asking moral questions about the justification of employee and CEO compensation, the use of “sweatshop” labor, the rationale for protectionist policies, and the legitimacy of using economic boycotts to make policy change. On the contrary, we hope this book will show how important morality is in economic life. But our primary concern is with economic theory rather than directly with economic life. Our job will be to show clearly the role that ethics has in economics and policy analysis and to demonstrate how knowing moral philosophy helps one do economics and policy evaluation better. In our view, the main value of moral theories does not lie in prescribing what to do in particular situations. Moral theories are not cookbooks for good behavior. Their main purpose is to help people to understand what morality is, where it fits into human life, and why people assign it the 1

2

Ethics and Economics?

importance they do. Moral theories have a practical role in guiding people’s reflection on the moral principles they accept and in thereby helping people decide what to do when their moral principles conflict with each other. Similarly, understanding ethics can help economists to think productively about the moral dimensions of policy problems, and it can bolster their confidence in recognizing and dealing with these moral issues. Knowing some ethics can help economists and policy analysts to improve their methods of policy evaluation and to understand how people’s economic behavior is influenced by the moral dimensions of their lives. Moral insights are, to be sure, more important to some parts of economics than others. Though not entirely irrelevant to any human choices, moral ideas are of little help in forecasting the price of wheat or in refining theories of exchange rate determination. Moral ideas will be more important to economists who face problems such as improving the standard of living in poor countries, increasing tax compliance, or helping citizens think through the trade-​offs between environmental protection and economic growth.

1.1  Economics and Morality in Contemporary Controversies This third edition of our book makes its appearance roughly a decade after the dramatic economic events that unfolded in the wake of the bursting of the housing bubble in 2007. Questions raised by these events, as well as the greater attention to the growing inequality between the rich and everyone else they have helped provoke, illustrate the close relationship between economic analysis and moral judgment that this book explores. Serious thinking about economic policy demands both economic and moral sophistication. The U.S. economy (followed by those of much of the rest of the world) experienced a spectacular disruption in 2008. The value of financial assets plummeted, and unemployment soared to levels not seen since the Great Depression. The shock was all the greater because it followed a period of more than twenty years of relative stability in prices and employment sometimes labeled as “the Great Moderation.” During this period of macroeconomic stability under Democratic and Republican governments, interrupted only by the intense, but brief, “dot-​com” recession of 2001–​2002, mainstream economists grew confident that deregulation of financial markets coupled with skillful control of the money supply by the Federal Reserve had secured stable (if modest) economic growth, employment, and

1.1  Economics and Morality in Contemporary Controversies

3

prices. A great many economists regarded the great problem of maintaining a steadily growing economy as solved. The progressive deregulation of the financial system that began in the 1980s led in the new century to a bewildering variety of financial instruments that were supposed to allow for an optimal distribution of risk among those who were best equipped to bear it. But once a housing boom (spurred by irresponsible mortgage lending) collapsed, it quickly became apparent that the entire economy had been very heavily leveraged on risky debt that no central agency was monitoring. The resulting crisis destroyed trillions of dollars of assets on the books of homeowners, businesses, and large financial institutions. The financial contraction led to rapid reductions in business investment and consumer spending, and it plunged the real economy –​that is, employment and production –​into the deepest recession since the Great Depression. These rapidly cascading developments confronted legislators and policy makers with pressing economic and moral challenges. Consider these three: 1. What help, if any, should the government provide to homeowners who were suddenly unable to pay the mortgages on their homes, whether through job loss, deceptive lending practices, or because of having gambled on purchasing an unaffordable home with the expectation of reselling it at a profit? Was it fair to help those who took the biggest risks, while doing nothing for those who financed their homes more conservatively and therefore did not face foreclosure, but who still lost much of the value of their homes? Was it fair to let the victims of the housing crisis (some more “innocent” than others) suffer? Releasing them from their obligations might create what in the world of insurance is called a “moral hazard”: encouraging reckless borrowing on the belief that the government would bail people out if their speculations miscarried. (In the end, federal efforts to help homebuyers proved to be limited and halting.) 2. In the event of nonpayment for most mortgages, the lender can seize the house, but the lender has no legal claim over any other of the homeowner’s assets. That means that homeowners whose property is worth less than what they owe can walk away from their debts by simply stopping paying their mortgages and letting the bank seize the property (often after staying on for free for some additional months). There may be prudential reasons for homeowners not to do that, including harm to their credit ratings. But is there a moral obligation to keep the repayment promise implied in taking the mortgage?

4

Ethics and Economics?

Bankers and community leaders tended to argue that there is such an obligation, both as a fundamental moral requirement and on the grounds of the harm that abandoning homes causes to neighborhoods (White 2010). 3. The third moral challenge requires some further background. In 2008, the drop in the value of the very risky assets held by major banks and financial institutions made them illiquid (without the cash needed to maintain day-​to-​day activities) and in some cases insolvent (with debts they could not pay). Was it appropriate for them to receive financial assistance (a “bailout”) from the federal government in weathering the storm? In letting Lehman Brothers fail, rather than finding some way to save parts of the firm (as the government had in the earlier case of another investment bank, Bear Stearns), officials invoked the principle that investment banks should accept the downside of the serious risks they chose to take. As it turned out, permitting Lehman Brothers to fail broke the confidence of the market and led to a furious effort by other financial institutions to sell off their risky investments and to try to build liquidity. This effort drove down the prices of those investments, reduced the liquidity and risked the solvency of these institutions, and nearly caused the collapse of the entire financial system. Federal officials facing the crisis, and chastened by the consequences of their decision regarding Lehman Brothers, concluded that the risks to the financial system as a whole were too great to permit any of the remaining large financial institutions to fail. The government consequently loaned large amounts of money to the major banks and nonbank investment firms to enable them to keep operating. This decision, made rapidly and with little public deliberation, was widely seen as unfair, if not outrageous –​basically rewarding the companies and individuals that had caused the crisis. The contrast between the feeble response to homeowners with troubled finances and the lavish assistance provided to wealthy bankers and their firms sharpened the sense of unfairness. Federal officials defended the action not on grounds of fairness but of economic necessity, asserting that failure of these institutions would have caused intolerable economic hardship across society. Were they right? In answering this question both moral values and empirical consequences come into play. Many people, not unreasonably, regarded the decision to insulate the banks and their leaders from the consequences of their high risk investment strategies as unfair. In addition, the

1.1  Economics and Morality in Contemporary Controversies

5

bailouts created or strengthened a serious moral hazard. If banks believe that they are “too big to fail” –​that their importance to the financial system implies that they will always be bailed out when they face large losses –​then they will be more inclined to undertake reckless investments. Thus, even if major financial institutions could not be allowed simply to collapse, there were economic as well as fairness reasons to make the terms of a “rescue” of the banks less favorable to the banks than they in fact were. The government might, for example, have pursued a federal takeover of major banks and have compelled both a change of leadership and a limit or ban on bonuses. Financial institutions could have been required to establish procedures to adjust unaffordable mortgages. Clearly, any alternative would have had pros and cons and uncertainties of its own. But the result was that tens of millions of Americans saw the government as protecting the wealthy and powerful malefactors responsible for the financial crisis while turning its back on the vulnerable. Partly as a result of this perception, the Great Recession focused public attention on the huge gap in income and wealth between Wall Street financiers and ordinary Americans –​attention spurred by the short-​lived Occupy Wall Street movement with its slogan “We Are the 99%.” While inequality in incomes and wealth was already increasing significantly during the last quarter of the twentieth century, those who believed that a growing economic pie offered a larger piece for everyone found that reality easy to ignore. As the Nobel Laureate Robert Lucas (2004) put it, “Of the tendencies that are most harmful to sound economics, the most seductive, and in my opinion the most poisonous, is to focus on the question of distribution… The potential for improving the lives of poor people by finding different ways of distributing current production is nothing compared to the apparently limitless potential of increasing production.” The perceived disparity in the treatment of rich and poor in the wake of the financial crisis, coupled with new findings on the distribution of income, wealth, and economic opportunity, undermined the dismissal of distributional concerns that Lucas espoused. In particular, the economists Emmanuel Saez and Thomas Piketty, in partnership with others, reported historical data on trends in income and wealth inequality for the United States and other countries that revealed that these inequalities, which dropped precipitously during the Great Depression of the 1930s and remained relatively low until the 1970s, had been increasing rapidly, with incomes and wealth at the very top expanding most rapidly, while declining or holding steady for a large majority. For example, in the 1960s, CEOs earned princely incomes roughly 25 times as much as that of the average

6

Ethics and Economics?

workers in their firms. They now make imperial incomes 250 times as much. The top 20 percent of the U.S. population has seven-​eighths of the wealth, and half of that belongs to the top 1 percent. These data are of concern both to citizens and to scholars. New work has taken up the challenge of explaining the reversal of a generations-​long pattern of declining inequality and providing policy suggestions for limiting inequalities. For example, Piketty argues in his best-​selling Capital in the Twenty-​First Century that capitalism has an inbuilt tendency toward increasing inequality that was interrupted by the extraordinary events of two world wars and a great depression. These exceptional occurrences destroyed enormous amounts of wealth held by the very affluent and strengthened the hand of workers and soldiers in claiming benefits from the economic system. In Piketty’s view, the best solution to the inequalities that capitalism engenders lies in government intervention in the market, largely in the form of taxation at the top and benefits at the bottom. The theoretical and empirical issues raised by inequality are far from resolved and engage challenging questions of morality and justice that are entwined with them. Here are some of the remaining points of controversy: 1. In explaining the rapid growth of inequality of income and wealth over recent decades, some analysts have emphasized changes in technology and the growth of globalization, while others have emphasized political factors, including changes in the tax-​transfer system and antiunion efforts. (The “Reagan revolution,” which weakened unions and cut the top tax rates by half, coincided with the beginning of the great divergence between the fates of those at the top and the rest of the population.) Who is correct? And, if growing inequality is the result of “impersonal” forces of technology and globalization, are the outcomes any less morally problematic than if the greater inequality results from political decisions? Should the source of these inequalities matter to our moral judgment of them? What is wrong with income and wealth inequality? Is such inequality a problem only for its effects on productivity and efficiency, as some have argued? Or for other reasons, including reasons of fairness? 2. Some scholars have argued that focusing on the extremes of wealth and income –​the 1 percent –​distracts attention from the real problems of intergenerational mobility and equality of opportunity. There is, for example, considerable evidence that growing income and wealth inequalities are systematically related to growing disparities in educational resources and students’ performance in school (Duncan

1.1  Economics and Morality in Contemporary Controversies

7

and Murnane 2011). This is not simply a problem for the tail ends of the distribution: the historically high earnings premium associated with added years of schooling contributes to growing earnings gaps and low intergenerational mobility across the entire range of incomes (Autor 2014). Should the focus be on opportunity and mobility? According to Saez and Zucman, as reported in the Economist on November 8, 2014, “The 16,000 families making up the richest 0.01%, with an average net worth of $371million, now control 11.2% of total wealth –​back to the 1916 share, which is the highest on record.” Is the problem with these inequalities the extent to which they limit opportunity for so many others, or instead should we be worried about the power that such wealth confers on a small number of individuals? 3. It is often suggested that great fortunes are less morally problematic if they are generated by an entrepreneur in his or her own lifetime than if they have been inherited. Piketty anticipates that in the twenty-​first century inherited wealth will grow in importance relative to new wealth, leading to a “patrimonial capitalism” that he finds distasteful, although he does not elaborate an explicit moral view. Why might entrepreneurial wealth be less troubling ethically than inherited wealth? One reason is that encouraging the accumulation of vast entrepreneurial wealth provides incentives for innovation and investment. For example, Mark Zuckerberg might have been less motivated to develop Facebook if he had faced higher taxes. Others have argued that successful people “deserve” their returns, quite apart from any incentive effects. An opposing view grants that inequalities may be justified when they contribute to the well-being of the least well off, but argues that it is hard on these grounds to justify fortunes anywhere near as great as those the top .01 percent enjoy. The purpose of this section has been to underline the contemporary relevance of this book’s inquiry into economics and ethics by sketching the inescapable links between the ethical and economic questions raised by the financial crisis and contemporary inequalities. We have not tried to answer these questions, but rather to suggest that addressing them requires systematic analysis and reflection of the kind to which a large part of this volume is devoted. Economic Analysis, Moral Philosophy, and Public Policy is not, however, concerned only with the interplay between economics and ethics in

8

Ethics and Economics?

the assessment of public policy. It also discusses the ways in which moral assumptions are embedded in purportedly purely scientific economics, the relationship between morality and rationality (particularly as understood by economists), the moral limits on markets, some of the many moral frameworks that economists tend to avoid, and the tools that economists have provided for ethical inquiries.

1.2  What Are Moral Questions and How Can They Be Answered? Moral questions and moral reasoning can be difficult to understand, and we have found that students often hold very skeptical or even cynical views. One hears claims such as “Morality is just a matter of how you feel.” “There’s no rational way to resolve moral disputes.” “Moral claims cannot be true or false.” “Morality is just a matter of social convention or prejudice.” These views seem to have some foundation. • It might seem that morality is just a matter of individual feeling and that moral disagreements cannot be rationally resolved, because it is hard to understand how moral claims can be tested, confirmed, or disconfirmed. • It might seem that moral claims cannot be true or false, correct or incorrect, because moral claims are often prescriptions that concern how things ought to be rather than how they in fact are. • It is tempting to believe that moral claims are conventional or relative, because members of different societies disagree about morality. Yet these skeptical conclusions are exaggerated, and they yield implications that are hard to accept. To see why, consider a genuine moral question that might face an individual: Sarah’s abortion decision. Sarah is a young woman attending college who becomes pregnant and is trying to decide whether to have an abortion. Although some of her friends see no moral problem with doing so and advise her to do whatever she thinks best for herself, Sarah has moral qualms. She is in doubt about whether abortion is morally permissible.

Saying that Sarah has “moral” qualms means that she is not just asking what would please her family or conform to the norms of her community. She is asking what would be the right thing to do for anyone whose circumstances are in the relevant regards just the same. Sarah’s problem is not a legal problem. She knows that abortion is in fact legal. But this does not tell her whether it is morally permissible. It is legal to be rude

1.2  What Are Moral Questions and How Can They Be Answered?

9

to your parents or to pretend to love a person in order to seduce him or her, but that does not mean that these actions are morally permissible. Second, notice that Sarah’s question is not one that a sociologist can answer. Even if she reads that 62.37 percent of her fellow citizens think that an abortion is permissible in circumstances like hers, her problem has not been solved. Sarah still needs to decide whether she ought to have the abortion. The third point to recognize is that Sarah’s question is a real question, something that she might agonize over. Whether reflecting by herself or talking over her dilemma with friends or family or counselors, she will be thinking about reasons why she should conclude that abortion is or is not morally permissible. Whether or not one believes that morality is subjective (in some sense of this ambiguous term) or that morality depends in some sense on feelings, there is unquestionably a huge potential role here for argument and judgment. It seems that some answers to Sarah’s question are better than others  –​for example, they may be more logically consistent or better supported by evidence –​and it is possible to think rationally about which answers are better and which are worse. There are genuine moral questions about social policy as well as about individual choices. For example, the question about whether abortions should be legal cannot be decided by ascertaining what the law currently is. The moral question of what the law concerning abortion ought to be must also be distinguished from questions about whether laws permitting or banning abortion are constitutional. Before the Thirteenth Amendment was passed, the U.S. Constitution specifically permitted slavery. That made slavery constitutional, but it did not make it just. Questions about what the Constitution ought to say are moral questions. One also cannot decide whether abortions ought to be legal by means of sociological research, such as taking a poll. A poll can determine what most people believe, but it will not determine whether they are right. Those who believe that abortions ought not to be legal cannot be refuted by results of polls showing that most people believe that they should remain legal. One addresses moral questions instead by making arguments. Once one recognizes these truisms  –​that moral questions have better and worse answers, and that arguments can sometimes help people find out which answers are better –​one can see that the cynical or relativistic conclusions concerning morality are exaggerated and unjustified. • It is not true that there is no method of resolving moral disagreements. There is a method: one can make arguments; that is, one can look for

10

Ethics and Economics?

premises that others agree on and then use logic to try reaching agreement on the issues in dispute. To be sure, when people stand to benefit from doing evil, they may be deaf to rational argument. It took a civil war in addition to the arguments of abolitionists in order to eliminate slavery in the United States, but without those arguments (to which there were, in fact, no good responses) there would not have been a movement opposing slavery. While some moral disagreements may remain, others can be resolved through argument and evidence. • The fact that some moral judgments are prescriptive  –​that they say how things ought to be rather than how they are  –​may imply that some moral judgments cannot be literally true or false, but it does not follow that one cannot sensibly consider whether moral judgments are mistaken. When people assert that “killing people for fun is wrong,” they certainly take themselves to be saying more than just “I’m personally not into killing people for fun.” The idea that killing is a moral wrong plays a different role in their thought and action than a mere whim or feeling would. Moreover, even if moral judgments are not descriptive assertions, the reasons for those judgments often include empirical claims that can be criticized and investigated. • Though moral questions are not always easy to answer and though difficult questions give rise to persistent disagreement, there is also a lot of agreement in ethics. Because people only argue about those claims upon which there is disagreement, it is easy to overlook how much people agree about. Few people approve of torture for any purpose, and even fewer approve of it for the purposes of entertainment. The claim, which is often heard, that morality is “relative” is ambiguous. In one sense morality clearly is relative: what is right to do depends on (is relative to) what the facts are. Whether it is permissible to knock over a frail old man depends on whether one knocks him over to see whether his bones are brittle or whether one knocks him over to prevent him from being run down by a truck. But to recognize that one does not have a well-​defined moral prescription until one has specified all the facts is perfectly consistent with the idea that moral questions can have better and worse answers. What people mean by claiming that morality is relative is often something altogether different: It is sometimes suggested that whatever a person (or a society) believes is right is automatically right (“for that person or society”). But when Sarah is trying to decide whether it is morally permissible to have an abortion, she is not trying to find out what her beliefs already are; she is trying to find out which answer to her question she should affirm.

1.2  What Are Moral Questions and How Can They Be Answered?

11

Similarly, when thinking about whether abortion should or should not be legal, people are not trying to find out what they (or others in their society) already believe but instead whether the law ought to permit abortions. If whatever people believed about ethics were automatically right, then there could be no moral disagreement. To disagree with someone about a moral question commits you to believing that people’s ethical beliefs can be incorrect. Similarly, if a social consensus guaranteed its own correctness, then defenders of unpopular views would automatically be mistaken. One would not need to argue with defenders of minority views, since they could not possibly be right. But iconoclasts cannot be refuted with polls, and social consensus is not proof of correctness. We recognize how tempting it is to think that there is no fact of the matter about morality and that, even if there were, people could not know it. Morality seems in large part a human construction, so it is easy to jump to the conclusion that it is mere social convention or, more radically, that individuals determine what is right or wrong by what they believe or feel. But these temptations lead either to moral nihilism –​the denial that anything is right or wrong –​or to other views that cannot be sustained. If you think that anything is right or wrong, good or bad, morally praiseworthy or blameworthy, then you are not a moral nihilist. And if you take any moral claims seriously, wonder whether they are correct or incorrect, and sometimes disagree or argue with others, then you cannot believe that all moral views are on a par and that there can never be any reason to accept some and reject others. Sometimes people feel that it is intolerant or dogmatic to believe that their moral convictions are correct. This is a mistake. Dogmatism is not believing that one is right; it is believing that one could not be wrong. There is nothing inconsistent about holding moral convictions while recognizing that you are fallible and might learn something from those who hold different views. Intolerance, unlike dogmatism, concerns action. Whether tolerance is a virtue and what views and actions should be tolerated are questions within morality. Some moral theories are tolerant while others are intolerant, just as some people –​whether the subject matter be morality or sports –​are dogmatic and others are ready to listen and learn. Short of giving up morality altogether, including all concerns about tolerance, there is no alternative to taking one’s moral beliefs seriously. People who are genuinely tolerant are not moral skeptics: they believe that those who are intolerant are wrong to be intolerant. Tolerance is tied to an appreciation of the richness of different cultures and different life experiences, to a respect for others, and to a willingness to take their perspectives and arguments

12

Ethics and Economics?

seriously. If there really is nothing right or wrong in moral matters, then there is nothing wrong with intolerance. There is nothing dogmatic or intolerant about believing that some answers to moral questions are better than others and that rational argument can help one to judge which answers are better. These beliefs are implicit in our everyday individual moral judgments and in our policy making, and it is hard to deny them without denying that there is any such thing as morality.

1.3  How Is Moral Philosophy Relevant to Economics? The idea that studying ethics could help people to do economics or policy analysis may seem far-​fetched. Why not consult Tarot cards instead? Many people –​indeed, probably some of you reading these very words –​doubt that moral philosophy can help one do anything better (except perhaps to spread confusion and cure insomnia). If one is seeking clarity, why look in a swamp? Our hope is that readers of this book will not leave with the impression that argument in moral philosophy is obscure, unworldly, or boring. It is, to be sure, intellectually demanding, abstract, and often intricate; like economics itself, moral theory is loaded with controversies and unresolved issues. We want most of all to show that moral reasoning can help people gain a surer grip on serious problems about how to make their lives and our society better. But even if moral philosophy clarifies morality, why should it clarify economics? Many economists would draw a sharp distinction between evaluative questions and the “positive” science of economics, which is concerned with facts, not values. In the 1930s, Lionel Robbins expressed this view as follows:  “It does not seem logically possible to associate the two studies [ethics and economics] in any form but mere juxtaposition. Economics deals with ascertainable facts; ethics with valuations and obligations” (1935, pp.  148–​9). Robbins is drawing on a commonsense distinction (which is maddeningly difficult to make precise) between factual claims and evaluative claims. Intuitively, there is a huge difference between describing how many tons of steel the United States imported in 1999 and saying whether importing foreign steel is a good or a bad thing. Although there is, we believe, no way to draw the distinction between facts and values precisely, it is worth describing how some influential philosophers and economists have distinguished them. Table 1.1 summarizes the contrasts. We should stress that these contrasts are highly controversial

1.3  How Is Moral Philosophy Relevant to Economics?

13

Table 1.1.  Exaggerated contrasts between facts and values Factual claims

Evaluative claims

Disagreements can be resolved by evidence Relatively little disagreement Descriptive: say how things are True or false Objective Independent of evaluative claims Bear on action, only given motives

No good way to resolve disagreements Relatively little agreement Prescriptive: say how things ought to be Not true or false Subjective Dependent on factual clams Guide action: logically tied to motives

and that (as we shall argue in the Appendix) the distinctions are exaggerated and in some cases mistaken. Drawing these contrasts in this overly sharp way makes it easy to see why economists have thought it was so important to argue that economics is and should be “value neutral”: that it makes and should make no evaluative claims. Many economists believe that only factual claims can be studied by the methods of science, and their status as scientists is important to their self-​understanding and to their social prestige. In this view, economists trespass beyond the boundaries of science and of their proper social role when they take stands concerning evaluative matters. This view of the separation between scientific claims, which are the subject matter of economics, and evaluative claims, which cannot be the subject matter of any science, might be expressed as follows:  “Economics is science or engineering. It shows how to arrive at certain goals but, unlike ethics, it does not prescribe what goals one should have. Economics provides technical knowledge that has no more to do with ethics than does geometry or physics. No matter how sensible and well-​conceived ethical theories may be, they have nothing to do with economics and cannot possibly help one understand economies.” This entire book is a response to the view that ethics and economics have (and should have) nothing to do with each other. The best case for relating economics and ethics is to show that linking the two subjects is productive in the practice of those disciplines. We begin to make this case in Chapters  2 and 3 by showing through examples how unproductive it is to try to “cleanse” economics of the evaluative content of familiar economic ideas such as efficiency, welfare, and freedom. Instead of beginning this book with abstract philosophical considerations concerning the relations between economics and ethics, we have postponed that discussion

14

Ethics and Economics?

to an Appendix in which we directly challenge the claim that economics and ethics should be sharply separated. There we also criticize explicitly the “engineering” vision that portrays economics as entirely value neutral. In the main text of this book we try to be constructive, showing concretely how evaluative and factual matters are entangled in economics and policy analysis.

1.4  Organization of the Book Chapters 2 and 3 present four examples that illustrate ways in which moral questions arise in economics, and Chapter 15 returns to these examples and applies the concepts, distinctions, and principles developed in the intervening chapters. Chapter 16 reflects upon some of the new challenges that recent economic transformations and their consequences pose for mankind and shows how the analyses of this book help to clarify them. The chapters in between –​Chapters 4 through 14 –​are divided into four parts. Part I focuses on rationality, morality, and markets. Like morality, rationality is normative. One ought to be moral and one ought to be rational. One is wicked if not moral and foolish if not rational. Rationality, unlike morality, plays an explicit and conspicuous role in contemporary economic theory. Economists usually deny that economic theory presupposes any ethical views, but they freely admit that it presupposes a great deal about rationality. However, economists cannot have it both ways. Endorsing their theory of rationality, we will argue, already commits them to controversial moral principles. In defending their model of rationality, economists wind up espousing fragments of a moral theory. It also turns out, we shall argue, that exploring the connections between morality and rationality leads to criticisms of economics, because the moral principles implicit in standard views of economic rationality are implausible. When these principles are stated explicitly, few people would endorse them. The standard views of rationality held by economists also make it hard to understand the rationality of norms and of morality itself. Without an appreciation of the importance of morality and the constraints it imposes on self-​interest, it is hard to understand the limits that all societies impose on the extent of the market. Parts II and III then zero in on concepts and theories of economic evaluation. Which economic policies and institutions are best? How should they be judged? Part II focuses on the standard theory of welfare and on methods of evaluating economic outcomes and institutions in terms of their consequences for welfare, such as welfare economics and utilitarianism. We

1.4  Organization of the Book

15

shall criticize both the preference satisfaction view of welfare that appears to be implicit in normative economics and contemporary work that takes welfare to be happiness. We shall show how economists can avoid committing themselves to any philosophical theory of welfare. Although we argue later that economists should be concerned with other ethical considerations, such as freedom, rights, equality, and justice, in addition to welfare, we do not doubt that welfare is of great moral importance, and a major aim of Part II is to clarify its role. Part III is mainly concerned with freedom, rights, equality, and justice. These notions are important in the evaluation of economic policies and institutions, and moral theories have been built around them. When one shifts one’s attention away from an exclusive preoccupation with welfare, new vistas appear. There arise new questions, new aspects of economic arrangements to consider, and new methods of thinking about morality. We shall in particular explore libertarianism, egalitarianism, and contractualism, each of which provides a way of making sense of morality in which the consequences of policies for individual welfare are not necessarily decisive. Whereas Part II uses the concepts of standard welfare economics, Part III presents alternatives to the questions that welfare economists ask and to the terms in which they answer them. Part IV provides an introduction to technical work in economics that is directly guided by ethical concepts and relevant to moral theorizing. The payoffs from knowing something about both economics and philosophy do not go only to economists. Philosophers have lessons to learn from attempts at formalizing moral concepts and exploring their consequences. The concluding chapters consider the benefits of harnessing the combined powers of economics and ethics in addressing important issues. In Chapter 15, we revisit the four cases discussed in Chapters 2 and 3 and try to show, through these and other examples that we discuss in Chapter 16, how the ideas and tools explained in this book might help identify good policies and principles for citizens and governments to adopt. At the end of each chapter (except the final two), we provide a brief discussion of relevant literature in Suggestions for Further Reading, and each chapter is followed by Questions for Study and Discussion. A Glossary of relevant terms is assembled at the end of the book. When you are finished with this ramble through the lush forests of moral philosophy and the brushland where it meets verdant plains of economics, we believe that you will see that economics remains partly, and necessarily, a moral science. It cannot be done without moral presuppositions, and it is hard to do it well without addressing moral issues

16

Ethics and Economics?

intelligently. Similarly, moral philosophy cannot be done without engaging with human psychology and the structure of human interactions, and it is hard to do it well without knowledge of the kind that economists seek. Like those who would disavow the culture of their parents, economists sometimes try to deny their philosophical lineage. Although they can reform and improve their philosophical inheritance, they cannot escape it, and attempting the escape renders their theories hollow. Neither can the philosophical parents of today’s social disciplines successfully repudiate their offspring. Moral philosophy and economics have much to contribute to each other.

Suggestions for Further Reading The financial crisis of 2008 and the ensuing recession have produced a voluminous literature. A lucid analysis of the problem of the risky behavior of banks and the resulting moral hazard is Admati and Hellwig (2013). An excellent general account of the events of the crisis years is Blinder (2013). Thomas Piketty’s best-​selling book of 2014 provides a rich economic history of wealth inequality in a number of countries and a case for expecting inequality to grow persistently in capitalist societies. Atkinson (2015) analyses income inequality in the United States and United Kingdom and advocates policies to reduce it. Stiglitz (2013 and 2015) argues that excessive inequality undermines the economic growth and stability of societies and advocates policies to reduce inequality. Evidence about growing impacts of material inequality on inequality of educational opportunity is presented in a number of studies in Duncan and Murnane (eds.) (2013). A case for the claim that social mobility is likely to fall drastically in coming decades is made in Putnam (2015). For a brisk introduction to ethics with a useful discussion of the ways in which it is objective, see Shafer-​Landau (2003). For another book-​length treatment of the relations between economics and ethics see Wight (2015).

Questions for Study and Discussion 1. It strikes many people that bailing out the financial institutions that caused the economic collapse of 2008 was unfair. If doing so was important for prosperity, should we worry about its unfairness? How important do you think that fairness is?

Questions for Study and Discussion

17

2. Do you think that it is morally permissible for homeowners to walk away from their mortgages and not repay the money they borrowed, if they can? 3. Should the same moral rules apply to banks and investment firms that apply to individuals? 4. Should leaders of major financial institutions be held morally culpable as individuals for decisions they make in their corporate roles? 5. What do you think of Lucas’s claim that distribution is unimportant? 6. Why should we care whether a relatively small portion of the population is growing fabulously rich? Is not doing so just a case of envy? 7. Should it matter to our moral assessment of inequalities whether they result from inheritance or entrepreneurial activity? 8. What does possessing great wealth –​let us say wealth in excess of $100 million –​contribute to life? After Mark Zuckerberg already had $100,000,000, what more could money do for him? In what ways would those who are extremely rich be wronged or harmed by extremely high taxes? Even if a ceiling on the wealth an individual could amass did that person no harm, are there reasons why such a ceiling might hurt other members of society? 9. Chapter 1 mentions that many students feel that • Morality is just a matter of how you feel. • There is no rational way to resolve moral disputes. One can only fight. • Moral claims cannot be true or false. • Morality is just a matter of social convention. Would you accept one of these claims? How would you defend it? 10. A question such as “Should there be laws limiting greenhouse gas emissions?” is concerned with the law, but it is a moral question, not a legal question. What is the difference? 11. There appears to be a consensus in some societies, such as Saudi Arabia, that women’s lives should be largely controlled by men and that women should have fewer rights than men. Do you think that the answer to the question of whether women are morally equal to men depends on the society in which the question is asked? 12. Suppose that at the grocery store you see a man severely beating a five-​year-​old. You try to intervene, and the man says that his religion calls for him to beat Satan out of his children, and that his son took a box of cookies off the shelf without his permission. Would it be intolerant of you to intervene? Should you intervene even if it is intolerant to do so? Does intervening mean that you are an

18

Ethics and Economics?

intolerant person? Does intervening commit you to suppressing this man’s religion? Does it matter what country you are in when this is happening? 13. Consider the following statement:  “Markets are the most efficient way of distributing goods and services.” Is this a factual or an evaluative claim?

T WO

Ethics in Welfare Economics

This book will be filled with arguments, but we shall begin with examples to bring them down to earth. One good example may do more to clarify how ethics matters to economics than would a hundred pages of argument. Furthermore, ethical reasoning is not just a matter of logic. Emotion has its part to play, too, and examples help to engage the emotions. In this chapter and the next, our concern is not to argue that ethics matters in economics but to exhibit through examples how important ethics is. In this chapter we will focus on two examples, which will enable us to identify the main moral assumptions that characterize mainstream normative economics. The first example caused an uproar.

2.1  A Shocking Memorandum Consider the following confidential World Bank memorandum, which was published in The Economist (February 8, 1992, p. 62). Larry Summers, then the World Bank’s chief economist, absorbed the criticism the memorandum provoked, but he was probably not the author: Just between you and me, shouldn’t the World Bank be encouraging more migration of the dirty industries to the LDC’s [less developed countries]? I can think of three reasons: (1) The measurement of the costs of health-​impairing pollution depends on the foregone earnings from increased morbidity and mortality. From this point of view a given amount of health-​impairing pollution should be done in the country with the lowest cost, which will be the country with the lowest wages. I think the economic logic behind dumping a load of toxic waste in the lowest-​ wage country is impeccable and we should face up to that. (2) The costs of pollution are likely to be non-​linear as the initial increments of pollution probably have very low cost. I’ve always thought that under-​populated 19

20

Ethics in Welfare Economics

countries in Africa are vastly under polluted; their air quality is probably vastly inefficiently low [sic] compared to Los Angeles or Mexico City. Only the lamentable facts that so much pollution is generated by non-​tradable industries (transport, electrical generation) and that the unit transport costs of solid waste are so high prevent world-​welfare-​enhancing trade in air pollution and waste. (3) The demand for a clean environment for aesthetic and health reasons is likely to have very high income-​elasticity. The concern over an agent that causes a one-​in-​a-​million change in the odds of prostate cancer is obviously going to be much higher in a country where people survive to get prostate cancer than in a country where under-​5 mortality is 200 per thousand. Also, much of the concern over industrial atmospheric discharge is about visibility-​impairing particulates. These discharges may have very little direct health impact. Clearly trade in goods that embody aesthetic pollution concerns could be welfare-​enhancing. While production is mobile the consumption of pretty air is a non-​tradable. The problem with the arguments against all of these proposals for more pollution in LDCs (intrinsic rights to certain goods, moral reasons, social concerns, lack of adequate markets, etc.) could be turned around and used more or less effectively against every Bank proposal for liberalisation. (quoted in the Economist, February 8, 1992, p. 66).

The memorandum is worth reading carefully, and we urge you to read it a second time before proceeding further. It is a provocative exploration of the implications of “economic logic” rather than a serious proposal for a World Bank program to export pollution to the LDCs. What makes this memorandum particularly worth studying is that its economic logic appears to be exemplary, yet its conclusions are alarming to many people.

2.2  Eight Distinctive Features of Welfare Economics The World Bank memorandum does not purport to be just a study of facts, which bear on policy only in the way that engineering bears on policy. On the contrary, this note makes claims about what the World Bank “should” be doing, and it describes some facts as “lamentable.” The memorandum clearly makes evaluative claims, and it would be excluded from economics by those who insist that economics must be free of any value judgments. Yet this memorandum obviously seems to be concerned with economics. One way to recognize this, while still insisting on the importance of distinguishing between factual and evaluative claims, is to maintain that there are two kinds of economics: “positive economics,” which deals only with matters of fact, and “normative economics,” which is concerned with the evaluation of economic states of affairs, processes, and institutions. This memorandum is clearly an instance of normative economics.

2.2  Eight Distinctive Features of Welfare Economics

21

We would like to draw your attention to eight features of this note, which are, we maintain, typical of mainstream normative economics or “welfare economics.” Each of these features represents a choice: its way of thinking about economic states of affairs and policies is just one of many possible ways available to us. Once one recognizes what distinguishes this way of thinking about outcomes and policies from other ways, one understands a great deal about this kind of normative economics. 1. We have already mentioned the first of these eight features: the memorandum evaluates economic states of affairs and recommends how to improve them. The focus is on economic outcomes rather than processes. 2. The memorandum assumes that there is a single framework for economic evaluation, which it takes for granted. It does not state this framework explicitly or argue for it. Though the author probably would not state the matter this way, he relies on an ethical foundation that he believes his readers share. 3. The memorandum takes the form of an argument in which premises concerning costs and demand are supposed to establish conclusions concerning what the World Bank should do and what facts are lamentable. That argument seems to draw on rather intricate economic and ethical reasoning. The ethical conclusions are not based on appeals to emotion or prejudice. 4. The memorandum considers how policies and states of affairs bear on individuals. No questions are asked about the significance of their other effects, such as those on the environment or local cultures, except insofar as those in turn affect the welfare of individuals. The note assumes that what matters consists entirely in the consequences of the policy for individuals. 5. The memorandum evaluates economic states of affairs in terms of their consequences for individual welfare rather than with respect to any other feature of individuals. It is concerned about whether transfers of pollution would be “welfare-​enhancing.” Owing to the prevalence of this feature, mainstream normative economics is typically called “welfare economics.” 6. In measuring welfare, the memorandum implicitly accepts the way that markets evaluate states of affairs, when (competitive) markets exist. 7. Although the memorandum focuses exclusively on welfare, it does not attempt to add up welfare gains and losses or to compare the welfare

22

Ethics in Welfare Economics

of different people. It does not claim that trade in pollution would maximize total or average welfare. Indeed, total or average welfare is never mentioned. 8. In addition to focusing exclusively on the welfare implications of shifting pollution, the memorandum suggests that there is a qualitative difference between the “impeccable” “economic logic of dumping a load of toxic waste in the lowest-​wage country” [our emphasis] and miscellaneous and unspecified ethical objections in terms of “intrinsic rights to certain goods, moral reasons, social concern, lack of adequate markets, etc.” The implication is that the welfare arguments are rigorous and worth taking seriously, while the miscellaneous ethical objections can be disregarded. (Otherwise, the claim that these objections “could be turned around and used more or less effectively against every Bank proposal for liberalisation” should lead him to criticize the World Bank’s proposals for liberalization rather than to make an additional one!) Welfare economists rarely deny explicitly that other moral considerations are relevant to evaluating policies and outcomes –​and indeed this memo makes no such argument. Welfare economists are, however, often suspicious of, impatient about, or even contemptuous of other ethical concerns. Some of these eight features of welfare economics are widely shared in the thought and culture of modern liberal democracies, while others are more distinctive to mainstream economics. Thus the presumption that social policies are subject to rational evaluation through argument, as in items 1 and 3, is widespread (though hardly universal), while the tendency to focus on outcome evaluation and downplay process is more peculiar to economics. Similarly, “individualism” –​in the particular sense of item 4, that the only sources of ultimate value are human individuals –​is a familiar feature of liberal social theory, while the idea that only the welfare of those individuals ultimately counts (item 5) is more specific to normative economics. These features of welfare economics, even those that are widely shared with liberal social theory more generally, are not inevitable. Each involves a choice, and each feature could be questioned or changed. These choices are both methodological and ethical. Although welfare is obviously very important, so is freedom and so is equality, and normative economics might focus on them in addition to or instead of welfare. There are alternatives, and to choose among them requires ethical reflection. One of our tasks, which will occupy much of Parts I and II of this book, is to lay bare the philosophical principles underlying welfare economics and

2.3  The Economic Benefits of Exporting Pollution to LDCs

23

then examine to what extent these principles constrain the character of normative economics. Another task, which will occupy Part III, is to consider alternative approaches. Rather than passing judgment, we aim to reveal the philosophical complexities implicit in policy evaluation and the range of choices that economists might reasonably make.

2.3  The Economic Benefits of Exporting Pollution to LDCs The last three of the eight features of the memorandum we highlighted are all concerned with the economic benefits of transferring pollution to less developed countries. In Part II we shall have a good deal to say about the notion of an economic benefit or cost, but a few words here will help clarify the argument in the memorandum. Air and water pollution lessen the quality of life in many ways, yet most kinds of pollution have no market prices. One cannot go to the hardware store and purchase for $19.95 a 20 percent decrease in the toxic chemicals in the air one breathes. Three factors explain why there are few markets in pollution limitation. First, it is not possible to locate all the sources of air and water pollution. (If your neighbors decide to dump left-​over weed killer down their drain, who will ever know?) Second, even if it were possible, it would be prohibitively expensive to strike a deal with all the polluters in order to improve your air or water. Third, any deal you strike with polluters will affect your neighbor, and vice versa: while walking to the corner, you have to breathe the same air your neighbor breathes. Any effective deal will require cooperation among your neighbors. Thus some collective action is often needed to deal with pollution. One way economists can help with the problems of controlling pollution is by imputing costs to it. The goal is to figure out what pollution costs would be, if there were markets where pollution could be bought and sold. Economists may attempt to impute pollution costs by examining housing prices in communities that are much the same, apart from their air quality. They can draw inferences from how much people pay for air filters, water filters, or bottled water. They can collaborate with biologists in determining and assessing the costs of damage to health caused by pollutants. In such ways, economists may be able to estimate how much people in developed countries would be willing to pay to lessen pollution in their environment and how much people in LDCs would have to be compensated in order to be willing to accept more pollution. People may be ignorant of the harms caused by pollution. People in a LDC may be willing to accept toxic wastes for very little compensation,

24

Ethics in Welfare Economics

because they are unaware of the contents of the wastes, the harms they might do, or the prospects of the wastes escaping and poisoning groundwater. Any willingness to accept more pollution that is based on such mistakes does not truly reflect what serves a person’s interests. The memorandum avoids this difficulty, because it argues that the willingness of those in LDCs to accept pollution for less compensation than those in developed countries are willing to pay rests on differences in the consequences of pollution in LDCs and developed countries. Economists employ complicated techniques to determine economic costs when things are not literally bought or sold, and it will simplify the discussion here if we engage in some make-​believe. Suppose pollution could be bought and sold on the market. Since pollution is a “bad,” consumers would want to buy its absence. Let us think of reductions of pollution that consumers could buy as units of a single hypothetical commodity, “environmental quality.” We can then think about what the price of units of environmental quality would be in this make-​believe world in which individuals could separately purchase and consume environmental quality while firms could separately produce and market units of it. One can then take the memorandum to be arguing that –​were it possible for individuals easily to buy, sell, and transport “environmental quality” –​ there would be active trading between the developed and less developed nations of the world, and pollution would be pouring out of the developed nations and into the less developed nations. Wealthy people would be buyers of environmental quality and poor people would be sellers for the same reasons that wealthy people are buyers of house cleaning services and poor people are sellers of them. But units of environmental quality cannot (“lamentably”?) be individually appropriated, bought, and sold, and it is hard to transport pollution between nations. Consequently, the World Bank can enhance welfare by simulating what a hypothetical market would accomplish and facilitating transfers of polluting industries to LDC’s.

2.4  The Argument of the World Bank Memorandum The three points in the memorandum aim to show that the amount individuals in developed countries would be willing to pay to lessen their pollution is more than the amount that individuals in LDC’s would demand in order to accept more pollution; also, there would be more pollution in LDCs if pollution were easily exchangeable. But why should one conclude that it is “lamentable” that pollution is not easily exchangeable? Why should one conclude, “Clearly trade in goods that embody aesthetic pollution concerns could be welfare-​enhancing [our emphasis]?” What makes it an

2.4  The Argument of the World Bank Memorandum

25

improvement to approximate more closely how things would be if pollution could be easily traded? How is one supposed to reach the conclusion that “the World Bank [should] be encouraging more migration of the dirty industries to the LDC’s”? How does one get from claims about hypothetical markets or willingness to pay to claims about welfare and from claims about welfare to claims about what the World Bank ought to do? We need to clarify the logic implicit in the memo. We suggest that the argument can be spelled out as follows: 1. Rational agents in LDCs would accept pollution from developed countries for less compensation than rational agents in developed countries would be willing to pay to get rid of the pollution. In other words, for some compensation C  –​that lies between the least that agents in LDCs will accept and the most that agents in rich countries will offer –​all rational individuals, whether in developed countries or in LDCs, would prefer to transfer pollution from a developed country to a LDC (premise 1). 2. Whatever individuals prefer makes them better off or increases their welfare (premise 2). 3. Shifting pollution to LDCs from developed countries and paying some compensation C makes everyone better off (from 1 and 2). 4. One should adopt policies that make people better off (premise 3). 5. One should adopt policies that shift pollution to LDCs and pay compensation C (from 3 and 4). If one assumes that the jobs and revenues provided by dirty industries are “reasonable compensation,” then this reconstruction captures what is meant by the claim that “the World Bank [should] be encouraging more migration of the dirty industries to the LDC’s.” In fact, as the memorandum notes, many transfers of pollution are not feasible, and individual units of pollution typically cannot be bought or sold. Some economists might argue that it would be a good thing to shift pollution to LDCs regardless of whether there is any compensation –​on the grounds that there is a “net benefit” in such shifts. But it takes more controversial moral premises to establish this stronger conclusion, and we will not comment on arguments that rely on the notion of a net benefit until Chapter 9. Statements 2 and 4 appear to be moral premises, and in later chapters we shall examine them at length. Premise 2 is particularly important in forging the link between market evaluation and welfare by identifying welfare with the satisfaction of preferences. This feature lies at the very core of mainstream normative economics. It is through that identification (plus the connection that positive economics establishes between preferences

26

Ethics in Welfare Economics Table 2.1.  The moral framework of normative economics

1.  What should economists appraise?   a. Outcomes   b. Processes 2.  What method(s) of appraisal should economists use?   a. Single method of appraisal   b. Multiple ethical perspectives, depending on problem 3.  What matters about outcomes?   a. Consequences for individuals   b. Consequences for groups, or the environment 4.  Which features of outcomes for individuals matter?   a. Welfare   b. Freedom   c. Rights   d. Justice 5.  What is welfare?   a. The satisfaction of preferences   b. Some mental state, such as happiness   c. “Objective” goods –​e.g., achievements, personal relations, health 6. How does welfare (as preference satisfaction) bear on the evaluation of outcomes?   a. Market evaluation and the Pareto concepts   b. Add up preference satisfaction 7.  What role do other ethical notions play?   a. Independent; important, but not a concern of economics   b. Their importance is derivable from their consequences for welfare   c. Must be integrated into the economic appraisal   d. Of no importance Note: Responses in italics are those of orthodox welfare economists

and market prices) that the memorandum links premises about costs and demands to conclusions about what outcomes are good or bad and what policies will enhance welfare. We especially want to emphasize that the tone of the memorandum (which is addressed to fellow economists, not to the public at large) is misleading. It suggests that the three numbered paragraphs make a “scientific” case, while the last paragraph mentions wishy-​washy moral objections. But the moral content does not wait for the last paragraph to make its appearance. The three numbered points of the memorandum above are already part of a moral argument. Table 2.1 lists some of the alternatives normative economists face and the specific choices orthodox welfare economists have made.

2.5  Should the World Bank Encourage Migration of Dirty Industries

27

2.5  Should the World Bank Encourage Migration of Dirty Industries to LDCs? The uproar caused by this memo suggests that many people are not willing to accept its conclusion. This may be a thoughtless first reaction. Why should not the World Bank encourage migration of dirty industries? Since the argument appears to be logically valid, those who reject the conclusion have to reject at least one of its premises. Though we cannot speak for all those who find the conclusion objectionable, here are five possible objections. (We will consider others in Chapter 15.) 1. Premises 1 and 3 tacitly assume that the consequences of pollution are local and that the amount of pollution is independent of where the pollution takes place. Both these assumptions are questionable. Encouraging dirty industries to migrate to LDCs might lead to more total pollution. Developed countries have both the incentive and the administrative capacity to enforce pollution controls, while LDCs have weaker incentives to limit pollution, and they may have a harder time enforcing pollution regulations. Since many of the consequences of pollution are global, the consequences of encouraging the migration of dirty industries to LDCs may be very bad. Although the first objection  –​that migration of dirty industries will increase worldwide pollution –​is a serious one, it need not challenge the memorandum’s moral framework, and we shall have little to say about it. To the extent that the effects of pollution are not localized, the arguments concerning the benefits of shifting pollution are not valid. If the harms of pollution are not mainly local, then one cannot conclude that the migration of dirty industries will be beneficial –​no matter how attractive they may be to the particular traders. 2. Even if people in both developed economies and LDCs would be happy to shift pollution to LDCs for some reasonable compensation, the exchange is unfair. Consider an agreement between a billionaire and a beggar whereby the beggar agrees to work sixteen-​hour days in exchange for gruel and a straw mattress. Both parties may be rational and well informed, and enter the agreement “voluntarily.” (As we will see in the Chapter 3, the idea that any exchange not directly involving physical force is voluntary is a controversial moral claim.) The fact that an exchange may be voluntary does not by itself justify it. For example, many might object that the billionaire is exploiting the beggar.

28

Ethics in Welfare Economics

This analogy may understate the injustice of transferring pollutants in LDCs, because some of the poverty in LDCs has been caused by the developed countries. For example, agricultural subsidies in rich nations have devastated agriculture in poor countries. Exploitation seems more egregious if the wealthy party caused the other’s poverty. This second objection would not surprise the author of the memorandum, who alludes to such objections in the last paragraph, but offers no response apart from the rather alarming observation that such objections “could be turned around and used more or less effectively against every Bank proposal for liberalisation.” This unfairness objection shows that premise 4, that policy should make people better off, needs qualification. It is plausible to maintain that, other things being equal, it is a good thing to make people better off. But other things are not always equal. It may not be good to make people better off if doing so is unfair. To point out that shifting pollution to LDCs may be exploitative does not show that it is wrong. Most people are prepared to put up with some injustice in exchange for a sufficiently large increase in welfare. Before deciding what is best, one needs to know how large the welfare benefits will be and how much unfairness will result. But normative economics only talks about welfare; it has nothing to say about how to think about trade-​offs between justice and welfare. One misleading feature of this memo is its suggestion that, apart from some moral bellyaching, the policy implications are simple. 3. The memorandum’s analysis compares only one alternative to the status quo: that of shifting pollution to LDCs. What about other policies that might be better still? Notice that the case for transferring pollutants depends on the huge income disparities between rich and poor countries: without those disparities why would people in, say, Africa pay less to avoid pollution than people pay in the United States? Should this status quo income disparity be simply taken as given? A much more effective policy for enhancing welfare might be to transfer income rather than pollution from rich to poor countries. 4. Satisfying preferences does not automatically increase welfare. People may prefer things that wind up being bad for them. They might do this because they lack information about the costs and benefits of different options, because they are choosing between options all of which have negative consequences for them, or because they have adapted their preferences to fit their circumstances. Voluntary exchange need not always be mutually advantageous.

2.5  Should the World Bank Encourage Migration of Dirty Industries

29

The basic point in this objection is hard to deny, but economists can argue that, given the actual consequences of increased pollution in LDCs, rational individuals in LDCs should be willing to accept pollution for less money than individuals in developed nations should be willing to pay. The reasons are that wages are lower in LDCs, low levels of pollution do little harm there, and poverty in LDCs makes clean air and water relatively less important. The greater willingness to accept pollution does not derive from miscalculating the effects of pollutants. The standard reasons why one might question whether satisfying people’s preferences would make them better off do not apply. 5. What about premise 1, that all rational and well-​informed agents would prefer to make the exchange? This premise is itself the conclusion of an argument from the fact that the economic costs of pollution are lower in LDCs than in developed countries. But do the economic costs and benefits capture what is morally relevant? Must rational and well-​informed individuals accept the market’s evaluation of the consequences of the pollution? Is premise 1 not a controversial moral premise, too? The moral relevance of market prices is questionable, even when they do not reflect any ignorance or irrationality. Individuals need not accept the market’s evaluation of the consequences of the pollution. If there were markets on which all varieties of environmental quality could be bought and sold, there would be massive transfers of pollution to LDCs. But those transfers would not necessarily increase human well-​being. Let us see why. People in LDCs are willing to pay relatively less to avoid pollution, because pollution causes relatively less harm in LDCs. It causes relatively less harm (as indexed by willingness to pay) because wages and productivity are lower, because people are more likely to die of other causes before they can be harmed by pollution, and because there are other pressing needs on which individuals will spend their money first. But is willingness to pay the right way to measure economic benefits? Are economic costs and benefits a good guide to what is harmful and beneficial? Given the current unequal distribution of wealth, people in rich countries will pay much more to prevent crippling injuries than people in poor countries will. But the moral significance of crippling injuries arguably should not depend on whether the victim lives in a wealthy country. A middle-​income person in a rich country might pay five hundred dollars to add a larger flat-​screen TV to her den, while a poor person in an LDC might decline to –​or be unable to –​pay the local equivalent of five hundred dollars to fix his broken leg. Willingness to

30

Ethics in Welfare Economics

pay is a dubious guide to what ought to be done. To regard costs and prices, as determined by actual markets, as markers of moral significance is a dubious moral presumption. For a telling example, consider the following:

Market Value vs. Moral Value In some cases market value seems to differ drastically from what most people would regard as moral value. The drug eflornithine is a highly effective “miracle” cure for sleeping sickness, which infects tens of thousands of people in Southern Africa. Until 1999 eflornithine was produced by a U.S. subsidiary of the Aventis company, but when eflornithine proved ineffective against cancer (its intended target), Aventis stopped making the drug, which is expensive to synthesize, and gave the production license to the World Health Organization. People who contract sleeping sickness are very poor, and there was no money in producing a cure for the disease. Only in early 2001, when stocks of the drug were almost exhausted, was the WHO able to find drug companies to manufacture it –​and then only because the companies hoped to make profits from marketing eflornithine in developed countries as a cream that removes facial hair. It is currently sold under the brand name Vaniqa. By market criteria, eflornithine is of value because it removes facial hair, but as a cure for sleeping sickness it is not worth producing. In fact, economists do not typically identify the value of a human life with the loss of expected earnings or with the differing amounts different people would pay to prevent a death. But why not? If economic costs and benefits are a good guide to what is harmful and beneficial, then they should be a good guide to the allocation of risks of death and injury; and if they are not a good guide to the allocation of risks of death and injury, why should one believe that they provide an acceptable way to measure benefits or harms? The World Bank memorandum is an amalgam of economic analysis and moral philosophy. Or perhaps it would be less misleading to say that its economic analysis already contains ethical assumptions. We do not pretend to have proven that ethics is mixed up in all of economics and cannot be separated out; indeed, in the Appendix we shall consider proposals for isolating a pure “value-​free” science of economics. Our point here is illustrative. Economics of the kind exemplified by the World Bank memorandum is shot through with controversial ethical premises. Table  2.1 illustrates this point by listing a number of questions that must be faced by anyone

2.6  School Vouchers

31

concerned with the evaluation of policies and by presenting some of the alternatives to the answers mainstream economists have generally given.

2.6  School Vouchers Our second example involves a controversy concerning primary and secondary education: the proper role of competition and market incentives in the provision of schooling. Economists distinguish between the case for public funding of some product or service and the case for direct public provision of that item by government workers (Schultze, 1977). The argument for public funding is usually cast in terms of some public purpose in expanding production and consumption of the service above what the market will provide. What mechanism should be used to accomplish this purpose is a separate question, and economists often argue that markets have advantages in terms of efficiency and freedom compared to direct provision. Even though markets often possess these admirable features, it may not always be best to rely on them, both because markets sometimes fail (as in the case of pollution) and because, even when they succeed, markets might undermine other values. However, economists are on the whole inclined to believe that markets are neutral with respect to other values and to take market failures as the only justification for government interference. Economists typically regard the market as serving given preferences –​as a means to furthering people’s ends, whatever they may be. This is a substantial moral assumption, which is more reasonable in some circumstances than others, as the case of public elementary and secondary education illustrates. Can it be right to rely on preferences to evaluate a practice such as schooling, which so strongly shapes preferences? The question of public provision versus public funding has arisen in a particularly sharp way in the context of elementary and secondary schooling in the United States. More than a half-​century ago, Milton Friedman (1962, ch. 6) argued that the social interest in ensuring that all young people receive an education does not imply that the government should operate schools. He proposed that government funding for schooling should be provided instead in the form of vouchers. Vouchers are certificates that families can use to purchase schooling and that schools then redeem for cash. There are many possible variants on a voucher system. Vouchers could be a fixed amount per student, or they could vary with family income or the differing costs of educating students with special needs. Some voucher proposals would permit parents to supplement the value of the voucher, while others

32

Ethics in Welfare Economics

would restrict families’ spending on schools to the amount of the vouchers. Some voucher proposals would permit schools to select among their applicants, while others would require them to accept all applicants on a first-​come first-​serve basis. The main attraction of vouchers, as an alternative to government schools, is that they would induce schools to compete for students and thereby (it is argued) provide education of better quality at lower cost. To consider the case for or against vouchers, it helps to step back and ask why governments need to encourage elementary and secondary schooling, beyond the amounts and types families would choose themselves. What makes education different from “private” goods such as automobiles? We distinguish three reasons supporting government funding or provision of primary and secondary education: paternalistic, distributive, and political. Paternalist Justification The most obvious reason for public support of education rests on the fact that primary education and secondary education are crucial to people’s life prospects. They determine the range of career opportunities as well as cultural and personal horizons available to people. Furthermore, the recipients –​children –​are not capable of making informed choices themselves. Most parents take an interest in the future well-​being of their children, and at least in developed societies, that interest gives parents a strong reason to see to it that their children become educated. But the interests of parents and children are not the same, and the children’s prospects should not depend entirely on the responsibility, interest, and capacities of their parents. Government has a (paternalistic) responsibility to children. Distributive Justification By itself, the paternalist reason for public concern might be satisfied by requiring parents to school their children, leaving it to the market to supply the demand occasioned by that requirement. But this line of thought leads immediately to a distributive concern, for parents may lack the resources to provide adequate schooling for their children. Allied to this may be a broader distributive issue: a concern for distributive fairness or equality of opportunity demands an adequate level of education for children from disadvantaged families. Political Justification Both the paternalistic argument and the distributive arguments focus on the influence of education on individual outcomes and well-​being. A third strand of argument is directly political. In his case for public funding of education, Friedman stresses the fact that how well children are educated has consequences for society at large. A society’s schools benefit many others in addition to those whom they educate by helping to create thoughtful and well-​informed citizens, helping to protect and to enrich the society’s culture, and providing the skills and knowledge for a

2.6  School Vouchers

33

productive economy. As we saw in the discussion of the World Bank memorandum, there is a case for interference with the market when activities have costs and benefits for others that do not involve costs or benefits to the individual. So there is a narrowly economic case for a role for government in ensuring that children are educated. How do these three types of reason  –​paternalistic, distributive, and political –​bear on the question of public funding vs. public provision of schooling? A paternalist who worries that parents would not spend enough on schooling might also worry that parents will make poor choices among schooling options. Governments may be better able to make schooling choices than parents, who lack expert information and are not always well motivated. But in most societies, relations between parents and children are governed by a strong presumption in favor of parental authority. This presumption rests on the reasonable grounds that in matters that deeply concern parents, such as education, parents will try hard to figure out what is best. Parents know their children better than governments do and are typically more actively concerned with their welfare. Although paternalistic concerns justify strong regulatory limits on parental discretion, they do not provide a strong case for government provision. The fact that children cannot decide for themselves justifies limits to parental authority, but it does not undermine the case for decentralized market provision, provided schools supplied on the market are regulated. Distributive concerns lie behind other arguments for government provision. Many have argued that a voucher system would be unfair to poor families. If parents were allowed to supplement vouchers, then vouchers would tend to reinforce inequalities even if vouchers were equal for children of all family backgrounds. Although the problem is attenuated as the size of the standard voucher grows, permitting parents to supplement vouchers would tend to undermine political support for generous vouchers, since these are less advantageous to the relatively affluent, who have the greatest political influence. As unfair as this variant of a voucher system would likely be, it might be objected that it would be less unfair than the current system of funding public education in the United States, which, thanks to its reliance on local taxes (which generate more revenue in wealthier areas), devotes a substantially larger share of public funds to educating children of those who are well off than to educating poorer children. Since a voucher system would probably be funded at a state rather than a local level, and since it would be difficult to justify providing a larger voucher to richer children than to poorer, even a uniform and supplementable voucher system might turn out to be fairer than the current system in the United States. (But more equal funding within states does not address the problem of

34

Ethics in Welfare Economics

inequalities in funding across states, which are substantial.) Voucher policies that direct resources toward those who are in more need of help would be fairer still. Whether any such systems are politically feasible is a separate question. The most challenging and the philosophically most interesting difficulties facing voucher schemes relate to the last reason for public support of schools: their influences on the development of moral and civic “character” among students. As we observed previously, judgments about the kinds of beliefs and values to promote among schoolchildren are political and moral judgments. At the most abstract level, there might be a reasonable degree of agreement on the values to be promoted –​tolerance, mutual respect, support for democratic participation –​but schools will be pressed to make decisions that are more fine-​grained and controversial. Should Muslim girls be allowed to wear headscarves in publicly funded schools in France? To what degree should American history emphasize the contributions of members of minority groups? At what age is sex education appropriate, and what should children be told? Questions like these, easily multiplied, are important for two reasons: first, they concern the content and not simply the amount of education children receive, and, second, the content questions are ones in which the society has a central interest. Thus, while it is plausible, on grounds of both liberty and efficiency, to say that a child’s parents are generally as well positioned as anyone to judge which educational investments are best for that child, there is no similar case that the parent should have a uniquely important role in deciding what kind of civic preparation his or her own child should receive. The question is not only what is best for the individual child but also what is best for the community. Even though a liberal state aims ultimately to serve the interests of individuals, it can only do so if it can engender support for its institutions. It seems unlikely that this concern can be met by regulation alone. Private religious institutions whose staff are unenthusiastic about secular states are unlikely to provide appropriate civic education. Of course, what some people call political education, others call political indoctrination. To the degree that the civic purposes of schooling are taken seriously, the question of who gets to decide what will be taught becomes crucial. Economists who regard markets as neutral with respect to values other than efficiency and freedom would argue that these questions, as important as they may be, are independent of the question of whether there should be government provision as well as government funding of

2.6  School Vouchers

35

education. Either way, one faces the same questions concerning the content of civic education and who is responsible for determining that content. In a voucher system could not government require whatever civic education is judged appropriate? But a voucher system is not neutral. It is bound to diminish the influence of local school districts, because parents will now have wider choices, and it is bound to diminish the influence of teachers, who as employees of private firms will have fewer common interests, less employment security, and consequently less responsibility and power. To what degree can elements of decentralization and competition be introduced into public schools compatibly with appropriate democratic control? It is too simple to characterize the options as simply traditional public schools and complete market provision through vouchers. Governments can hire private contractors to provide schooling without offering families choice, and a variety of devices, including charter schools, can provide elements of decentralization and parental choice in systems that do not feature profit-​seeking firms providing schooling. However, it would be naive to opt for thoroughgoing market provision through vouchers and then assume that government regulation will protect the character of democratic education. Because the development of students’ moral and civic values is important, decentralization and competition raise serious concerns. The view of the market as a morally neutral mechanism is not innocuous (Gutmann 1987 and 2000). The shift from a system of government provision to a voucher system is a momentous social choice, because establishing it might significantly change the character of individual, social, and political life. Individuals do not interact only via the market, and political life does not consist only of voting. Recent discussion of vouchers, while often citing civic preparation as a basic rationale for schooling, focuses primarily on the effects vouchers would have on the efficiency of schools, with some secondary concern about the distribution of education. In empirical studies, including those based on social experiments, the assumption is sometimes maintained that scores on standardized tests are an adequate proxy for the “output” of schooling. As a consequence, questions about the content of schooling are ignored or reduced to the influence of differences in content on test scores. Although questions of efficiency and distribution are surely important, reducing questions about vouchers (or about different systems of school governance more generally) to these matters begs important questions. In addition to the issues concerning civic education, one needs to consider

36

Ethics in Welfare Economics

the effects of dismantling publicly provided education on social solidarity and on the existence of a common culture. Classrooms are already often segregated by class, race, and religion, and voucher systems have the potential to exaggerate that segregation. Although many parents want their children to get to know children from different backgrounds, other parents would prefer that their children associate only with others of the “right” class, religion, or race. Overt discrimination by schools can be prohibited, but self-​selection is difficult to control, and the possibility that vouchers will help to bring about a future in which children grow up wearing the blinders of narrow communities is not far-​fetched. Such a system might be splendidly efficient, and possibly even fair, while at the same time causing the dissolution of community or the undermining of national identity. Society does not consist solely of market exchange within political constraints, and markets are not morally neutral apart from their contributions to freedom and efficiency. Awareness of the ways in which markets compete with other institutions and of their bearings on values such as justice, solidarity, and community can help economists contribute more thoughtfully to debates about schooling. We take up these issues in more detail in Chapter  6 when we discuss the question of whether or not the market has moral limits and whether everything should be treated as a market good.

2.7 Conclusions Our purposes in discussing the World Bank memorandum and school vouchers are, to repeat, only illustrative. Our goal is not to pronounce on whether school vouchers are a good idea or on whether pollution ought to be exported from wealthy countries to LDCs, and it is certainly not our intention to maintain that welfare economics is misguided. Indeed, it is the unabashed clear-​headedness of the applications of familiar economic ideas that makes the examples powerful. The point of the examples is to show that mainstream normative economics rests on substantial and controversial moral premises and to make our claim plausible: that knowing ethics can contribute to doing economics.

Suggestions for Further Reading General overviews of the relations between economics and ethics can be found in Allen Buchanan (1985), Hamlin (1986), Sen (1987a), and Gaus (2007).

Questions for Study and Discussion

37

There has not been much scholarly discussion of the World Bank memorandum, although it raises general issues concerning cost–​benefit analysis, which have been extensively discussed; see Chapter 9. For an excellent discussion of the notion of exploitation, see Wertheimer (1996). Satz (2010) also uses the memorandum as an example of the need to enrich the evaluative basis of economics. The classic statement of the case for vouchers as a mechanism for the public finance of education appears in Milton Friedman (1962, ch. 6). For a very different statement of the case for vouchers, see Gintis (1995). An influential empirical study is John Chubb and Terry Moe (1990). Criticisms of vouchers from the standpoint of democratic theory are advanced in Amy Gutmann (1987 and 2000) and from the standpoint of justice in Harry Brighouse (2000). A  comprehensive treatment of charter schools can be accessed at www.publicagenda.org/​media/​charter-​schools-​in-perspective-​ pdf (2014).

Questions for Study and Discussion 1. What do you think of the three arguments in the World Bank memorandum for exporting pollution to poor countries? 2. As summarized in Table  2.1, this chapter points out that there are alternatives to the many aspects of the ethical framework deployed in the World Bank memorandum and that welfare economics rests on choices concerning what to appraise and how to appraise it. Consider how welfare economics would change if just one of the many choices were made differently. 3. The argument in Section 2.4 is logically valid. That means that if you accept all of the premises, you must, on pain of inconsistency, accept the conclusion. If you do not accept the conclusion, then you need to deny at least one of the premises. What do you think of the premises of that argument? 4. An argument very much like the one in the World Bank memorandum can be made for eliminating some of the limits many societies place on markets, such as laws prohibiting prostitution, the consumption of recreational drugs, the buying and selling of organs for transplant, or the buying and selling of rearing rights to newborns. In each case, economists can argue that allowing these transactions will make both parties to the transaction better off. Try to construct a rigorous argument for this conclusion analogous to the argument in Section 2.4.

38

Ethics in Welfare Economics

If you do not accept the conclusion, which of the premises would you criticize? 5. Which of the many choices that constitute standard welfare economics seems to you especially defensible or especially indefensible? 6. Which of the objections in Section 2.5 to exporting pollution to less developed countries seems to you the strongest? Which seems the weakest? 7. One might argue that the actual effects of globalization have not been all that different from what the World Bank memorandum espouses. Lots of polluting industries have moved to LDCs, and they have brought a good deal of economic development with them. Is it so obvious that the conclusions draw in the World Bank memorandum are repugnant? 8. What is your view of changing from state-​provided public education to a voucher system? 9. How, if at all, has the discussion of vouchers in this chapter changed your views concerning their advantages and disadvantages? 10. The results of voucher systems depend heavily on details of their design. In the light of the issues raised in this chapter, how would you design a voucher system? 11. The results of state-​provided education systems also depend on details of their design. What are the leading improvements you would recommend for the system of state-​provided education that now exists in the country you live in? 12. As this chapter emphasizes, economists evaluate policies mainly in terms of efficiency. On the basis of the case studies in this chapter, what would you say are the main advantages and disadvantages of assessing policies mainly in terms of their efficiency?

THREE

Ethics in Positive Economics: Two Examples

Although some economists have envisioned the possibility of a value-​ neutral normative economics that merely investigates the consequences of policies for the satisfaction of preferences, most would be willing to concede that moral judgments do play a role in normative economics. For that reason, they might also concede that when evaluating policies economists should pay attention to matters of ethics. But when it comes to positive economics  –​the attempt to predict and explain economic outcomes and processes –​few economists see any role for ethics. Although it is possible (even in economics) to investigate features of the world without evaluating them, it is often hard to do so. When matters of economics bear strongly and immediately on people’s interests, those interests and the moral considerations that are relevant to them are likely to influence the questions economists ask and the answers they defend. Because of this, it is important to understand the values that are at stake. Familiarity with moral philosophy can help here. Although values will be most obvious in debates over practical policy questions –​consider debates over energy policy, prescription drug benefits, agricultural policy, or tax policy –​moral considerations can have a much broader influence. We shall show this influence in two examples. Both concern matters of fact about economies that bear on policy questions. The first example, which concerns whether there is any involuntary unemployment, is likely to seem odd to readers who are not familiar with economics. How, one might ask, could anybody doubt that some people are unable to get a job? We shall shortly see. Whether workers are unemployed voluntarily or involuntarily seems to bear on the question of what sort of unemployment compensation they should be paid, but its normative implications and presuppositions go deeper. The second example, which involves Paul Samuelson’s overlapping generations model, might appear to have no bearing on policy; but as we 39

40

Ethics in Positive Economics: Two Examples

shall argue, its normative implications have in fact driven what is purported to be a purely theoretical debate in positive economics.

3.1  Is Unemployment Involuntary? “Macroeconomics”  –​which treats the institutions and processes governing the economy as a whole –​was born during the worldwide economic depression of the 1930s with the publication of John Maynard Keynes’s General Theory of Employment, Interest, and Money. Among the central goals of macroeconomics have been explaining and figuring out how to remedy involuntary unemployment. During the last few decades, influential economists have argued that this inquiry rests on a mistake, because there is in reality no such thing as involuntary unemployment, except perhaps as a brief transitory occurrence. In everyday language, a person is “voluntarily” unemployed if she quits her job and chooses not to accept another one. A person is involuntarily unemployed if she loses her job and is unsuccessful in finding another one. People who are voluntarily unemployed are “at leisure”:  they choose not to be employed because they prefer some other set of activities to working under current labor-​market conditions. Indeed, in the way economic statistics are kept, many of those who are not working are not counted as unemployed at all –​to be considered unemployed you must be actively looking for work. How you define who counts as unemployed has significant consequences, especially in countries where only the officially unemployed can qualify for benefits under social insurance programs. The disputes between those who seek to explain involuntary unemployment and those who deny that there is any involve matters of economic theory and methodology. Yet the moral interpretation of the idea of “voluntariness” is significant as well, as we argue in more detail in this section. Whether economists deny that there is involuntary unemployment is linked to their views on policy. Most Keynesian economists, who believe involuntary unemployment is an important phenomenon, are also inclined to believe that activist monetary and fiscal policies are essential in managing modern economies. Thus, Shapiro and Stiglitz write, “To us, involuntary unemployment is a real and important phenomenon with grave social consequences that needs to be explained and understood” (1985, p. 1217). Those who are skeptical of involuntary unemployment tend to believe, in contrast, that individual activities coordinated by markets work out well and that government action is more likely to cause than to cure economic difficulties. Thus Robert Lucas suggests that unemployment is like any other economic choice:

3.1  Is Unemployment Involuntary?

41

S

w

w0

D w′

D′

N*

N0

N

Figure 3.1.  Supply and demand in the labor market.

The unemployed worker at any time can always find some job at once, and a firm can always fill a vacancy instantaneously. That neither typically does so by choice is not difficult to understand given the quality of the jobs and the employees which are easiest to find. Thus there is an involuntary element in all unemployment, in the sense that no one chooses bad luck over good; there is also a voluntary element in all unemployment, in the sense that however miserable one’s current work options, one can always choose to accept them. (1978, p. 354)

Lucas emphasizes the voluntary aspect: “By treating all unemployment as voluntary, this work has led to the examination of alternative arrangements which firms and employees might choose to adopt for dealing with fluctuations in product demand” (p. 356). The analytical and methodological issues raised by this dispute are complicated, and we can only sketch them here. The logic of standard microeconomic theory states that if demand for some product or service drops, producing an excess supply of it, then the price will promptly drop until the excess supply is eliminated. So, viewing labor as such a service, there should be no excess supply of labor or unemployment. Keynesian analyses claim that the self-​correcting mechanisms that smoothly adjust for fluctuations in supply and demand in commodity markets fail in the market for labor in the economy as a whole. Simplifying drastically, one might depict the market for labor in a given industry as in Figure 3.1. The supply curve, S, shows how much labor workers are willing to supply at each wage. We are oversimplifying and equating this with how many workers want to work. The demand curves D and

42

Ethics in Positive Economics: Two Examples

D´ show how much labor (how many workers) firms want to employ at each wage. The demand curves slope downward, because firms are willing to hire more workers when wages are lower. The supply curve slopes upward because workers are willing to work more at higher wages. Here D shows the demand for labor before the shocks that gave rise to the Great Depression, such as the Wall Street stock market crash of 1929. At the “equilibrium” wage w0, the number of workers whom firms want to employ and the number of workers who want to work both equal N0, and there is neither unemployment nor excess demand. There is then a shock that lowers demand to D´. Though there will (in reality) be confusion and transitory unemployment, the theory abstracts from this and shows that equilibrium will be restored, though at a lower wage of w´ and a lessened employment N´. Fewer workers are employed, but none are unemployed. Everyone who wants work at the new wage w´ can find work. Keynes argued that in the transition from an individual market to the system as a whole –​from micro-​to macroeconomics –​this story of how markets equate supply and demand breaks down. Unlike in an individual market, a drop in the wages of labor across the economy lowers workers’ purchasing power. With lower wages and fewer workers employed, the population will not be able to purchase as many goods and services, reducing aggregate demand. Exactly where and how the analogue of the microeconomic argument breaks down and whether there exist any mechanisms within the market system that will correct the imbalance in aggregate supply and demand are questions that have shaped the evolution of macroeconomic theory for three-​quarters of a century. But, at least in Keynes’s view, unemployment cannot be attributed to the unwillingness of workers to accept lower wages, and it falls to government to provide the proper cure, which is to stimulate higher aggregate demand via monetary or fiscal policy. It is difficult to incorporate involuntary unemployment into a coherent choice-​theoretical account of the economy (DeVroey 2004). As a simple illustration of this challenge, suppose one redraws the supply curve to represent the reluctance of workers to accept employment at wages below w0 (see Figure 3.2). This has been proposed as one way of characterizing the assumption of “rigid” or “sticky” wages, which on some interpretations is a key assumption of Keynesian economics. The unwillingness of workers to accept employment at wages below w0 is depicted by drawing the supply curve as horizontal to the left of the old labor market equilibrium (at w0 and N0). Because workers refuse to accept jobs for wages less than w0, employment drops all the way to N*, where the supply curve intersects the new demand curve. Orthodox theorists do not believe that a supply curve such

3.1  Is Unemployment Involuntary?

43

w D

D′ S w0

N*

N0

N

Figure 3.2. Sticky wages.

as the one shown would continue to reflect workers’ preferences, believing rather that workers’ expectations that they would find work at w0 would collapse. Although some theorists attempted to explain involuntary unemployment by means of a horizontal supply curve, it seems instead (as De Vroey argues) that, along the horizontal segment of S, workers must be indifferent between working and not working. They prefer not working to accepting jobs at wages below w0, and they prefer accepting jobs at wages above w0 to not working. The intersection of D´ and S is a point of market equilibrium, and at w0, workers do not prefer to supply more or less labor than N*. Suppose, for example, that w0 = $8.00 per hour. What the diagram says is that any individual worker willing to work for $7.99 per hour can readily find a job. The decisions of such workers to decline an offer of a penny less per hour would appear to be voluntary. Since workers are doing what they prefer to do and since nothing prevents them from accepting lower wages, Figure 3.2 does not depict a situation involving involuntary unemployment. (This pertains to the labor force as a whole. At a more detailed level, it is possible that some workers coercively prevent others from accepting jobs at lower wages.) Indeed, it is arguable that in Figure 3.2 there is no unemployment at all –​ at least as unemployment is conventionally measured. It is not the case that the N0 –​ N* workers who were employed before the shift in demand and who are not now employed would strictly prefer working at the going wage

44

Ethics in Positive Economics: Two Examples

to not working. Since they do not prefer working to not working at w0, they have no reason to seek work at w0 and thus, on one plausible construal of what it means to be unemployed, they are not literally unemployed at all! They are instead (perhaps unhappily) at leisure. Moreover, within the framework of economic theory, the situation depicted in Figure 3.2 is “efficient.” Workers and firms are behaving in accordance with their preferences. Nothing in the graph captures the fact that something has gone wrong; the enormous human cost of the Great Depression has simply disappeared. The attempt to comprehend involuntary unemployment by means of Figure 3.2 is neither plausible in itself, because workers’ expectations will adjust, nor adequate to the theoretical problem, because workers are not indifferent between working and remaining unemployed. In order for there to be involuntary unemployment, it must be the case that workers would strictly prefer to be working at the market wage but are unable to find a job –​or, in other words, that the jobs and wages available to those who have lost their jobs are substantially worse than those of workers with similar qualifications who are still employed. This means that actual employment at the market wage must be off the supply curve. In Figure 3.1, for example, the wage must somehow get “stuck” above w, even though there are workers who are willing to work for less. But what prevents the wage from falling? Except as a bit of transitory confusion, from which economic analysis abstracts, when demand shifts as it does in Figure 3.1 wages must fall unless (as in Figure 3.2) workers are indifferent between working and not working at the old wage w0. Simply to assert that wages are “sticky” does not explain why. Apart from totalitarian control over labor, how is it possible to be off the labor supply curve? However “obvious” it may be that a good deal of unemployment is involuntary, when one steps into the theoretical world of economics and focuses on graphs like the two shown here, it is hard to understand how involuntary unemployment is possible. This is the point that Lucas and other members of the “rational expectations” school focus on. Their view –​which is shared by many economists of other schools, too  –​is that an adequate analysis of economic activity must be built up from proper “microfoundations.” Theorists must assume that every actor is acting rationally on the basis of the information available to him and that every actor invests rationally in acquiring information. Unemployment results from rational search. In searching for a new job, an unemployed worker is viewed as continually weighing the quality of the best option she has found against the likelihood of finding a better option by searching longer. Periods of high unemployment occur when more people

3.1  Is Unemployment Involuntary?

45

are searching and the average search is longer. Lucas hastens to counter any “blaming the victim” accusation that might follow such a claim: “treating unemployment as a voluntary response to an unwelcome situation does not commit oneself to normative nonsense like blaming depressions on lazy workers” (1978, p. 356). Whether some version of the rational expectations program can ultimately succeed in reducing macroeconomics to microeconomics is a hotly disputed question. But even if a worker’s unemployment can be shown to be rational from her point of view, does that make it “voluntary”? Both Keynesians and “rational expectations” theorists agree that, in depressions, the alternatives facing workers become sharply worse from the worker’s own perspective. And both agree that the project of explaining in detail how labor markets work in these circumstances is a difficult problem, not fully resolved. In his denial that he is blaming the victim, Lucas shows his awareness that the word “voluntary” carries significant normative freight: if unemployment is voluntary, then it is the worker’s own responsibility and of lesser social concern. Similarly, in speaking of the “grave social consequences” of involuntary unemployment, Shapiro and Stiglitz ally themselves with a critical and activist social agenda. In our view, a better understanding of ethical theory has a good deal to contribute here. Both in everyday and in philosophical discourse, whether a choice is voluntary depends both on the quality of the alternatives and on the circumstances bringing about the occasion for choice. The choice posed by the robber who offers “your money or your life” is involuntary in the sense that the alternative to parting with one’s purse is unacceptable because the choice is brought about in an illegitimate way: the robber does not have a right to your purse, but you do. The fact that handing over one’s money to the robber is a rational choice –​and that it comes about through the agent’s own actions and intentions  –​does not make it voluntary in the relevant moral sense. People can be subordinated to others even though they still have choices. Robert Solow captures this understanding of voluntariness as follows: You can take the definition that is actually there in the General Theory, and loosen it up so that it says that someone is involuntarily unemployed if he or she would be prepared to work at a job that he or she knows how to do, at the going wage, and cannot find employment. That is good enough for me. You will notice that by this definition a person can be involuntar[il]y unemployed and employed at the same time. If you take an inferior job you can still be an involuntar[il]y unemployed skilled worker. (Snowdon and Vane 1999, p. 287)

46

Ethics in Positive Economics: Two Examples

Solow here denies that all rational choices are voluntary. To claim that people are involuntarily unemployed is not to claim that they cannot find any job at all. Unemployment should be considered involuntary when people do not have any job opportunities that are reasonably good by the standards of their society. What if the available jobs do not pay enough to support a minimally decent life? Or what if the only jobs available to the unemployed that pay well are illegal or carry stigma, such as smuggling or prostitution? Since whether the alternatives to unemployment are “reasonably good” is clearly a normative question –​and not just for the workers –​the economist’s concept of involuntary unemployment is itself, in part, a normative concept. (The normative criterion in Solow’s definition is that a “reasonably good” alternative for one who is unemployed is a job–​wage package that others with the same qualifications are actually enjoying.) Those who claim that  –​even in a depression  –​all unemployment is voluntary might then be understood in one of three ways. The first way is that they have made a mistake that could have been prevented by a greater familiarity with ethics: they are confusing “voluntary” with “rational.” From the perspective of unemployed workers, the available jobs, such as selling apples or shining shoes for a wage that cannot sustain them or their families, are not acceptable; their refusal to accept these jobs is not voluntary. In the same way that it can be rational for Ann to give her purse to a robber, that does not mean that her choice to do so was not objectionably constrained. Even a slave makes choices. But no one would deny that her servitude was involuntary. A second possibility is that the dispute between Keynesians, on the one hand, and Lucas and others on the other is fundamentally normative: they disagree about whether the alternatives to continuing to search are reasonably good, at least as temporary expedients. While Keynesians might look at working at extremely low-​wage jobs as unacceptable alternatives that should give rise to positive government action, those inclined to laissez faire might value the choice to accept such jobs as falling within a sphere of individual freedom. The third possibility –​closely related to the second alternative –​is that those who regard unemployment as voluntary understand spells of unemployment, even occasionally severe and widespread spells, as part of the normal workings of a market economy and (viewed perhaps in a lifetime perspective) understand that having to endure such spells is a reasonably good alternative to arrangements that would eliminate the risk of unemployment. To the extent that unemployed workers in effect have faced a hypothetical choice between secure jobs with no stretches of unemployment and

3.1  Is Unemployment Involuntary?

47

less secure jobs that are otherwise more attractive, they have had reasonably good alternatives and may be deemed to be unemployed voluntarily. The key point to note in this debate is that voluntariness is at root a moral notion, closely related to judgments of freedom and personal responsibility. If unemployment is involuntary, then it threatens economic freedom as well as wasting resources and lowering welfare. Emphasizing that workers are not responsible for their fate invites a search for governmental remedies. Robert Skidelsky’s (1992) biography of Keynes describes the importance Keynes attached to these emphases as he developed his arguments through the 1930s. If economists regard unemployment as voluntary –​a matter of searching longer rather than accepting an available job –​then it is easier for them to see it as part of the normal workings of a market economy, as relatively unthreatening to basic social values, and as less obviously a government’s responsibility to fix. On this view, if workers did not have unreasonably high expectations, then unemployment would disappear. The roots of these contrasting ethical perspectives lie in differences in attitudes toward market outcomes and processes as well as government intervention, which align plausibly (though not necessarily) with differences in theoretical view. Thus Keynes himself and Keynesian followers are inclined to see widespread unemployment as revealing a basic flaw in both market economies and microeconomic theory, and they have argued that a different theory is needed for the economy as a whole than for its parts. On the other hand, those who see unemployment as voluntary generally believe that market economies perform well without government interference and that macroeconomic theory can be built on axioms depicting every individual as choosing optimally under conditions of rational expectations. These disagreements led to heated controversies concerning how to respond to the severe economic downturn beginning in 2008, with Keynesians arguing for more government stimulus, and the descendants of rational expectations theorists arguing for government austerity to restore business confidence and lessen market distortions. The links we have noted between judgments about the voluntariness of unemployment, about the adequacy of microeconomic equilibrium theory to explain macroeconomic phenomena, and about the role government should play in reducing unemployment are not logically tight. One can believe that unemployment is both involuntary and the source of grave social consequences yet still think there is nothing government can or should do about it. In contrast, one can believe that unemployment is involuntary (in the moral sense just defined) and still think that a suitably developed theory of rational expectations equilibrium is the best approach

48

Ethics in Positive Economics: Two Examples

to explaining it. Still, the links among these views are not entirely coincidental, and the energy that goes into defending alternative positive theories of unemployment derives in large part from the normative views that frequently accompany particular theoretical positions. Though disagreements would obviously remain, a better understanding of ethics (in particular, of the nature of voluntary choice) would help to refine them.

3.2  Overlapping Generations In his celebrated paper “An Exact Consumption-​Loan Model of Interest with or without the Social Contrivance of Money” (1958), Paul Samuelson addresses the following problem: suppose individuals want to save for their old age, when they cannot produce anything, and suppose nothing lasts from one period to the next. All people can do if they do not want to starve is to strike a bargain during their working years so that those who are younger will support them when they are retired. In a world of endlessly overlapping generations of workers and retirees, what will the pattern of interest rates be? This appears to be a factual rather than an evaluative question, and obviously it is a very abstract issue. To isolate the effect of this desire to provide for one’s old age from other factors that influence the rate of interest, Samuelson formulates an extremely simplified model in which everyone lives exactly three periods. In each of the first two periods of their lives, people produce one unit of a completely perishable consumption good. In the third period people produce nothing and can survive only if those who, in the first two periods of their lives, support them. To simplify further, Samuelson considers first an absolutely unchanging economy and then one growing eternally with an unchanging rate of growth and an unchanging rate of interest. In these models a “biological” rate of interest equal to the rate of growth clears all markets and is optimal (in the sense that there is no way to make some people better off without making others worse off). For example, in the case of no growth at all, individuals can “save” one-​third of their output in each of the two periods in which they are productive and then maintain the same level of consumption (2/​3 unit) during retirement. But is this outcome socially feasible? Remember that output is perishable and cannot literally be saved. The one-​third of their output that young and mature workers do not consume constitutes the 2/​3 unit of the good that retired workers consume. What is here called “savings” is in fact a transfer to retired workers. Why should current workers transfer anything to the retirees? After all, the retirees never gave anything to these workers, and

3.2  Overlapping Generations

49

the retirees are in no position to retaliate if the young or mature workers refuse to “save.” It seems as if the biological interest-​rate solution (which in this simple example with a zero rate of growth involves a zero interest rate) is impossible. All that remains –​if people are to avoid starving when retired –​is for the workers in the second period of their lives to transfer goods to the younger workers in return for repayment in the next period when this period’s mature workers will be retired. Hence it seems that, on a competitive market, the only bargains people can strike will result in lavish living during the first period of life and penury during the remaining parts. At this point Samuelson’s essay takes a surprising turn. He simply drops the question he began with concerning the effects on the rate of interest of the desire to save for one’s retirement and focuses instead on (1) a theoretical demonstration that people would be better off with the “biological” rate of interest (equal to the rate of growth) than with the competitive solution and (2) an explanation of how such a biological interest-​rate regime could be achieved. He argues that an arrangement like the one just described, whereby everybody consumes 2/​3 unit in each period, could be achieved by means of a social contract in which individuals collectively agree to an obligation to support retirees or people agree to accept “fiat money” from retirees –​that is, money that has value only because it is common knowledge that individuals are willing to accept it as payment for goods. (Samuelson argues that fiat money is, in fact, very much like a social contract.) Samuelson treats his conclusions as important normatively, but an outsider might wonder why anybody should care about the efficiency of a biological interest rate in a fictitious economic model. The immediate responses to Samuelson’s essay by Abba Lerner (1959a, b) and William Meckling (1960a, b) are surprising. They purport to find analytical mistakes in Samuelson’s work (erroneously –​see Samuelson 1959 and 1960), but in fact they are driven by normative concerns. From a perspective to the “left” of Samuelson, Lerner denies the ethical desirability of a rate of interest equal to the rate of growth. From a “utilitarian” perspective (which we will explain later) Lerner argues that maximum average consumption (and welfare) in any given period is achieved if and only if consumption is equal across the three generations. In the case of a zero interest rate, Lerner’s utilitarian plan corresponds to the distribution that would result from a biological interest rate, but the two distributions differ for any positive rate of growth. Lerner argues that Samuelson is mistaken in regarding provision for the retired as savings because, in Samuelson’s model, nothing can be

50

Ethics in Positive Economics: Two Examples

invested or held over from one period to the next. Social security should instead be regarded as a program in which workers transfer income to retirees. Normative policy disagreements concerning whether social security should be conceived of as a savings program or as a transfer program are central to the controversy between Samuelson and Lerner. Meckling, writing from a more conservative perspective, denies that workers will voluntarily provide goods to retirees, because the retirees never previously gave those workers anything. So the only way to save for retirement involves mature workers making loans to young workers, who will then consume more than they produce their first year and make repayment the next year. Why? Meckling’s main argument against the possibility of a social contract is that it is against the self-​interest of young workers to transfer income to retirees and hence that they will not voluntarily do so: “the zero-​interest-​ rate equilibrium can prevail only if the sheriff is retained on a permanent basis” (1960b, p. 84). But the sheriff is just as necessary if there is no social contract, for whether in repayment or as required by the social contract, it is never in anybody’s self-​interest (in this model) to transfer anything to retirees. Contracts to do so require enforcement. Furthermore, if an individual believes that later generations will respect the social contract or accept fiat money only if she does so, then it is in her interest to honor the social contract and accept the fiat money. So it is doubtful whether a very tough sheriff is needed after all. What is really bothering Meckling, as well as Lerner and Samuelson, is a normative rather than an analytic issue. In this hypothetical instance, perfect competition does not lead to the most preferred result. Samuelson’s model suggests that government is needed to back fiat money or to enforce a social contract. As we shall see later, perfect competition is widely regarded as an ideal both by economists who support government intervention to remedy market failures and by economists who believe that government ought not to interfere in economic affairs. That is why Samuelson is so interested in demonstrating that the competitive solution without a social contract or fiat money is suboptimal and why Meckling is so disturbed by the demonstration. Lerner, in contrast, doubts that markets are the right way to make provision for retirement, and he seeks to shift the question to one of establishing welfare-​maximizing levels of transfer payments to retirees. What appears on the surface to be a dispute about a theoretical question in positive economics actually turns on the evaluative presuppositions that economists have toward competitive markets and their outcomes.

Questions for Study and Discussion

51

3.3 Conclusions That the arguments economists make about important matters include an ethical component should be neither a surprise nor a cause for regret. As our second example shows, this ethical dimension is present even in a controversy concerning abstract overlapping generations models, as well as in contexts (such as unemployment or school vouchers) fraught with policy implications. So long as economists deal with socially consequential subjects, some (but certainly not all) entanglements of fact and value will be unavoidable and proper. The best way to cope with these entanglements is to make them explicit and to address both the economic and the ethical issues. If there is any implicit criticism of the economists in the examples discussed in this chapter, it is that they have not always achieved the same level of analytical clarity and critical self-​awareness in the parts of their work that connect with ethics as they have sought in the rest of their work. Our aim in the following chapters is to clarify different ethical approaches and to help readers to think about them rigorously.

Suggestions for Further Reading Our discussion of involuntary unemployment was heavily influenced by De Vroey (2004). The debate about whether unemployment is involuntary is connected to the larger question of whether the transactions that compose a market economy are voluntary. A classic defense of that view is in Friedman (1962, ch. 2). An economist’s argument that market systems have coercive elements can be found in Okun (1975). On voluntariness in market transactions see also Peter (2004). A good overview of the literature is Yellen (1984). Paul Samuelson’s (1958) overlapping generations model has been much discussed. For an overview of its influences and of the permutations the basic model has undergone, see Hausman (1992, ch. 7).

Questions for Study and Discussion 1. Economists who maintain that unemployment is voluntary usually favor less government intervention in the labor market, whether in the form of unemployment insurance, minimum wages, or government hiring. They also typically hold an unfavorable view of unions.

52

Ethics in Positive Economics: Two Examples

Economists who maintain that unemployment is involuntary typically favor more government involvement and assistance and are more favorably disposed toward unions. Why? 2. If the only jobs available to the unemployed do not pay enough for them to feed themselves or find a place to live, then it seems implausible to maintain that they are voluntarily unemployed. But if the unemployed have the option of taking jobs that do not imply starvation or homelessness, should we regard their unemployment as voluntary? What about someone who has trained to be an intensive-​care nurse and who turns down a lesser job as a cardiac-​care nurse? How should economists distinguish between voluntary and involuntary unemployment? 3. While they were working, those who are currently retired accumulated claims to goods and services, whether in the form of cash or other assets, that could be exchanged for the goods and services retirees consume. Meckling asks the question “Why should those who are now working give any of these goods and services to those who are no longer working?” How would you answer him? Does your answer explain why materially self-​interested individuals supply goods to retirees? 4. Should we think of the social security program as a mandatory form of individual savings for retirement, or should we think of it as a social program whereby the young support the old? In other words, is it a savings program or a transfer program? Or is this a false dichotomy? 5. There are deep disagreements concerning the slow current rate of growth of most advanced economies, with some economists arguing that the problem lies with too much government debt, too much taxation, too much government regulation, and a consequent lack of confidence on the part of businesses, who are reluctant to invest. Others argue that the problem lies with inadequate overall demand and that more government borrowing (which can be done at low cost, owing to low interest rates) can improve infrastructure and lead to increased investment and growth by increasing demand. These are disagreements about factual matters. What, if any, influence do you think that ethical considerations have on these disagreements? 6. Do you think that it is possible to formulate an economic theory that, unlike the examples in this chapter, shows no influence of normative concerns? Why? Can you think of an example?

PART ONE

RATIONALITY, MORALITY, AND MARKETS

When we say that it is morally right to relieve famine victims, we are expressing our approval of famine relief and at least suggesting that, all things being equal, people ought to go to the aid of those in danger of starvation. Morality is normative as opposed to merely descriptive. So is rationality  –​although, unlike morality, only conditionally so:  actions are rational or irrational, given the agent’s beliefs and desires. When we say that it is rational for individuals to have medical insurance, we are expressing approval of having it. Similarly, to characterize a choice as irrational is usually to condemn it, and not simply to describe it. Not only are morality and rationality alike in these ways, but “rational” is often used (as in the previous example) as a synonym for “prudent,” and prudence is a morally admirable character trait or moral virtue. At the same time, morality and rationality are not the same. How are morality and rationality related? Is it always rational to be moral? These general philosophical questions are critical to understanding the relations between ethics and economics, because economics is built around a theory of rationality. We explore the relations among economics and rationality, norms, and morality in Chapter 5. To prepare for this exploration, Chapter 4 presents the standard theory of rationality and its extensions to circumstances of risk and uncertainty; it also considers some of the objections to which the theory has been subject. Chapter 4 is the only technical chapter in the first three parts of this book. Chapter 5 examines the role of the standard theory of rationality in both positive and normative economics, and it explains how the theory ties positive and normative economics together and makes both appear to be more plausible than they otherwise would be. Chapter 5 also deals with questions raised by the effects of people’s morals on their behavior. How (if at all) can economic theory take account of these effects? What is the relation between 53

54

Rationality, Morality, and Markets

morality and preference? Can one (or should one) develop a conception of economic rationality that permits a distinction between acting morally and maximizing utility? If the two diverge, is it then sometimes irrational to act morally or immoral to act rationally? Chapter 6 shifts focus from questions concerning the bearing of morality on individual behavior to the bearing of morality on markets themselves. Some things should not be for sale. Which? For what reasons? How do markets affect the social relations among individuals and their morals? Is participation in competitive markets corrupting, or does it encourage honesty and trustworthiness? How large a role should markets play in social relations?

FOUR

Rationality and Utility Theory

Mainstream economics portrays agents as choosing rationally. Many generalizations in economics concerning how people do in fact choose are also claims about how they ought rationally to choose. If people are, to some reasonable degree of approximation, rational, then a theory of rational preference, belief, and choice will constitute an approximate theory of people’s actual preferences, beliefs, and choices. The fact that economics is built around a theory of rationality distinguishes economics from the physical sciences, where quarks and polymers do not choose at all and whose theories have no comparable normative dimension. In common usage, someone is rational if and only if her motives, beliefs, and choices are appropriately sensitive to reasons. Spelling out this vague idea is a huge philosophical challenge. “Folk psychology” –​the everyday theory of human rationality  –​takes actions to derive from constraints, beliefs, and a wide array of motivational factors such as urges, emotions, habits, and desires. So, for example, when one rainy Friday night a hungry student named Ellen takes a frozen pizza out of the refrigerator, unwraps it, puts it in her stove, and turns knobs on the stove, we folk psychologists explain Ellen’s action by the constraints (pizzas do not get hot by themselves), Ellen’s beliefs  –​including especially her beliefs that turning the knobs will cause the stove to heat the pizza –​and her desire to eat hot pizza. This sort of explanation is familiar but not very satisfactory. Ellen might also like to eat her pizza frozen, or she might also have a desire to reheat some leftover meatloaf. Or she might rather skip dinner and keep studying decision theory. What explains her action is not merely wanting to eat hot pizza (plus possessing the requisite beliefs) but also wanting to do this as much as or more than she wants to carry out any of her feasible alternatives. The theory of rational choice that dominates economics is less ambitious than folk psychology: it aims only to specify formal conditions 55

56

Rationality and Utility Theory

on preferences, beliefs, and choices. Economists tighten up the folk-​ psychological account of action by replacing the noncomparative notion of a “desire” with the comparative notion of a “preference.” They can then explain the interaction between Ellen and her stove in terms of physical constraints, Ellen’s beliefs about the outcomes of the alternative actions she can undertake, and her overall ranking of those outcomes. Economists would explain the actual pizza warming by showing that Ellen ranks it at least as highly as any feasible alternative. In taking “preferences” to incorporate everything relevant to choice –​as total comparative evaluations –​one has arrived at the view of choice that economists have generally adopted. Economists regard choices as actions that arise from constraints, preferences, and expectations (or beliefs). The concept of “preference” is central to economic theory. Economists typically take preferences to be predetermined or “given” facts about individuals and, for their purposes, not in need of explanation or subject to substantive appraisal. Economic analyses begin with an individual’s preferences, whatever they may be. Folk psychology and the refinements by economists that we have just sketched are “positive” theories –​that is, they are theories that explain and predict people’s choices. But it is plausible to regard various conditions on preference, belief, and choice and on the relations among them not merely as generalizations about people, but as formal conditions on rationality. Choice is rational when it is determined by a rational set of beliefs and preferences. The rationality of sets of preferences and beliefs is defined within “utility theory.” Although this is not a technical book, our discussion cannot proceed without some discussion of the technicalities of utility theory.

4.1  Certainty and Ordinal Utility Theory Suppose one is concerned with the choices, preferences, and beliefs of an agent named “Q.” In circumstances of complete certainty, Q chooses rationally if her preferences are rational (which we define in this section) and if there is nothing available that Q prefers to what she chooses. Except in the case of ties, one can simplify: Q is rational if her preferences are rational and she chooses what she most prefers from among the available alternatives. Q’s preferences are rational if they are transitive and complete. Q’s preferences are transitive if and only if, for all objects of choice or evaluation (“alternatives”) x, y, and z, if Q prefers x to y and y to z then Q prefers x to z. Similarly, if Q is indifferent between x and y and between y and z, then she is indifferent between x and z. It is plausible to require that rational

4.1  Certainty and Ordinal Utility Theory

57

preferences be transitive. Here is a famous argument for the conclusion that intransitive preferences are irrational. Suppose: 1. Q has intransitive preferences (she prefers x to y and y to z, but she prefers z to x); 2. Q already has some of y; and 3. Q is willing to pay a penny to trade what she has for something she prefers. Q will then pay a penny to trade y to get x, another penny to trade x to get z, another penny to trade z to get y, another penny to trade y to get x, and so on, until Q realizes that her preferences leave her vulnerable to manipulation. Hence Q will not cling to intransitive preferences. On the assumption that rational choices should further one’s ends, intransitive choices are not rational. This argument showing the irrationality of intransitive preferences is called “the money pump argument.” It is persuasive, though some would question it (Schick 1986). Q’s preferences are complete if, for all alternatives x and y, either Q prefers x to y or Q prefers y to x or Q is indifferent between x and y. If Q’s preferences are complete, then Q is never unable to rank x and y. Completeness is an idealization, because sometimes there are alternatives among which people have no preference. For example, Q may not know enough about some alternatives to know which she prefers. Such problems will not arise in conditions of certainty, and completeness may be a reasonable simplification when there are no uncertainties. But it is questionable whether completeness is a condition of rationality (Levi 1980) or a precondition for the capacity to choose rationally. Note the difference between being indifferent and being unable to rank. Indifference is ranking equally, not an inability to rank. In folk psychology, Q’s preference ranking is a subjective state of Q:  it reflects Q’s evaluation of states of affairs with respect to everything relevant to choice. Q’s preferences and Q’s beliefs then jointly explain her choices. Many economists would disagree. In accordance with the theory of “revealed preference,” they instead attempt to define preference in terms of choice. Provided that consistency requirements are met, choosing x when one might have had y at a lower cost reveals a preference for x over y, whether or not it is an actual subjective state of Q. The theory of revealed preference is subject to serious criticisms. If Q mistakenly believes that y is not available, then Q may choose x while actually preferring y. Moreover, if preferences are defined by choices, then Q’s preferences are restricted to those things Q can choose. According to

58

Rationality and Utility Theory

revealed preference theory, individuals may prefer to vote for a candidate promising a balanced budget, but the theory of revealed preference makes it impossible for individuals to prefer that the budget be balanced, because whether there is a budget deficit is not something that a voter can choose. Economists also need to be able to predict how people’s choices will change when they acquire new information, and in order to do this, they need a less behaviorist notion of preference. When corporations report large unexpected quarterly losses, economists predict that the price of their stock will go down. Why? To say that people now reveal a preference to sell this stock simply shifts the question. To explain why people want to sell the stock, economists must cite the preference for profitable investments. But this is not a revealed preference. Its connection to action is via belief (about which investments are profitable) rather than direct. Completeness and transitivity together establish a weak ordering of any finite set of alternatives. What this means is that one can put the alternatives in a long list, with the alternatives Q prefers in higher rows and the alternatives among which Q is indifferent in the same row. Completeness implies that every alternative goes into some row, and transitivity ensures that alternatives go into only one row. Given such a list, one can assign numbers to options such that preferred options get higher numbers and indifferent options get the same number. If Q prefers a particular portable computer to $500 and $500 to a sandwich, then economists can represent Q’s preferences by assigning numbers to these three alternatives. For example: computer $500 sandwich

12 7 2

Any other three numbers in the same order  –​such as (100, 97, 4)  or (2,000,000, 92, 91) –​would do just as well. This may seem bizarre, because people are accustomed to attaching significance to the magnitudes of numbers, not just what they indicate about ordering. But the sole purpose of the numbers here is to show the ranking. Any such assignment of numbers is an ordinal utility function. “Utility” here does not refer to usefulness or pleasure. A utility function is only a way of representing a preference ranking –​that is, a ranking of alternatives with respect to everything relevant to choice. The ordinal representation theorem states that if an individual’s preferences are complete, are transitive, and satisfy a further “continuity” requirement, then they may be represented by a continuous real-​valued

4.1  Certainty and Ordinal Utility Theory

59

ordinal utility function (Debreu 1959, pp. 54–​9). (Continuity is a technical condition that we do not discuss.) The theory of rationality in circumstances of certainty can thus be restated: Agents are rational if and only if their preferences can be represented by ordinal utility functions and their choices maximize utility. We intentionally avoided saying that they act “in order to maximize utility,” because in contemporary economic theory, utility is merely an indicator of preference: maximizing utility is just doing what one most prefers to do. Although the utility language was inherited from the utilitarians –​some of whom thought of utility as a pleasurable sensation with a certain intensity, duration, purity, or “propinquity” (Bentham 1789, ch. 4) –​there is no such implication in contemporary economic theory. To speak of individuals as “aiming to maximize” utility or as “seeking more” utility may misleadingly suggest that utility is an object of choice, some good thing that people want in addition to healthy children, longer lives, or good food. But the theory of rational choice is a formal theory that says nothing substantive about what people want or ought to want. Because the theory of rational choice is indifferent to the content of what people want, it does not imply that agents always act out of self-​interest. Jill may prefer that some benefit go to Jack without automatically concluding that it is better for her that Jack benefit. When economic theories treat individuals as self-​interested, as they often do, they add substantive claims concerning what people want to the formal theory of rationality. Utility theory lays down formal conditions that choices and preferences ought to satisfy. It is not a positive theory because it says nothing about the extent to which people are rational, and it is not merely a model or definition because rationality is itself a normative notion. To define rational preference and choice is ipso facto to say how one ought rationally to prefer and to choose. Consider the following example. The psychologists Amos Tversky and Daniel Kahneman (1981) presented a hypothetical story to two groups of experimental subjects: “Imagine that the U.S. is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed.” The first group, consisting of 152 subjects, was told that the exact scientific estimates of the consequences of alternative responses were as shown in Table 4.1, and 72 percent preferred program A. The second group, consisting of 155 subjects, was instead given the information shown in Table 4.2, and 78 percent preferred program D. But the choices differ only in how they are described. It is irrational to prefer A to B and at the same time prefer D to C. In saying this we are not

60

Rationality and Utility Theory Table 4.1.  The unusual Asian disease

Program A Program B

200 people saved 600 people saved with probability 1/​3 No one saved with probability 2/​3

Table 4.2.  The unusual Asian disease (again) Program C Program D

400 people will die No one will die with probability 1/​3 600 people will die with probability 2/​3

denying that many people have these preferences, for that is what the experiment shows. Nor are we simply stipulating that the term “rational” should not be used for preferences like these. In saying that these preferences are irrational, we are criticizing them. In this example the subjects’ decision making is influenced by “framing” factors that ought to be irrelevant. Since utility theory says nothing about what individuals prefer, it has a much wider scope than most economic theories, which depict agents as seeking more commodities or larger net returns. Thus, utility theory is employed by psychologists, statisticians, philosophers, and sociologists as well as by economists. To hold that choices are rational if determined by preferences and that preferences are rational if they are complete and transitive does not seem to demand much. Yet many theorists maintain that it is still too demanding, and there are interesting weaker concepts of rationality (see McClennen 1990, ch. 2).

4.2  Expected Utility Theory In a good deal of traditional economics, theorists abstract from the problems that arise when agents are not certain about the consequences of their actions. Frank Knight (1921) suggested that one speak of circumstances of risk when all the alternative outcomes and their probabilities are known. Someone playing roulette with a wheel that is known to be unbiased is in a circumstance involving risk. In circumstances of uncertainty, on the other hand, some of the probabilities and perhaps even some of the possible outcomes are unknown. Global warming involves uncertainty, because no one knows all the possible outcomes or their probabilities. In speaking of known or unknown probabilities, we are talking about so-​called objective probabilities, such as relative frequencies. For example, the probability of

4.2  Expected Utility Theory

61

a coin landing heads can be regarded as the limit of the ratio of heads to total flips as the number of flips grows ever larger. As we shall see shortly, there is also a subjective notion of probability as “degree of belief ” that is important in the theory of decision. In situations of risk and uncertainty –​ when actions do not lead with certainty to any particular outcome –​actions may be regarded as lotteries with outcomes as prizes. Actions undertaken in circumstances of certainty can also be regarded as lotteries, though of a boring sort, since they always pay off the same prize. To extend the theory of rational choice to circumstances involving risk and uncertainty, standard theories assert that preferences among lotteries are complete, transitive, and continuous. As noted before, completeness is not very plausible as a condition of rationality in conditions of uncertainty. Economists have postulated some technical conditions relating complex lotteries (which have lotteries as prizes) to simple lotteries. Finally, to extend the theory of rationality to conditions of uncertainty, one needs an additional condition, which is known in the literature as the “independence condition” or the “sure-​thing principle.” Suppose that Q, like most people, prefers (other things being equal) more money to less and that Q has a choice between two gambles concerning the flip of a coin. Gamble 1 pays off $10 if the coin comes up heads and nothing if the coin comes up tails. Gamble 2 pays off $9 if the coin comes up heads and nothing if it comes up tails. It seems to be a condition of rationality that Q prefer the first gamble with the higher prize, though there is no violation of completeness, transitivity, or continuity if Q prefers the gamble with the lower prize. A preference for the second gamble violates the independence or sure-​thing principle. That principle says that if two lotteries differ only in one prize (which may itself be a lottery), Q’s preferences between the lotteries should match her preferences between the prizes. It is called the sure-​thing principle because Q is at least no worse off no matter what the outcome of the lottery, and with one of the outcomes Q is better off. Axioms such as completeness, continuity, and independence are more controversial in the case of uncertainty than in the case of risk. Indeed one might argue that talk of probabilities and of lotteries is obscure when probabilities and even the range of possible outcomes are not known. Some disagree: so-​called Bayesians are more comfortable applying these axioms in circumstances of uncertainty than are non-​Bayesians. For Bayesians believe that the probabilities that are relevant to decision making are subjective degrees of belief. People may have degrees of belief about events for which relative frequencies cannot be defined, such as the outcome of a particular World Cup soccer match. A particular World

62

Rationality and Utility Theory

Cup is played only once, and the probability that a particular team will win is not the limit of the frequency of its victories in repeated World Cups. Bayesians interpret a claim such as “The odds that Brazil will win again are 3 to 1” as expressing instead a subjective degree of belief. An argument resembling the money-​pump argument shows that degrees of belief should obey the formal axioms of completeness, transitivity, and continuity. Whether one should apply expected utility theory in conditions of uncertainty remains controversial. The characterization of actions as lotteries is controversial because it is arguable that what matters when an agent is choosing among alternative actions are the probabilities that the actions will cause various outcomes and not simply the conditional probabilities of the outcomes given the actions. Robert Nozick (1969) provided a memorable hypothetical problem (“Newcomb’s problem”) to illustrate what is at issue here. Suppose there are two boxes on a table in front of you. One is transparent and contains $1,000. The other is opaque. You can either take the opaque box and leave the transparent box, or you can take both boxes. You know that the opaque box contains $1,000,000 if and only if an almost infallible predictor predicted that you would choose only the opaque box. The monetary expectation of choosing one box is thus close to $1,000,000, while the monetary expectation of choosing both boxes is close to $1,000. If one thinks of the actions as two lotteries, it seems as if choosing one box is the better bet. Yet, unless causation goes backward in time, whether the $1,000,000 is in the opaque box is already determined; it does not depend causally on whether you choose one or two boxes. Whether the million dollars is in the opaque box or not, you always come away with $1,000 more by choosing both boxes. So there appears to be a compelling argument for taking both boxes. Complications such as these have led some decision theorists to argue for what they call “causal decision theory,” wherein our decisions among actions depend on a consideration of their causal consequences rather than on conditional probabilities (for further discussion see Gibbard and Harper 1978; Skyrms 1980; Lewis 1981; Eells 1982). If Q’s preferences satisfy all three axioms, then they can be represented by a cardinal utility function where the magnitudes of the utility numbers, rather than just their order, are significant. The (cardinal) representation theorem says that if an agent’s preferences are complete, transitive, and continuous and if they also satisfy the independence condition (and other technical conditions concerning compound lotteries), then those preferences may be represented by a utility function that has two special properties (see

4.2  Expected Utility Theory

63

Ramsey 1926; von Neumann and Morgenstern 1947; Herstein and Milnor 1953; Savage 1972; Harsanyi 1977b, ch. 3). First, this utility function possesses the expected utility property:  the (expected) utility of any lottery is equal to the utilities of its outcomes multiplied by their probabilities. Suppose Q is offered a gamble over the roll of a fair die. Q will win $60 if the die comes up 1 or will lose $12 if the die comes up with any other number showing. The expected monetary value of the gamble is 1/​6 times $60 minus 5/​6 times $12, or $0. The expected utility of the gamble to Q is 1/​6 times the utility to Q of winning $60 plus 5/​6 times the utility to Q of losing $12. Depending on how rapidly Q’s utility increases as a function of money, the expected utility of the gamble may be positive, negative, or zero. If Q’s preferences satisfy the axioms, then there is a way of representing Q’s preferences so that the expected utility of a lottery is the sum of the utilities of the prizes weighted by the probabilities of winning them. One sense in which the expected utilities are “cardinal,” rather than merely ordinal, is that the precise utilities of gambles can be calculated from probabilities and utilities of the outcomes. Second, the cardinal representation theorem states that if Q’s preferences satisfy the axioms, then an expected utility function (a utility function with the expected utility property) that represents them is “unique up to a positive affine or linear transformation.” To say that a function U is “unique up to a positive affine or linear transformation” is to say that any U´ = aU + b, where a is a positive real number and b is any real number, will do just as well as U itself. If a function U represents an agent’s preferences and has the expected utility property, then so will all positive affine transformations of U. Conversely, no utility function that is not a positive affine transformation of U will provide a cardinal representation of the agent’s preferences. There will be other utility functions that will also represent Q’s preferences, but they will not be cardinal representations with the expected utility property. To understand what this means, it helps to compare the measurement of expected utility to the measurement of temperature. Alternative temperature scales, such as the Fahrenheit and the Celsius scales, are positive affine transformations of one another (TF = (9/​5)TC + 32). Because of this, measurements of temperature on the Fahrenheit scale convey the same information as temperature measurements on the Celsius scale. A claim such as “temperature differences between night and day are larger in deserts than in wet climates” is true regardless of which scale one uses because the two scales are positive affine transformations of one another. For the case of ordinal utilities, in contrast, differences between utilities are entirely arbitrary, since any numbers in the same order represent the

64

Rationality and Utility Theory

preferences just as well. The fact that the difference between the ordinal utilities U(x) and U(y) is larger than the difference between U(c) and U(d) tells one nothing about the agent’s preferences. The relative magnitudes of the numbers are just an accident of the particular utility function chosen. But matters are different with cardinal utilities. The cardinal representation theorem establishes that comparisons of utility differences (like comparisons of temperature differences) are not arbitrary; they do not depend on what scale is chosen. If U(x) − U(y) > U(c) − U(d) and if U´ is a positive affine transformation of U, then U´ (x) − U´ (y) > U´ (c) − U´ (d).1 Although expected utilities are cardinal and thus comparisons of utility differences are in this way “objective,” nothing said so far permits one to make comparisons between the utilities of different individuals. If Q’s cardinal utility is 100 and M’s cardinal utility is 89, one cannot draw any conclusion about whose preferences are better satisfied. Since any positive linear transformation of these numbers will represent the preferences of the individuals just as well, their relative magnitude reflects only an arbitrary choice of one particular utility function for each individual. As in the case of utility theory concerning choice under certainty, economists relate choice to preference by asserting that individuals choose whatever available option is highest in their preference ranking –​or, in the case of ties, one of the options tied for first place. Unlike ordinal utility theory, which is a theory of preference and choice only, expected utility theory also has implications concerning belief. The axioms of expected utility theory imply that people’s degrees of belief concerning uncertain outcomes satisfy the axioms of the probability calculus. Expected utility theory is thus not only a theory of rational preference and choice; it is also a partial theory of rational belief.

4.3  Questions about Utility Theory Expected utility theory demands more than does utility theory under certainty and is much more controversial. Psychologists and economists have found ways of testing whether people’s preferences satisfy the conditions of expected utility theory, and they have found many examples where preferences appear not to satisfy them. If people do not behave the way that 1 Proof: We have U ′(x) − U ′(y) = aU(x) + b − aU(y) − b = a(U(x) − U(y)). Similarly, U ′(c) − U ′(d) = a(U(c) − U(d)). So U ′(x) − U ′(y) U ′(c) − U ′(d) if and only if a(U(x) − U(y)) > a(U(c) − U(d)). And since a is positive, this will be true if and only if U(x) − U(y) > U(c) − U(d).

4.3  Questions about Utility Theory

65

Table 4.3.  Allais’ problem Payoffs Problems

Choices

I

A B C D

II

Red (1)

$1 million $0 $1 million $0

White (89)

$1 million $1 million $0 $0

Blue (10)

$1 million $5 million $1 million $5 million

expected utility theory says they ought to, then it may be the case that people are irrational. People may break the rules, even if they are good rules. But if choice behavior that violates the axioms of expected utility theory seems reasonable, then one may question whether expected utility theory is a correct theory of rationality. Empirical anomalies, along with general critiques of expected utility (Allais 1952; Ellsberg 1961; Sen 1985b; Levi 1986; McClennen 1990), have stimulated the formulation of alternative theories of rationality under uncertainty. Testing a normative theory of rationality is less straightforward than testing a positive theory of actual choice, and controversy can easily persist. We cannot survey all of the controversies concerning expected utility theory and its competitors here, but we will give one famous instance that is of some importance to moral philosophy as well as to decision theory. This example provides some sense of the character of these controversies. In the early 1950s, Maurice Allais formulated the problem shown in Table 4.3 (Allais and Hagen 1979). A ball is drawn from an urn containing 1 red ball, 89 white balls, and 10 blue balls. So the probabilities of drawing a ball of any color are known. Depending on the color and the choice of A or B in problem I or of C or D in problem II, the player receives one of the payoffs listed in the figure. For example, if Q faces problem I and chooses A, then she gets $1 million for sure. If Q chooses B, then she has a 1  percent chance of getting nothing (if a red ball is drawn), an 89% chance of winning $1 million, and a 10% chance of winning $5 million. Many people are inclined in problem I to prefer option A to option B and in problem II to prefer option D to option C. Even the prominent Bayesian statistician Leonard Savage at first expressed these preferences (1972, p. 103). But these preferences violate the independence principle, for the only difference between the pair (A, B) and the pair (C, D) is in the magnitude of the payoff if a white ball is drawn,

Rationality and Utility Theory

66

which should be irrelevant.2 Yet many people still prefer A in problem I and D in problem II. Intuitively the reason is that in choosing B, one is giving up a million dollars for sure, while in choosing D one is not. This reason is in conflict with the independence condition, which says that the prize if a white ball is chosen is irrelevant to the choice. It is possible to reconcile a preference for A and D with the independence condition by arguing that the outcomes have not been properly described. The outcome of choosing B if a red ball is chosen is not $0, but $0 plus severe regrets. (“I could kill myself. I gave up a million dollars for sure!”) Although some decision theorists (see Broome 1991b, ch. 5) have defended the independence condition in this way, others have worried that such redefining of the outcomes makes expected utility theory empty, since apparent disconfirmations can always be explained away by redescribing the outcomes. For example, in Tversky and Kahneman’s Asian disease example, one could say that there is no irrationality in people’s choices, because 200 of 600 people living is a different outcome than 400 of 600 people dying. Broome’s response is to argue for substantive principles of rationality that specify whether it is rationally permissible to distinguish between the outcomes when a red ball is drawn when one chooses B (and thus might have had $1  million for sure) and when a red ball is drawn when one has chosen D. (Such principles would deny that it is rationally permissible to distinguish 200 of 600 dying from 400 of 600 surviving.) Most decision theorists and economists do not follow Broome here, and they regard preferences for A in problem I and at the same time for D in problem II as being inconsistent with the independence condition and irrational. Here’s why: Let L be the lottery that pays off $5 million with probability 10/​11 and nothing with probably 1/​11. One can then recast Figure 4.3 as follows. 2

Payoffs Problems I II

Choices A B C D

White (89) $1 million $1 million

Red or Blue (11) $1 million L $1 million L

The independence condition says that A is preferred to B if and only if $1 million is preferred to L, and it also says that C is preferred to D if and only if $1 million is preferred to L. So A is preferred to B if and only if C is preferred to D. For those who prefer some algebra, let V be the utility of $5 million, U the utility of $1 million, and 0 the utility of $0. Then EU(A) = U, EU(B) = 0.89U + 0.1V, EU(C) = 0.11U, and EU(D) = 0.1V, where EU denotes expected utility. If A is preferred to B, then EU(A) > EU(B). Hence U > 0.89U + 0.1V, or 0.11U > 0.1V. So if A is preferred to B, then C must be preferred to D.

4.3  Questions about Utility Theory

67

If one takes the choices of A in problem I and D in problem II to be inconsistent with the independence condition, then are these choices evidence against the independence condition or rather evidence of human irrationality? It is hard to see how to make conclusive arguments in favor of one of these alternatives, but we believe that either the payoffs need to be redescribed or there is a case to be made against the independence condition. There is no other evidence of irrationality among those preferring A and D, and plausible alternative theories of rationality have been defended that rationalize choosing A and D. Apparent anomalies such as Allais’ example are relevant to moral philosophy as well as to economics, because expected utility theory is involved in arguments for moral conclusions –​such as Fleming’s (1952), Harsanyi’s (1955), and Vickrey’s (1945) purported proofs of utilitarianism. As we will discuss in Chapter 7, utilitarianism is the view that actions and policies are morally permissible if and only if they lead to no less total “utility” than any alternative. There are also nonutilitarian ethical systems that are built around expected utility theory, such as David Gauthier’s (1986) or Ken Binmore’s (1994, 1998) bargaining theories of justice. These issues concerning expected utility theory are not merely theoretical, because alternative theories may yield different recommendations about what is the “rational” thing to do. For example, in the case of global warming (which obviously involves uncertainty rather than merely risk), should policy makers rely on their best guess as to the future consequences and maximize expected utility? Or should they, as Isaac Levi (1980) urges, recognize their ignorance of relevant probabilities and use some other principle of choice? In addition to the controversies concerning whether expected utility theory is too narrow or too demanding, some moral philosophers have maintained that there are such things as substantively irrational preferences quite apart from any questions about consistency. This is a classical view found in Aristotle and Plato. More recent defenders include Thomas Nagel (1970) and John McDowell (1978). Derek Parfit, for example, holds that it is “irrational to desire something that is in no respect worth desiring” (1984, p. 123). He gives the example of a person who suffers from “future-​Tuesday indifference.” This individual prefers a greater pain on a future Tuesday over a lesser pain on a Wednesday, simply because he does not care what happens on future Tuesdays. Parfit remarks:  “The fact that the agony will be on a Tuesday is no reason for preferring it. Preferring the worse of two pains for no reason, is irrational” (1984, p. 124). Although the decision theory most economists accept makes no substantive claims about what preferences are rational or irrational, the

68

Rationality and Utility Theory

economic theories they accept typically identify rationality with material self-​interest and hence implicitly judge that preferences for things that are advantageous for oneself are rational and that other preferences are irrational (Walsh 1994). John Broome argues that expected utility theory itself must rely upon judgments that some preferences are intrinsically irrational. Consider, for example, an instance of intransitivity: Q prefers x to y, y to z, and z to x. If Q wanted to make trouble for the decision theorist, she could argue that the apparent intransitivity results from misdescribing the objects of her preferences. She might say that “z when compared to y” is a different alternative from “z when compared to x.” If one calls the first z1 and the second z2, then Q may have a fully transitive preference ordering such as z2, x, y, z1. Such a reply to the charge of having intransitive preferences can always be given, and so transitivity turns out to be empty as a condition of rationality, unless one is willing to argue that it is irrational to prefer z2 (z when compared to x) over z1 (z when compared to y). Expected utility theory has no “bite” unless one is committed to the irrationality of such preferences. Like Broome, David Lewis and Susan Hurley have argued in this way that expected utility theory presupposes substantive principles of rationality. Although it is good to know something about the ways in which expected utility theory may be too strong or too weak, what is most important for our purposes is to understand clearly what the different varieties of utility theory have to say about rationality. In simple terms, utility theory sees rational choice as choice that is determined by rational preference and belief, and it sees rational preference as demanding that individuals be able to rank all options. It is remarkable that such a bare-​bones formal theory of rationality has given rise to so much controversy and (as we shall demonstrate) that it has such striking ethical implications.

Suggestions for Further Reading The basic model of rationality and its extension in expected utility theory have been widely discussed. Some of the sources that we have found particularly accessible and helpful are Sen (1971), Harsanyi (1977b, ch. 3), Loomes and Sugden (1982), Machina (1987), Gärdenfors and Sahlin (1988), McClennen (1990), and Hollis and Sugden (1993). For philosophical discussion on the plausibility of restricting questions about rationality to questions about consistency and about the rationality of means, see Hume (1738, 1748), Nagel (1970), McDowell (1978), Parfit

Questions for Study and Discussion

69

(1984, pt. 2), Gauthier (1986, ch. 2), Griffin (1986, chs. 3, 4), Hurley (1989, ch. 5), Broome (1991b, ch. 5), Hampton (1998), Sen (2002). For empirical investigations of possible irrationality, see Lichtenstein and Slovic (1971); Kahneman, Slovic, and Tversky (1982); Camerer et al. 2004; Lowenstein (2008); and Thaler (2015).

Questions for Study and Discussion 1. How is it possible for utility theory to function both as a normative theory concerning how one ought rationally to choose and as a theory purporting to describe how people actually choose? Why does standard economics include a theory of rationality? 2. In what ways is folk psychology inadequate, and how does the theory of rational choice that economists accept attempt to address these inadequacies? 3. Why is it misleading or incorrect to say that an individual aims to get as much utility as possible? What is utility, as it is understood by economists? 4. The standard theory of rationality says that individuals are motivated exclusively by their own preferences. Does this imply that they are self-​interested? 5. What does the theory of revealed preference say? Why does it imply that it is impossible for voters to prefer that Trump win the election against Clinton? 6. Do you think that you would give different answers to the question concerning how to treat the unusual Asian disease depending on how the questions were phrased? Why? If you are inclined to give different answers, depending on how the question is phrased, how would you go about deciding which of your conflicting answers was correct? 7. If you were faced with Newcomb’s problem, would you take one box or two boxes? Why? What can unrealistic hypothetical cases such as this one teach us? 8. If you were faced with Allais’ problems, would you choose in accordance with what expected utility theory predicts and normatively requires? Why? 9. Do you think that future-​Tuesday indifference is irrational? Why? Can a purely formal theory of rationality succeed, or is it necessary to include some claims concerning which preferences or beliefs are substantively irrational?

FIVE

Rationality and Morality in Positive Economics

Positive economics is concerned with the explanation and prediction of economic phenomena, while normative economics is concerned with their evaluation. It is hardly surprising that moral considerations are important in normative economics. But as the examples of involuntary unemployment and overlapping-​generations models illustrate, moral considerations are important within positive economics, too. Rationality is itself a normative notion concerning how people ought to choose, prefer, believe, or reason and, as this chapter explains, it plays a very large role in positive economics. As we shall see later, rationality is also crucial to normative economics, and its use in both domains forges a deep connection between positive and normative economics.

5.1  Rationality and Positive Economics Economists regard people’s preferences as rational if they are complete and transitive, and economists regard choices as rational if they conform to people’s preferences. If one adds to this theory of rationality the generalization that consumers are in fact rational and the assumption that they prefer more commodities to fewer, then one has the central principles of the theory of consumer choice. Similarly, the traditional theory of the firm maintains that firms or entrepreneurs are rational and that they combine inputs so as to maximize the difference between revenues and costs. Although the theory of consumer choice and the theory of the firm make additional claims concerning the motives of consumers and entrepreneurs, they both take the theory of rational choice to be the foundation of their accounts of actual choices. Positive economics on both the consumer and the producer side can be formulated without using the word “rational.” Rather than first defining 70

5.1  Rationality and Positive Economics

71

“rational” and then stating that individuals are rational, economists can assert that the preferences of consumers are complete and transitive and that individuals choose whatever affordable bundle of commodities they most prefer. On the production side, economists can simply assert that firms aim to maximize profits. But the identification of what is actual with what is rational remains. This identification reflects the fact that economics simultaneously provides a theory of the causes and consequences of people’s economic choices as well as a theory of the reasons for them. In explaining actions, people sometimes cite only the agent’s beliefs or only the agent’s desires, since the other is obvious from the context. For example, when asked why he robbed banks, Willie Sutton replied, “Because that’s where the money is.” His explanation makes us laugh because, unlike most of us, Sutton thinks that the desire to steal is too obvious to need mentioning. Nevertheless, folk psychology, the pattern of explaining action in terms of the agent’s beliefs and desires (given constraints), is ubiquitous. Economic explanations of choices in terms of utility maximizing fit this pattern. As we explained in Chapter 4, economists typically talk in terms of “preferences” or “utility” rather than “desires,” and they often do not mention the agent’s beliefs because they assume that the agent knows the relevant facts. But these are minor complications. In the main, explanations in mainstream economics of the choice behavior of individuals conform to the pattern of everyday folk psychology. Economists are more interested in the market consequences of individual choices than in the choices themselves, but individual choices are the causal intermediaries that connect “shocks” such as crop failures or a new tax to their consequences for prices or quantities exchanged. The relations between microeconomics and folk psychology are worth emphasizing because folk psychology offers an account of the reasons for actions. Indeed, a number of philosophers, especially in the 1950s, contrasted folk psychological explanations in terms of reasons to causal explanations of behavior in terms of conditioning or brain chemistry. In their view, explanations for actions in terms of beliefs and desires were exclusively reason giving and not causal. If they were right, then there would be a radical difference between explanations in microeconomics, which (like folk-​psychological explanations) cite reasons, and explanations in the natural sciences, which cite physical causes. We disagree:  the reasons that explain actions must also cause the actions that they explain. Donald Davidson (1963) supplied the crucial insight here. Consider, for example, a real-​estate agent who attends church regularly. One reason is that the agent believes in God and wants

72

Rationality and Morality in Positive Economics

to express devotion to God. Another reason is that the agent meets potential clients at church. Although some might find the first reason more admirable than the second, both reasons justify going to church. Both reasons make going to church a rational thing for the real-​estate agent to do. Davidson points out, however, that even though good reasons justify actions, they do not automatically explain them. One needs to know which reasons actually “led” the agent to go to church. Both reasons might explain the church going, or the agent might attend because of piety and only incidentally profit in other ways from attending. On the other hand, the agent may be a hypocrite and attend church only as a way of attracting clients, or it might be that the explanation is actually something else. Perhaps the agent loves to listen to the church organist. Davidson’s point is that, in order to explain an action, one needs to know which reasons are causally responsible for it. Reasons, unlike physical causes, are subject to normative appraisal. Piety is a good reason to attend church. “Good” here seems to mean both “(morally) admirable” and “sensible.” The desire to make lucrative contacts is a sensible reason to go to church, though not as admirable. The desire to prevent climate change is an admirable aim though not, according to most people’s worldview, a sensible reason to attend church:  it would be more effective to engage in political action. The desire to resurrect Joseph Stalin is neither a sensible nor an admirable reason to go to church. Reason-​giving explanations of human behavior invite questions about whether the behavior of individuals is justified by their reasons and whether their reasons themselves are justified. In microeconomics courses, word problems often ask students to figure out the hidden rationality behind seemingly irrational behavior, such as airlines charging lower fees for longer flights, as they often do when there is more competition on the longer flight, even though the costs are higher. Such explanations help defend a firm’s behavior against accusations that it is irrational and that it is unfair. As we noted earlier, psychologists and behavioral economists have produced a great deal of evidence that apparently shows that people’s preferences often violate the axioms of utility theory and that people do not care only about their own wealth or consumption. Consider that in preference-​ reversal experiments (Lichtenstein and Slovic 1971, 1973), subjects regularly judge that one bet J is definitely better than another bet K yet express a willingness to pay more for the inferior bet, K. These experimental subjects do not have complete and transitive preference rankings. Such evidence disconfirming central theoretical propositions of positive economics might be worrisome to economists (Grether and Plott 1979).

5.1  Rationality and Positive Economics

73

But most economists stick with the standard theory. Why? One reason is that choice phenomena such as preference reversals are irrational and can be exploited by others. In one study, Chu and Chu (1990) sold K-​bets to experimental subjects for the price they stated they were willing to pay, had them exchange the K-​bets for the J-​bets (which the subjects claimed to prefer), and then purchased back the J-​bets for the lower price that subjects claimed the J-​bets to be worth. At the end of the series of transactions the subjects were back where they started, but poorer. In such a transparent “money pumping” cycle (see also Berg, Dickhaut, and O’Brien 1985), subjects readily figure out that they are being exploited, and they adjust their stated preferences or the prices they are willing to pay for the bets. The vulnerabilities that such irrationalities create tend to place limits on their continued viability. A defender of the standard theory might then maintain that the standard theory of choice must be a good first approximation, because a theory that portrays choice as irrational reveals opportunities for exploiting this irrationality, which in turn leads people to change their behavior. Such a theory of human choosing thus undermines itself. Similarly, economists have argued that profit-​maximizing firms will tend to drive those that behave otherwise out of business, and corporations that use their assets inefficiently will be targets for takeover. Economists can then argue that surviving firms will be profit maximizers. But, even if a firm aims to maximize profits, there is no reason to believe that it will consistently make just the right choices needed to succeed, and it can take a long time to drive firms out of business. At any rate, the fact that the mainstream positive theory of choice is simultaneously a theory of rational choice may bolster economists’ allegiance to that theory. We have claimed that the theory of rationality also serves to link positive to normative economics. If economists maintain that people are in fact self-​ interested (as is common in positive economics) and that they are rational, then a person will prefer one outcome to another if and only if she believes that it is better for her. If, in addition, economists suppose that people’s beliefs are correct, as is also common in positive economics, then she will prefer one outcome to another if and only if it is in fact better for her. Given these assumptions, economists can read off what promotes individual’s well-​being from examining their preferences. This link does not, of course, show that whatever satisfies people’s preferences is in fact good for them, if for no other reason than the possibility that people make mistakes, but it does show that positive and normative economics are tightly connected. We discuss normative economics in Part II of this book.

74

Rationality and Morality in Positive Economics

5.2  Self-​Interest, Rationality, and Morality Where does or should morality enter into the account that positive economics offers of individual choices? Every society has moral norms that can influence the choices that individuals make. People approve of behavior that conforms to the dictates of their morality and disapprove of behavior that violates established moral norms. Those who violate moral norms typically experience guilt or shame as well as resentment and blame from others. Moral norms enable people to coordinate their actions more efficiently than would be possible without a shared morality; such norms can be economically productive. It is in everyone’s interest to live in a society governed by moral norms requiring that people tell the truth, keep their promises, refrain from actions of violence toward others, and treat others with respect. Yet it can sometimes be costly to individuals to do what is morally right. If one is on a sinking ship without a life jacket, there are advantages to being without scruples about hitting some weaker passenger over the head and stealing his life jacket. Cheating on one’s taxes may pay. One might feel guilty later, but people can sometimes overcome their guilt. Although (as we shall explain) going to one’s death in order to comply with a moral norm against theft is consistent with the formal theory of rationality discussed in Chapter 4, it might appear to be substantively irrational: how can it be rational to do what is plainly not in a person’s self-​interest? Can economists incorporate morality into their models without abandoning their view of individuals as both rational and self-​interested? We cannot hope to settle ancient philosophical puzzles about the compatibility of morality, self-​interest, and rationality, but we will explore their ramifications for economics. In particular, we maintain that moral considerations deserve a place in economic modeling and that omitting them has unattractive consequences. This in turn requires relaxing the assumption of self-​interest. It is easy to misinterpret the standard theory of rationality as implying that all agents are self-​interested. One might reason as follows: according to utility theory, it is rational to choose whatever one most prefers. Rational choices are determined by one’s own preferences rather than by anyone else’s. So rational choices are self-​interested. It apparently follows that the individual who, in accord with her own preferences, gives almost her whole income to charity, like the individual who spends every cent on herself, acts in her own interests. Thus, the person who steals someone else’s life jacket is no more self-​interested than the one who chooses death over immorality.

5.2  Self-Interest, Rationality, and Morality

75

The conclusion is paradoxical, and the argument is fallacious. To be self-​ interested is to have preferences directed toward one’s own good, not simply to act on one’s own preferences. What people who are not entirely self-​ interested judge to be best overall does not always coincide with what they judge to be best for themselves alone (Sen 1977). What distinguishes people who are self-​interested from those who are altruistic or malevolent is what they prefer, and utility theory says nothing about the content of preferences. Utility theory leaves open the question of the extent to which individuals are self-​interested. There appears, then, to be no inherent incompatibility between the dictates of altruism or morality and the standard view of rationality. Moreover, if we omit both moral and altruistic reasons from our models of human action, we will miss the fact that individuals can act in ways that decrease their welfare for the sake of others and for the sake of their moral commitments. Some philosophers have held that there is a close association between self-​interest and a more substantive view of rationality. In his classic work The Methods of Ethics, Henry Sidgwick (1901) attempts to derive utilitarianism (which aims at maximizing total well-​being) from intuitions concerning the nature of rationality, but in the end he concludes that what he calls “the method of egoism” –​that is, self-​interested conduct –​is just as rational as utilitarian morality. Many economists would go further than Sidgwick and identify rationality and self-​interest. For example, Robert Frank writes, “I will use the terms ‘rational behavior’ and ‘self-​interested behavior’ to mean the same thing” (1988, p. 2n). If rational behavior and self-​interested behavior always coincide, then moral or altruistic behavior is either irrational or somehow self-​interested after all. We doubt that Frank accepts this implication of his view. “Rational” has many meanings. One of those is “prudent” and hence self-​interested. In this sense altruism is irrational. But there are clearly other senses of “rational” in which those who devote themselves to helping others are perfectly rational. Explaining or predicting human actions requires understanding what motivates people. Postulating that people have complete and transitive preferences does not by itself suffice to explain their actions. As we observed, economic models typically supplement utility theory with the assumptions that consumers care only about the goods and services they can consume and that entrepreneurs care only about their net returns, assumptions that leave little room for altruistic or moral considerations. If these assumptions are necessary for economic modeling, then economic models are incapable of capturing the complexities of human motivation. However, although it

76

Rationality and Morality in Positive Economics

is often useful and enlightening to depict economic agents as entirely self-​ interested, it is not necessary to do so; and it is possible to extend the reach of economic modeling and to enhance its power by incorporating other motives, including moral motives. Refusing to do so leads to explanatory mistakes and, ironically, may have moral implications. Amartya Sen, a Nobel laureate in economics, who is one of the most influential contemporary commentators on economic theories of welfare and choice, has argued persuasively that the economist’s model of choice, which admits only three factors (constraints, beliefs, and preferences), is far too simple to capture the nuances of human behavior. He argues that economists should expand their models to include the possibility that people’s actions are influenced by sympathy and commitment (1977, 1985a). When the consequences of actions for others make an individual, Q, feel better or worse, then one has a case of what Sen calls “sympathy.” When Q has preferences among actions that do not depend exclusively on her expected self-​interested benefits, then one has a case of what Sen calls “commitment.” Sen also maintains that economic models should allow for the possibility of counterpreferential choice: he argues (2007) that people sometimes impose constraints on themselves and do not choose what they most prefer. Sen is right to insist that economic models should be sensitive to a larger range of the motives that influence people’s choices beyond those of self interest. At the same time, we believe that many of his concerns can be accommodated within economic models by recognizing a much wider range of motives as influences on preferences. For example, altruistic concerns can be modeled as competing with self-​interested objectives to influence preferences. Nonegoistic concerns are often more complicated than a simple concern with the well-​being of others. Consider that people may be concerned not only with the level of another person’s welfare or consumption, but also with how it is attained. In particular, someone may wish to be an active contributor to another’s well-​being. Parents not only want their children to acquire good moral characters; they also want some of the responsibility for shaping their upbringing. As Kenneth Arrow has noted, the desire to contribute to the welfare of others may have substantially different implications than mere altruism. If many rich people prefer a higher level of consumption for the poor, there may be a free rider problem: individuals want poverty eliminated, but they would like others to pay for it. If instead each wishes to contribute to helping the poor, the free-​rider problem disappears (Hochman and Rodgers 1969). Among the factors influencing our preferences are desires to do certain things rather than simply to enjoy the consequences of their being done.

5.2  Self-Interest, Rationality, and Morality

77

Moral commitments pose more difficult problems for economic modeling. Morality is not identical to altruism. Consider that an altruistic concern for the members of one’s own family is consistent with racist attitudes toward others in society. Or consider that many people regard prudence as a moral virtue even though it concerns only oneself. A commitment to property rights constrains not only those who would steal for themselves but also would-​be Robin Hoods who would steal to give to the poor. People see their moral commitments as sometimes conflicting with their concerns for the interests of others as well as with their own interests. One way to model the complexity of choice involves assuming that individuals have different preference rankings or utilities depending on whether they consider matters from a self-​interested, an altruistic, or a moral perspective. Alternatively, an economic model can attribute to individuals “metapreferences” –​that is, preferences concerning what their preferences should be. Multiple preference systems very naturally model internal conflict concerning such personal choices as whether or not to smoke (Schelling 1984, pp. 57–​112) as well as moral choices such as whether to contribute to charity (Frankfurt 1971; Etzioni 1988, pp.  177–​80). Some have argued that the influence of morality on choice is better represented in such models (Margolis 1982; Etzioni 1988; Sen 1977). However, while such an approach may make sense of certain choices, it is not clear within the multiple-​preference framework how to explain which ranking will prevail in which circumstances. If that determination is itself made by a consistent preference ranking at a higher level, then it makes sense to reinvoke standard utility theory. Sen’s notion of counterpreferential choice raises two different concerns. First, allowing that rational choice may depend on other factors in addition to constraints, beliefs, and preferences weakens the explanatory and predictive power of economic models. It would no longer be possible to infer what people will do from their beliefs and preferences and the constraints. (Of course, the fact that explanation becomes more difficult does not mean that the model is less accurate.) Second, if self-​imposed constraints really are constraints, then the choices people make are not counterpreferential after all. Someone who chooses to walk rather than to fly unassisted, as she would prefer to do, is not choosing contrary to her preferences. Nor is someone who refuses to take the most comfortable chair, because of a self-​ imposed constraint of courtesy. Whatever modeling strategy one adopts, we have argued that mainstream economists need to broaden their theory of economic behavior in order to take account of the influence of moral norms. Doing so requires

78

Rationality and Morality in Positive Economics

recognizing that economic agents are not always self-​interested and that self-​interest is not always aimed at material goods. Consumers do not care only about the commodities and leisure they may consume, and entrepreneurs do not care only about the bottom line. Sometimes they also want to benefit or to harm other people, and even when self-​interested they may care more about the esteem and affection of others than about the size of their houses or the style of their clothing. For example, it would be as foolish for those designing a health-​care system to suppose that doctors and nurses do not care whether their patients live or die as to suppose that doctors and nurses do not care about their salaries (Hausman and Le Grand 1999). Economists need to throw off the straitjacket of assuming that preferences are always self-​interested.

5.3  The Moral Danger of Ignoring Morality People sometimes cheat on their taxes. What motivates some individuals to cheat while other individuals scrupulously pay their taxes? Economic analyses typically focus on the factors that influence the gains from cheating (marginal tax rates, most obviously) and the costs (probability of detection and probability and severity of legal penalty if detected). These cost–​benefit considerations provide reasons for or against cheating on taxes and help identify circumstances when cheating can be expected to promote an agent’s material self-​interest. But the expectation of gain and the fear of being caught are not the only relevant factors: not everyone will cheat who expects it to be profitable to do so. Many people think it is wrong to cheat. Others may cheat, even when there is no expected monetary benefit, because they object to taxation or to the policies that their taxes fund. Economists who regard the self-​ interested reasons for cheating as rational and largely ignore moral reasons not to cheat may unintentionally justify cheating. If there is a sufficiently large expected return, then cheating is rational from the point of view of an individual concerned only with personal net income. On the assumption that everyone is materially self-​interested, those found evading their taxes are simply incautious or unlucky. Although illegal, their actions are perfectly sensible. Focusing only on the tangible costs and benefits of cheating may suggest that there are no other reasons to comply. Economists can thereby find themselves implicitly endorsing conclusions that are compatible with the cynical view that moral commitments are rare or foolish. Indeed, Marwell and Ames (1981) and Frank et al. (1993) found that those who studied economics were more

5.3  The Moral Danger of Ignoring Morality

79

likely to be self-​interested and to free ride on the efforts of others in cooperative situations than noneconomists. Explanations of choices in terms of preferences for larger bundles of commodities or greater net revenues are often entirely adequate: no heavy moral lifting is involved in predicting Honda’s decision about the color schemes for next year’s new cars on the basis of its expectations concerning sales and ultimately profits. But when economists proceed to analyze morally fraught choices in the same manner, larger issues are at stake. It is one thing to assume, as Adam Smith did, that market incentives would determine behavior “within the rules of justice” (1776, bk. II, ch. 4). It is quite a different thing to assume that self-​interested calculations will determine whether people obey those rules or observe any moral limitations on their choices. Indeed, there is an odd feature of many economic models. For at the same time that economists suppose that consumers and firms are self-​interested, they typically suppose that agents obey the law, whether or not doing so furthers their interests. (Studies of tax compliance are in this regard atypical.) Why does Homo economicus obey the law rather than cheating, lying, or stealing whenever it suits him? Suppose that a firm is considering dumping toxic wastes into a stream and poisoning an area of marshland. If doing so is illegal, economists typically assume that it is not among the feasible alternatives. If it is legal, then most economists will take the decision to depend on what course of action maximizes the firm’s net revenue. If morality has no effect on the behavior of firms, why should legality? To the degree that economists assume that the only reasons to be sought in explaining the firm’s behavior are self-​interested reasons –​like those justifying car color choices –​their analysis will tend to justify the firm’s dumping toxic waste if and only if the costs of punishment outweigh the benefits to the firm. It is as if the only thing wrong with Willie Sutton’s bank robbing was that he underestimated the risks of being caught or that the government failed to make bank robbery sufficiently unprofitable. A common argument an economist might offer for ignoring moral restraints on a firm’s polluting behavior would be that competitive pressures do not permit firms the luxury of moral scruples: if a firm took the high road with respect to pollution, then it would be driven out of business, and the less scrupulous surviving firms would continue to pollute. This argument is not conclusive. Notice first that the conclusion –​that taking the high ground never does any good  –​does not logically follow from the assumptions. Even if competitive pressures were to drive the nonpolluting firm out of business, those pressures may not lead competitors

80

Rationality and Morality in Positive Economics

to cause the same level of pollution. There are also serious challenges to the claim that the price of morality is bankruptcy. The conclusion does not hold if all firms show the same moral concern for the environment, for then none would be at a competitive disadvantage. So, the argument takes for granted that some firms do things such as poisoning marshes in order to increase profits. But should this behavior be taken as given? Furthermore, even if this assumption is granted, the strength of competitive pressures on a firm varies. For example, from the time of its formation in 1901 to 1991, the market share of the United States Steel Corporation constantly decreased, yet it survived (Mueller 1992, pp. 309–​10). All the competitive pressures argument shows is that not polluting has costs and may fail to achieve its purposes. To take avoiding bankruptcy as a sufficient reason for polluting a marsh presupposes that there is a limit on the weight that owners or managers give to moral considerations. This judgment is less objectionable than is the presupposition that moral scruples have no weight at all when there is a conflict between profit and social good, but it likewise tends to excuse or justify the firm’s behavior. If a firm facing bankruptcy hired gangsters to burn down a competitor’s factory or murder its owner, few economists would show the same attitude toward an explanation in terms of economic pressures. One might argue that just as it may be permissible to steal when an individual’s life is at stake, so it may be permissible for a firm to cheat or pollute when its very survival depends on doing so. But the survival of a firm is not on a par with the life of an individual. Sometimes people are better off with the demise of some firm than they would be with its survival. Economists have shown increasing interest in studying morally freighted activities such as theft, violent crime, environmental pollution, and trustworthiness. When assessing explanations in such areas, it is important to recognize that moral as well as self-​interested reasons can help explain behavior and that a commitment to explaining all behavior in terms of self-​ interested rationality tends to excuse or justify it. Most of us regard a desire for more commodities as a justification for working longer hours, but not as a justification for robbing banks. These moral and prudential judgments are reflected in our satisfaction with the explanation of the longer hours worked and our demand for further explanation of the bank robbing. We have argued that purely self-​interested explanations of immoral behavior can have the effect of defending that behavior. To see nothing remarkable in destructive behavior motivated only by the prospect of additional profits or the avoidance of bankruptcy is to endorse (or to be blind

5.4  The Influence of Moral Norms on Economic Behavior

81

to) a pernicious moral cynicism. Equating rationality with self-​interest and then explaining and implicitly justifying morally questionable behavior solely on grounds of rationality rely on questionable empirical assumptions; and should such explanations become widespread, they risk becoming self-​fulfilling prophecies.

5.4  The Influence of Moral Norms on Economic Behavior One reason not to exclude moral considerations from economic models is, as we have seen, that doing so can have important moral implications. Another reason is that these models will make false predictions, because morally motivated actions exist and have important economic implications. Consider the phenomenon of “efficiency wages.” Usually economists maintain that individuals are paid more because they are more productive. But paying people more may also induce them to work more productively. The idea of an efficiency wage is that sometimes simply raising the wages of a group of workers motivates them to work more effectively. Efficiency wages are wages above the market-​clearing rate that motivate workers to be more productive. Efficiency wage theory apparently helps to explain involuntary unemployment. If there are unemployed workers who are willing to work for lower wages, why wages do not fall to a level consistent with employing the entire labor force? Efficiency wage theory answers that reducing wages could cost firms more from lost productivity than they would gain from the savings in wage payments. Efficiency wage theory apparently also explains the existence of so-​called dual labor markets. The U.S. economy appears to have both a growing “casual” labor market characterized by high turnover and low (market-​clearing) wages (fast food workers and interns) and an increasingly tenuous “career” labor market characterized by lower turnover, higher wages, and rationed entry through queues of applicants for a limited number of jobs. The presence of efficiency wages in firms whose jobs are in the career labor market category may be part of the explanation. Firms could hire equally well-​qualified workers at lower wages, but at lower wages the workers would not be motivated to work as productively, and higher turnover would be costly. A number of economists have attempted to explain efficiency wages in models that assume rational self-​interested behavior on the part of both firms and workers. For example, Bowles and Gintis (1993) point out that it is difficult for firms to monitor and control the performance of their workers. If the only sanction firms have is the threat of firing, then firms are not

82

Rationality and Morality in Positive Economics

going to be able to motivate their workers unless firing is costly to the workers, and firing will not be costly to workers unless firms pay them more than market-​clearing wages. Notice that this account does not appeal to moral norms in its explanation of wages. George Akerlof, in contrast, has offered an explanation of efficiency wages in terms of what he terms “partial gift exchange” (1982). Akerlof revisits a case study developed originally by the sociologist George Homans. Homans examined “cash posters” at the Eastern Utilities Company, whose job was to record customers’ payments on ledger cards at the time of receipt. It was easy to measure the workers’ output and to compare it to the company’s performance standard. Of the group of ten workers, Homans found that two barely met the standard of 300 postings per hour, but the average number of postings per hour (353) was 17.7 percent above the standard. Despite considerable variation in performance (the range was from 306 to 438 postings per hour), all workers received the same wage (which was above the market-​clearing wage), differential promotion was not an issue, and there were no serious efforts to punish or threaten the slower workers. These facts are hard to reconcile with standard economic theory. If some workers prefer gossiping, coffee drinking, or daydreaming to recording payments on ledger cards, then those who post much more than the standard 300 should slack off or demand more pay. If the company, for its part, wants to increase its returns, then it seems that it should raise its performance standards. If the more productive workers exceed the standard without any reward, then they should be willing to increase their rate of postings further in exchange for additional compensation. There should be mutually beneficial arrangements in which firms pay higher wages to more productive workers, who in turn increase their output. Akerlof argues that moral norms explain these facts. Both the firm and the workers adhere to a norm of “gift exchange” for effort and wages above the minimum expected. Workers receive a gift of more than the market-​ clearing wages, and in return they give more than the minimum standard of effort. These transactions are governed by the norms of “a fair day’s work for a fair day’s pay.” To explain the absence of pay differentials and the firm’s willingness to maintain low-​performing workers in their jobs, Akerlof appeals to further norms of fair treatment among workers. Introduction of pay differentials or efforts to dismiss the less productive workers might result in withdrawal of some of the extra effort. (Akerlof reports evidence that slowdowns occurred in the past when the company tried to toughen up.) In the presence of a “gift” from the workers of 17.7 percent increased output, the firm has reason to respect the workers’ norms of fair treatment.

5.4  The Influence of Moral Norms on Economic Behavior

83

It is hard to determine whether Akerlof ’s explanation of the cash posters’ case is correct because, as Akerlof notes, it is possible to construct alternative explanations of these facts that avoid any appeal to norms and rely instead, as Bowles and Gintis do, on turnover costs, monitoring costs, measurement difficulties, and the like. Nevertheless, the sociological explanation is plausible: norms regarding gift giving are familiar in the United States, and the “ties of sentiment” that develop among workers and even between workers and their employing firm are a plausible ground for the emergence and maintenance of such norms. Akerlof ’s account raises questions about the origin and determinants of social and moral norms, questions that for now have no entirely satisfactory answers. It is, however, a huge leap from recognizing these questions to claiming that good answers to them will reduce everything to self-​interest and so produce an explanation of the sort mainstream economists prefer. Whether there is an adequate account in terms of economic self-​interest is precisely what is at issue here. As this example illustrates, normative aspects of work relations may influence the macroeconomics as well as the microeconomics of labor markets. In the view of some, the “efficiency wage” story offers a promising route to explaining the existence of involuntary unemployment in equilibrium. (In addition to Akerlof, see Solow 1981; Bowles 1985; Blinder and Choi 1990; and De Vroey 2004.) However, an adequate macroeconomic explanation of unemployment grounded in efficiency wage theory requires solving further puzzles. Why do not unemployed workers undercut the prevailing above-​market-​clearing wage by offering to work for less than those with jobs? Why do not employers actively seek such workers, perhaps differentiating the workforce to pay more to those with established jobs and less to the eager newcomers? Solow (1990, ch. 2) and Weibull (1987) suggest that further social and moral norms come into play –​for example, a solidarity norm among workers that discourages bidding against their fellows for jobs and a norm among employers against seeking to undercut their existing workforce. There are other areas of economic life in which moral norms appear to be influential. The regulation of negative externalities such as air pollution and littering is accomplished partly through legal sanctions, but norms against such antisocial behavior also play a significant role. The same may be said about the provision of public goods in general and about voting. People cast their ballots, sometimes at significant cost to them, even when the tangible benefits from their single vote are vanishingly small. Public transport in many European cities operates on an honor system with only sporadic

84

Rationality and Morality in Positive Economics

ticket checks. One cannot assume, however, that norms will always do the work: honor systems on public transit with sporadic checks work smoothly in much of Germany, Austria, and Switzerland, whereas cheating is common in France.

5.5  How Do Norms Motivate Human Actions and What Sustains Them? Automobile manufacturers, such as Volkswagen, are not above lying and cheating. Yet they also sometimes initiate costly recalls of their product when they are not required to do so. One possible explanation is that there are norms requiring that manufacturers repair flaws in merchandise. Thus, one might explain a particular voluntary recall by mentioning the norm and citing the generalization that agents tend to follow norms. Few economists would be satisfied with such an explanation. They would insist that one must explain why the management decided to comply with the norm and why the norm persists. Some economists would go further and maintain that the recall has not been explained until it has been shown that it is in the material or economic self-​interest of the car company. What most economists would like to show is that norm following increases the profits of firms or the consumption or leisure of individuals. Although it is not, in our view, reasonable to insist that only material self-​interest motivates, there are genuine questions about how norms are sustained. Norms do not descend from the heavens, as the Ten Commandments are alleged to have done; they are somehow created, modified, sustained, and enforced by people. Indeed, sometimes they break down. An explanation of behavior that merely cites a norm –​without considering what sustains and enforces it –​is likely to be shallow. Let us shift to an example that involves individuals and where an explanation in terms of moral considerations is more persuasive than in the case of an automobile recall. In a classic experiment, researchers left wallets containing identification and a small amount of cash on the streets of New York City (Hornstein, Fisch, and Holmes 1968). Almost half were returned with the cash intact. It is hard to understand how returning the wallets could have been materially advantageous for those who found them. It is troublesome to pack up a wallet and mail it. Presumably the reason why most people returned the wallet was either a direct altruistic concern for the owner of the wallet or a moral norm against illegitimately appropriating other people’s property. Since altruism tends to be weaker when one has no vivid impression of the other person, it seems reasonable to assume that

5.5  How Do Norms Motivate Human Actions and What Sustains Them? 85

the norm against theft is the main explanation for why people returned the wallets (Batson 1993). How does a norm motivate individuals to comply with it? One possibility is that the moral reasons justifying the norm motivate the compliance. People’s desire to do what is right can motivate them to follow norms, even when doing so will not benefit them materially. Economists and sociologists influenced by economists have tended to focus not on moral norms but instead on the individual sanctions  –​ the costs and benefits –​motivating compliance with norms. If the finder returns the wallet, she feels good about herself, and the receipt of a thank-​you note makes her feel even better. She can tell her family, friends, and associates and in response gain their trust and praise. Indeed, in this way it is possible for her to garner some material advantage (though usually not enough to compensate for the effort of returning the wallet). On the other hand, if she does not return the wallet, then, if she has a conscience, she will feel guilty. Moreover, if her failure to return the wallet becomes known to others, she may be censured or may suffer from knowing that she has lost their esteem, even if the censure of others is never expressed (Pettit 1990, p. 740). It is important to be careful about viewing guilt feelings as externally imposed “costs.” These feelings have an important cognitive aspect and would not exist if the person did not believe that he or she did wrong. The idea that these reactions are like penalties others impose on us treats conscience as if it were something external, like some sort of immaterial ball and chain that prevents us from doing what we want and that punishes us when we lapse (Nussbaum 2001). That said, it is surely true that sanctions, both internal and external, help explain compliance with norms: there are often benefits to complying with norms and costs to violating them. These individual rewards and costs are important, but they are not the whole story. To say that individuals comply with norms because of sanctions does not get one far. Sanctions have to be applied by people, and it is typically no more materially advantageous to punish or reward others for complying with norms than it is to comply with norms oneself in the first place (Coleman 1990). Given that others will punish violations and reward compliance, it may be advantageous to comply with norms. But this only explains why one person should comply with norms if others do. It does not explain why people sometimes comply with norms even in the face of widespread violation, and it does not explain why the norms persist. One might attempt to explain why norms persist by citing the benefits they provide to groups. Communities with strong norms of truth telling

86

Rationality and Morality in Positive Economics

and promise keeping usually do better than groups without such norms. But no single individual can have much effect on the norms obtaining in any sizable group or on average compliance. Although the social benefits of a system of norms may provide a moral reason for complying with them, individuals may do better by free riding on the norm-​following actions of others. Since there are few sanctions attached to not sanctioning others, rational individuals who are concerned only about their own profits, consumption, or leisure will not trouble themselves to punish or reward others for violating or complying with norms. So if one’s fellows are rational and self-​interested agents of the sort found in economic models, then one should have little to fear in the way of sanctions against violating norms and little to expect in the way of reward for complying with them. Norm following cannot be explained in terms of economic self-​interest alone. Consider “public goods” experiments, where subjects will earn more money if everyone contributes, but each does better by free riding on the contributions of others.1 In these experiments, there is initially a high rate of contribution that decays over time as the choices are repeated. These experiments suggested to economists that individuals are basically self-​ interested and that the initial cooperative behavior reflects confusion. But, if one modifies the experiments and allows group members to punish others (at a cost to themselves), one finds a very high rate of apparently cooperative behavior (Fehr and Gächter 2000; Camerer and Fehr 2006). Because some individuals are willing to punish selfish behavior at a cost to themselves, cooperative behavior can arise. The experimental results thus support the view that people’s motives are mixed. Some are self-​interested. Some are governed by norms. Others are willing to conform to norms only on the condition that others do, too. Of course, these experimental results are not an ironclad demonstration that norms can only be generated or sustained by a community if some of its members are not self-​interested. Involuntary attitudes of approval or disapproval may create the needed incentives to conform to norms (Pettit 1990; Brennan and Pettit 2004), or it may be that the costs and benefits of sanctioning sometimes make sanctioning rational (Coleman 1990, ch. 10). In any case, economists need to find a place within their theory of economic For example, members of groups of four may face an independent and anonymous choice to “invest” some or all of the $10 they have been given. The total investment is doubled and divided equally. If everybody invests, then each receives $20. But regardless of what the other members in the group does, each member does better for himself or herself by not investing anything. For example, if the other three members invest the full $10, and you invest nothing, you wind up with $25 ($60 divided by 4 + $10). 1

5.5  How Do Norms Motivate Human Actions and What Sustains Them? 87

behavior for a discussion of when norms will form, of how much influence they might have, and of what their consequences will be. In recent years, there has been a great deal of interest concerning the role that biological natural selection and broader conceptions of social evolution play in the development and persistence of moral norms. These lines of inquiry focus on the complex interplay between the advantages to groups and to their members that shared norms provide and the advantages that may accrue to individuals if they are able to free ride on the norm following of others. One way to achieve compliance with norms is to build it in biologically, and practitioners of “evolutionary psychology” have argued that receptiveness to following norms and even specific normative commitments –​such as caring about children or being disposed to act reciprocally toward those who treat us well –​have a basis in biological evolution. These arguments are controversial. Many of the processes that are responsible for norms are social rather than biological and are only roughly analogous to biological evolution. In addition, while natural selection has been compatible with the moral capacities and proclivities of human beings, regrettably it has not ruled out our aggressive and destructive tendencies. Scholars have asked how norm compliance of individuals affects how individuals and groups fare over time. The tools of game theory are revealing here. In repeated interactions, cooperative strategies may have a high survival value and tend to predominate, but there may be benefits to free riding, too. Temptations to free ride may be undermining norms within groups at the same time the cooperation that norms make possible may be helping those societies with strict norms and high compliance to supplant societies with weak norms. Norms obviously have social consequences, which may be beneficial or harmful. For example, Arrow (1974) and Reder (1979) have shown how the virtues of honesty and trustworthiness may promote economic efficiency in circumstances of uncertainty. Arrow has also argued more specifically that medical codes of ethics may be an efficient and socially desirable response to physicians’ opportunities to exploit their informational advantages over patients. When there are asymmetries in information, markets do not always work well; medical care is an especially difficult case, because the need for treatment may be urgent, preventing patients from shopping around or seeking out further information. Licensing, codes of ethics, and social norms all help justify the trust that patients often need to place in health-​care providers. Arguments of this kind may contribute to the justification of these norms by showing that they lead to good consequences, including consequences for economic well-​being. But the fact that norms

88

Rationality and Morality in Positive Economics

promote efficiency or group solidarity is not sufficient to justify them. For example, traditional codes of medical ethics often discouraged doctors from advertising their prices. While this may have helped discourage doctors from “cutting corners,” it also made it more difficult for patients to engage in comparative shopping. The good consequences of a moral norm are also not by themselves enough to explain its emergence and persistence. As Elster (1989a, b) has argued in his critique of functional explanations, to explain the existence or persistence of a norm also requires an account of the causal mechanisms by which the favorable consequences produce or sustain the practice (see Cohen 1982 for a rebuttal). Moreover, moral norms do not always have good economic consequences. Norms regarding work rules and wages may promote wage equality, high wages, and stable and productive working groups. Yet these norms may also contribute to unemployment, and more established “in-​groups” may benefit disproportionately from these norms while disadvantaged groups are excluded from the benefits. We have a great deal to learn about the causes and consequences of social norms, including moral norms, their moral justification, and the ways in which they motivate individuals. Economists can contribute to this project, but only if they acknowledge the importance of moral commitments and explore how the interaction of different motives determines people’s preferences and governs market outcomes.

5.6  Conclusions: Rationality, Morality, and Positive Economics Positive economics explains choices in terms of factors that are subject to evaluation. This does not mean that economists must evaluate them; they can cite the preferences and beliefs that explain an individual’s choice without appraising them. But whether or not the evaluative questions are answered or even explicitly mentioned, the questions are always there; and it is sometimes important to ask them. In addition to the intrinsic interest of finding out what drives tastes or perpetuates myths, the assessment of beliefs and preferences may suggest further topics for empirical research. So if, for example, economists explain some market phenomenon by a widespread belief among market participants that interest rates will remain steady, and if economists go on to evaluate that belief as unsupported by data and theory, then they will want to pursue the question of how these unsupported beliefs have propagated and persisted. The theory of “rational expectations” in finance and macroeconomics emerged from just this kind of inquiry.

Suggestions for Further Reading

89

Although some economists are willing to cite people’s moral beliefs in explaining some of their choices, many are strongly inclined to cling to the “solid ground” of material self-​interest in constructing economic explanations. This tendency may be at its strongest in the teaching of economics, where there is an interest in building up students’ skill in seeing the hidden hand of self-​interest behind behavior that is not obviously self-​interested. Alas: in combatting naiveté, economists may be cultivating cynicism rather than wisdom. Modeling behavior as if morality had no influence encourages the belief that morality has no influence and emboldens individuals to turn a deaf ear to their moral commitments. The idealized agent of contemporary economics is a rational agent. But the view of rationality that most economists endorse –​utility theory –​does not by itself provide a rich enough picture of individual choice to permit one to discuss the character, causes, and consequences of moral behavior. Economists need not, of course, aspire to provide a general theory of human action. Yet they should not shrug their shoulders at the difficulties in meshing moral behavior with economic rationality. For it seems –​and we have argued –​that moral behavior can have important consequences for economic outcomes, and economists’ application of utility theory itself has moral implications.

Suggestions for Further Reading An excellent discussion of the relationship between facts and values, with special attention to the views of the economist Amartya Sen, is Putnam (2002). There has been a good deal of discussion of so-​called folk psychology by philosophers of psychology. Discussions can be found in Stich (1983), Dennett (1987), Dretske (1988), and Churchland (1989). For discussions of reasons versus causes and of the possible difficulties involved in causal explanation of human behavior, see Winch (1958), Melden (1961), von Wright (1971), Rosenberg (1976, ch. 5), Davidson (1980), and Rosenberg (1995, ch. 2). There is a considerable literature concerning preference reversals. The phenomenon was first discussed in Lichtenstein and Slovic (1971), and the first discussion by economists is Grether and Plott (1979). More recent views can be found in Tversky, Slovic, and Kahneman (1990); Tversky and Thaler (1990); and Lichtenstein and Slovic (2006). An overview of economists’ reactions can be found in Hausman (1992, ch. 13). Various aspects of the relationship between economic rationality and morality are discussed in Frank (2004).

90

Rationality and Morality in Positive Economics

For the general contrast between “self-​interest” and “present-​aim” theories of rationality, see Parfit (1984, pt. 2). For more on the relations between rationality and self-​interest in economics, see Sen (1977) and Hausman (2012.) Useful discussions of norms can be found in Bicchieri (2006); Elster (1989a,b); Coleman (1990); Pettit (1990); Goldfarb and Griffith (1991); Frank (1993); Ben-​Ner and Putterman (1998); and Brennan and Pettit (2004). Efficiency wage theory is discussed in Akerlof (1984), Akerlof and Yellen (1986), and Solow (1990). An informed though polemical popular overview of evolutionary psychology can be found in Pinker (2002). The possible influence of moral factors in determining the efficiency of alternative ways of organizing the workplace is discussed in Reich and Devine (1981); McPherson (1983a); Putterman (1984); Bowles (1985); and Bowles and Gintis (1993). A related but distinct point is the role of trust in governing worker–​employer relations, which is discussed in Arrow (1974), McKean (1975), and Gambetta (1988). General problems concerning the explanation of public goods provision and the role that moral norms may play can be found in Schelling (1978); Kelman (1981); Taylor (1987); Elster (1989a); and Mansbridge (1990). For discussions of the existence of altruism and its possible importance in economics see Arrow (1972); Boulding (1978); Collard (1978); Becker (1981); and Batson (1993). Multiple-​preference rankings are introduced and discussed in Frankfurt (1971); Sen (1977a); Schelling (1984); Etzioni (1988); and Brennan (1989).

Questions for Study and Discussion 1. What, in your view, is the relationship between rationality and morality? Is morality necessarily rational? Can it be irrational to be moral? 2. What role do you think that morality plays in economic decision making? Why? 3. What is the difference between morality and altruism? 4. On one interpretation, Socrates argued that it can never be in one’s interest to behave immorally. Do you think he is right? Why or why not? 5. Does the claim that it is irrational to have intransitive preferences make it more likely to be true that people’s preferences are in fact transitive?

Questions for Study and Discussion

91

6. Why is the standard theory of rationality insufficient by itself to explain what people do? 7. What is the difference between the reasons for an action and the causes of an action? 8. If people are asked to choose between bets whose outcomes are nearly certain, such as winning $4.00 with probability 35/​36 or losing $1.00 with probability 1/​36 and bets that one is much less likely to win, such as winning $16.00 with probability 11/​36 and losing $1.50 with probability 25/​36, some people prefer the bet they are more certain to win, while others prefer the riskier bet. Preference reversals appear only among the former –​that is, people claim to prefer the more certain bet but are willing to pay more for the riskier bet. Can you think of any explanation for this strange phenomenon? 9. Why should economists suppose that people are motivated to obey the law, yet not motivated by moral commitments? 10. Should economists attempt to incorporate moral norms into their explanations of behavior such as recycling, career choices, or the employment practices of firms? What do you think? 11. What do you think of the argument that it is impossible for firms to take moral considerations into account without being driven out of business? 12. In order to explain how moral norms influence an individual’s behavior, is it necessary to show that following moral norms is in an individual’s self-​interest? 13. Suppose you had a chance to engage in a clearly illegal activity that paid off $10,000,000 if you got away with it, and penalized you $20,000,000 if you were caught. What is the probability of being caught that makes it unprofitable to engage in the illegal activity? Is there some probability at which engaging in the activity becomes morally justified? Does your answer depend on the nature and consequences of the activity? 14. Suppose you had persuasive evidence that it was in the interest of the richest citizens that low-​income families not vote. Would such evidence be sufficient to explain why low income people are much less likely to vote than those with higher incomes? Why or why not?

SIX

The Ethical Limits to Markets

In Chapter 5, we considered whether moral norms constrain what people do and whether they constrain, or ought to constrain, what it is rational to do. In this chapter, we turn from questions about how morality affects individual behavior to questions about how morality limits and ought to limit the use of and scope of markets. Markets are, of course, the central institution that economists study. Despite this, there is little attention paid to explaining exactly what a market is: most introductory economics textbooks do not even bother to define it. Markets may seem simple – a​ set of easy equations that can be depicted by supply and demand curves –​and natural –​a spontaneous form of ordering relationships between people. But in reality, Adam Smith’s “simple system of natural liberty” is a complex social institution. Any market presupposes property rights, so that trade and theft can be distinguished. All markets are governed by rules. These rules specify what the basic elements of a market are: Who counts as a trader? What counts as a commodity? What counts as a contract? These rules also specify how these elements may interact: How frequently may any one individual make trades? What kinds of actions may traders take against one another? In every developed economy that uses markets, law plays an important coordinating, regulating, and enforcing function (Satz 2010). The completely “free market” is an illusion. Most market transactions are also “arm’s length” in that the players in a market are strangers to one another. Because of this, and because it is often hard to judge the quality of tradable goods (think of used cars), markets depend on the availability of information and on social norms demanding trustworthiness. While Homo economicus may be self-​interested, he must not lie, cheat, or steal if markets are to function well. This point is often overlooked, but it is important, especially when exchanges are spread out in time and space. The creation of the more or less trustworthy, self-​interested 92

6.1  Are There Any Limits to the Reach of the Market?

93

trader is a social achievement, which plays an important role in the functioning of contemporary markets; untrustworthy agents can undermine markets, especially when they have disproportionate market power.

6.1  Are There Any Limits to the Reach of the Market? From the standpoint of modern economics, in theory almost any good or service can be distributed through a market, as long as the good or service in question is scarce, and it is possible to define and enforce property rights to it. Thus, air is not bought or sold on the market (since it is in effect limitless),1 but clean water is. Some goods are best distributed by nonmarket mechanisms or in highly regulated markets because of their technical features: for example, goods that form natural monopolies. It makes little sense to build two competitive water-pipe systems for a village, even though it is possible to do so. But aside from such cases, standard economics argues that efficiency provides a reason to distribute goods through the market, whether the goods in question be guns or butter, human hearts or automobiles, slaves, pencils, or toxic waste. The equations that will appear on the economist’s blackboard are indifferent to what is being equated. In fact, in its most abstract form, general equilibrium theory presupposes complete markets –​a price for anything and everything in all possible present or future states of the world. All societies have, however, limited what can be bought and sold. Human beings, votes, national sovereignty, human organs, criminal justice, and religious salvation are some of the items that many societies distribute, if at all, by nonmarket means. Whether to allow trades of some of the items on this list, such as human kidneys, are matters of current debate, and some have tried to extend the market’s reach as far as possible. Even as we write, there are online Web sites where adulterers can buy an alibi, individuals can offer their foreheads as advertising space for companies in return for a fee, and “immortal souls” have been for sale on Ebay. This expansionist philosophy contrasts with older views like those of Aristotle and Aquinas, who regarded some transactions as impermissible and some ways of making money as unnatural or as a form of vice. In addition to debates about particular markets and particular transactions, some social theories express a more generalized anxiety toward the market as an agent of social corruption:  as undermining a sense of the Although see http://​blogs.wsj.com/​chinarealtime/​2014/​03/​31/​bags-​of-​mountain-​air-​ offered-​in-​smog-​addled-​chinese-​city/​ 1

94

The Ethical Limits to Markets

sacred, as eroding the value of altruistic actions, as alienating and facilitating exploitation, and as weakening important human solidarities (see G. A. Cohen (2009), Hirschman (1985), Marx (1977), and Polanyi (1944)). We can see these anxieties at work in such cultural touchstones as Arthur Miller’s Death of a Salesman, where Willy Loman’s dismissal by his boss is depicted as both cruel and driven by a relentless market logic, “cause you got to admit, business is business.”

6.2  Market Virtues Notwithstanding these criticisms, a great deal can be said ethically on behalf of markets. First, markets are often the most efficient way to allocate goods and services. Through their reliance on a division of labor, markets allow people to specialize in distinct tasks and enhance efficiency. Although a focus on efficiency can narrow ethical considerations, efficiency is of great moral importance. An inefficient medical system means that people have to pay more for treatment or that they receive fewer treatments. Economic decline and stagnation seem invariably to hit the poor the hardest. Second, markets permit the simple freedom of being able to choose among alternatives as one pleases (provided that one has the means, of course). Because markets operate through millions of individual decisions, they empower people to act as choosers, at least within this domain. For example, you do not typically need anyone else’s permission in a market situation to decide to buy one type of food over another. You can be a vegan, consume nothing but meat, keep a kosher kitchen, or eat no grains. This makes market transactions an ideal form of social interaction among people who disagree with one another or have different tastes –​whether the disagreements are about food, or religion, or music, or the aims of life. When markets function well, they display important virtues, in addition to efficiency. Because there are consequences to each individual market choice, markets encourage a calculating rationality. It makes little sense for a person on the market to hold the intransitive preferences that would turn him into a “money pump.” In addition, the fact that markets depend on the behavior of a very large number of transacting actors has the effect of making no one dependent on any single overlord. Adam Smith observed that this erosion of “servile dependency” was among the market’s most powerful effects: “Each tradesman or artificer derives his subsistence from the employment, not of one, but of a hundred or a thousand different customers. Though in some measure obliged to them all, therefore, he is not absolutely dependent on any one of them” (Wealth of Nations, p. 420). However, when markets are close to

6.3 Market Limits

95

being monopolies, or when what is traded is urgently needed by one of the trading parties, then this independence is threatened. Markets also distribute information to people in the form of prices. Hayek exaggerated when he saw this mechanism as enabling a form of “spontaneous order,” since regulations are needed to deal with everything from transaction costs –​for example, with the frictions that arise when information is costly to obtain –​to specifying and enforcing property rights. Nevertheless, markets do decentralize many decisions and permit price information to influence the production and distribution of many goods. By providing the trading parties with all the relevant information (which sometimes is a big assumption), markets make it likely that they will only enter into trades of benefit to each of the traders.

6.3 Market Limits In an ideal market transaction, individuals are free to exchange or not to exchange, and their interactions have no effects on others. Many actual markets are, however, not like this. Interactions among individuals who are exchanging goods and services may provide benefits to third parties or impose costs on them. Some market exchanges generate externalities –​that is effects on others who are not party to the exchange. Climate change is an example of a massive externality; every time we three drive our cars we contribute to what appears to be a significant and harmful change in the Earth’s climate. In addition, some goods are indivisible and nonexcludable, such as military defense, and some markets are characterized by asymmetric information between the suppliers and consumers, such as health care and used cars. Where markets are inefficient for one or another of these reasons, freedom may be threatened, and it may be possible to make people better off by interfering with and regulating or preventing certain market interactions. Whether market regulation or alternative forms of allocation can in fact improve on outcomes needs to be considered on a case-​by-​case basis. Sometimes a little bit of inefficiency is preferable to a lot of centralized oversight, and government has its own inefficiencies and poses its own threats to freedom. Limiting markets limits some individual freedoms. People who have things they want to sell and the others who want to buy them are prevented from doing as they like. Sometimes such limitations may wind up enhancing freedom overall –​some people argue that this is the case with enforcing a minimum wage –​but the immediate direct effect of limiting markets consists in making some choices unavailable to people.

96

The Ethical Limits to Markets

Sometimes these limitations of freedom can be justified on the grounds of efficiency. When voluntary transactions cause harmful externalities  –​ that is, negative consequences on third parties –​then there is a trade-​off between protecting an individual’s freedom to engage in a particular transaction and protecting third parties from harm. In assessing this trade-​off, one needs to consider the costs of enforcing the limitations on individual interactions, which may be significant. Some markets that are limited, such as the purchase and sale of marijuana, involve few significant externalities, and the case for these limitations is more fragile. John Stuart Mill famously argued in On Liberty that if the voluntary interactions between competent individuals do not harm or risk harming third persons, then those interactions should not be limited by law or public opinion. Mill’s argument against interfering in actions that do not harm others remains controversial, and he himself pointed to some qualifications when he argued, for example, that individuals should not be allowed to sell themselves into slavery. Economic analysis cannot by itself resolve the major questions of whether and when to limit the use of markets for ethical reasons. For example, it might be objected that many of the problems with markets can be solved through markets themselves. Thus, we can price the externalities generated by pollution and sell the rights to pollute to companies. Once it is properly priced, the externality from pollution disappears. Ronald Coase argued that once property rights are clearly defined, transacting agents may be able to negotiate a mutually beneficial solution in the face of externalities and other market failures (1960). Some economists argue that most externalities can be dealt with in this way. The theory of market failure does not tell policy makers what the limits of the market should be. The fact that legislators can, at least in principle, respond to market failures by introducing new markets does not tell us whether it is a good thing to do so. Are there moral reasons to limit certain markets? For example, representative governments do not allow the buying and selling of votes even though there could be efficiency gains from allowing this sale, with those who do not value their right to vote exchanging it for a fee. Concerns about democracy rather than efficiency ground the objection to vote selling. Concerns about inequalities also ground objections to leaving the distribution of goods such as schooling or health care to the market, while market distribution of other goods, no matter how unequal, may seem unproblematic. Why? And what follows from this about redistribution? These are questions for moral and political philosophy.

6.3 Market Limits

97

Before turning to issues of ethics, however, it is important to note the contribution that economics makes to these debates about the moral limits of markets, by empirically assessing the costs and benefits of different allocation systems. Such assessment is hard to come by, since it will often involve some factors that are hard to measure as well as counterfactual effects. What, for example, would be the consequences of allowing a widespread market in women’s reproductive labor (“surrogate motherhood”)? It is difficult to determine this, but once determined, it can contribute to ethical debate and discussion. We can see one economic approach to the ways that moral values influence markets in recent work by the Nobel laureate Alvin Roth (2007). Roth argues that economists can treat the fact that people view certain markets as “repugnant” as a constraint on market design, akin to a technological or information constraint. He believes that economists should take this repugnancy constraint into account when studying how to improve certain markets, but should also challenge it: economists have “an important educational role of pointing to inefficiencies and trade-​offs, and costs and benefits [of the persistence of such repugnancy]” (p. 54). Consider the fact that many people view kidney markets as repugnant and most countries prohibit them. The one exception at the present time is Iran. (We should also note that, although illegal, the purchase and sale of kidneys happen anyway, especially in India, where there is a large black market.) At the same time, the number of people on transplant lists in the United States keeps growing; it was more than eighty-​five thousand in 2014. Our repugnance at selling organs is not only inefficient; it costs human lives. Roth’s work seeks to work around our repugnance to kidney markets by trying to design a system of kidney exchange that does not elicit repugnance  –​for example, a system of exchange that does not rely on prices. Most people admire living donors who give one of their kidneys to others; it is only when money or material reward is involved that repugnance is elicited. While it is in principle easy to arrange a one-​for-​one trade (for example, when two women, each of whom needs a kidney, has a donor incompatible with herself, but compatible with the other recipient), arranging a large number of matches poses a logistical problem coordinating hospital surgeries. Notice that working around people’s repugnance demands that economists understand the grounds for it. Presumably in designing a system about which people will feel a lesser repugnance, economists should attempt to respect those grounds that have the strongest ethical rationale. Roth’s kidney matching proposal eliminates the objection to markets in

98

The Ethical Limits to Markets

kidneys grounded in the unfairness of the rich being able to outbid the poor for kidneys. Market designers are now, in the age of the Internet and increased computing power, better able to engineer the platforms governing transactions among market participants. As a consequence, they can control variables such as pricing, liquidity, visibility, information disclosures, terms of trade, and transaction fees. Some examples of market design, associated with the work of Roth and others, include auctions for electricity and other commodities, tradable permit systems for pollution abatement, labor market clearinghouses, centralized systems for the allocation of organs such as kidneys, procedures for student assignment among urban schools, and other related matching and trading processes. Notice, however, that “market design,” like economics in general, cannot itself answer the ethical questions about what sorts of goods should not be traded. Moral analysis is crucial to understand “repugnance” to certain markets. Without having an explicit understanding of which values are relevant to which markets, it is impossible for economists to decide which allocation mechanisms are most desirable (whether the market, regulation, lotteries, rationing, etc.). When do we want to work around a restriction on market allocation, and when do we want to accept it? Should our repugnance to kidney markets be worked around, and if so in what ways? For example, should the state pay people to give up a kidney to a central distribution system, or should kidneys be allocated to the highest bidder? On the other hand, no one thinks that the repugnance to slavery should be treated by finding a “work-​around” to overcome our repugnance: slavery is instead rightly condemned.

6.4  Morals and Markets In the twentieth century, political philosophy had little to say about the nature of particular markets:  most philosophers who have written about markets have defended or criticized the market system as a whole. Egalitarians have concentrated on the role of markets in generating income and wealth inequality; libertarians have been concerned to show that markets protect and enhance freedom and promote wealth. In the past two decades, however, a significant literature has grown up around the topic of whether and why certain specific goods should not be allocated by market mechanisms. Some of the new theories proffer reasons that are consequentialist and contingent, while others draw categorical

6.4  Morals and Markets

99

distinctions between different kinds of goods or between the spheres of social life in which different goods belong. Theories that make categorical distinctions point out that our best understanding of certain goods is incompatible with trading these goods on the market. For example, Michael Sandel (2013) points out that a person who tries to buy a friend would by that very act transmute the nature of friendship:  whatever was bought, it would not be what most people mean by the term “friend.” Or, to use a more political example, consider criminal justice. If defendants could buy judicial decisions in the marketplace, justice would disappear. The market itself seems to depend for its stability on the fact that laws cannot be bought or judges bribed. (In fact, to our common shame, in reality, justice can to some extent be bought: when charged with the same crime, those who are poor are more likely to be convicted and, if convicted, are likely to face harsher punishments.) Proponents of the categorical nature of nonmarket goods have argued that goods such as justice, health care, basic sustenance, sexual and reproductive services, body parts such as kidneys or corneas, and even a clean environment are by their nature corrupted if market mechanisms are used in their allocation. The corruption charge against certain markets is intriguing, and it seems to apply powerfully to certain goods, such as human beings. But there are problems with citing risks of corruption as a general strategy to prohibit markets in certain goods. People disagree about the best understanding of goods. Consider sex. Some people think of sex as a sacred act, tied to marriage, while those who use the Internet to “hook up” for a night view it simply as a casual source of pleasure. How should policy makers determine whose view is better? The fact that views of goods are often controversial threatens another promising view of the moral limits of markets that attempts to tie certain goods closely to “personhood” and argues that goods of this kind should not be turned into commodities. Not only is there considerable disagreement about what goods are involved in personhood, but as one of the advocates of such a view herself notes, banning such markets in order to protect personhood can wind up threatening personhood instead. Prostitution might be objectionable, but banning prostitution can eliminate the best available option for a person and exacerbate her plight (Radin 1996; Nussbaum 1998). Elizabeth Anderson (1993) argues that markets are not appropriate vehicles for goods that should be honored or respected. In her view, certain markets should be limited because valuing certain goods appropriately

100

The Ethical Limits to Markets

rules out treating them as commodities. Thus, it should not be legal to buy and sell parenting rights, because doing so would fail to value children as they ought to be valued. As in the case of the corruption view, the expressive theory of goods powerfully illuminates certain cases. But there is often little direct connection between the way people view particular goods and their sale on the market. One of the great virtues of markets, noted previously, is that they allow people with sharply different views of goods to cooperate with one another. One person might view the Bible as the path to God, and another as a way to generate an income, but they can participate in exchange with each other. Jill can buy a Bible from Jack without accepting his particular understanding of its value. So it is not clear that the expressive view throws light on many of the cases when legislators limit or might want to limit markets. Libertarians argue that forbidding what Robert Nozick (1974, p.  163) refers to as “capitalist acts between consenting adults” interferes with their freedom and insults their dignity. While this may be true in some cases, it is false as a general statement. Nozick’s view overlooks the fact that, in an interdependent social environment, allowing two individuals to engage in a voluntary transaction can have harmful effects on others. Allowing Jill to sell her vote to Jack might be advantageous to each of them, but a large number of similar actions would undermine democracy. Legalizing prostitution could put pressure to turn to prostitution on women who are in need of funds. Allowing some people to work for whatever salary they want (e.g., having no minimum wage) affects the wages that other people are able to command. Since all market exchange presupposes property rights that restrict the freedom of nonowners, the important normative question is not how free individuals are to make whatever deals they please, but which system of entitlements increases net freedom and well-​being. Richard Titmuss (1971) argues that paying individuals for carrying out certain actions undermines their sense of civic duty. In particular, he claims that paying blood donors diminishes their willingness to donate blood, “crowding out” altruistic donors, by undermining altruistic motivations. Although Titmuss does not identify a mechanism that produces this effect, more recent empirical work suggests that the possibility of the crowding out of civic virtue should be taken seriously. For example, Gneezy and Rustichini (2000) report an experiment in which the introduction of a fine for parents who collected their children late from Haifa day care centers increased rather than decreased the rate of late collection (a phenomenon they interpret as being due to the fact that the possibility of paying for late

6.4  Morals and Markets

101

collection legitimizes doing so and reduces the perceived element of social disapproval). In other work, Bruno Frey and coauthors (Frey et al. 1996; Frey and Oberholzer-​Gee 1997) have suggested that willingness of individuals to contribute to public goods may be undermined by monetary payment. In particular, they draw on survey evidence of people’s willingness to accept privately noxious but socially necessary facilities (such as nuclear waste recycling plants). Surveys show that offering compensation does not increase the acceptability of such projects and sometimes decreases it. Where markets crowd out altruistic behavior, they will not necessarily increase efficiency. Perhaps if a market for kidneys were introduced, then the supply would decrease (although this seems unlikely given the low rates of altruistic donation). At the same time, carefully designed markets can sometimes “crowd in” socially responsible behavior. An interesting example is the use of monetary incentives to induce people to carpool. Not only did allowing people to buy access to carpool lanes in San Diego lead to increased revenue for the state, but it was correlated with increased use of the carpool lane by people who drove together and did not buy access (Strahilevitz 2000). At least one hypothesis about the increased number of nonpaying carpoolers is that people now felt good about getting something for free that others (the suckers?) were paying for. The crowding out literature does not provide categorical reasons to refrain from marketing certain types of goods. Instead it provides possible empirical and contingent reasons to be concerned about the effects of markets on civic values and prosocial behavior. And it underscores the importance of having a nonmarket realm to counteract these effects. The rule of law and civil and political equality are allocated on a nonmarket basis; these provide places where people can join together not as traders but as citizens. While some theorists offer what we have been calling categorical arguments that certain goods should not be for sale, others locate the problems with certain kinds of markets in their contingent features. While the textbook market features rational adult traders with full information and a multitude of options, many real world markets are not like this. Consider in this context most child labor markets in the developing world. Child labor markets depart from the idealized textbook market in at least four ways. In the first place, children generally do not offer their own labor on the market. As the economic historian Jane Humphreys (1999) has noted, there is no Infans economicus responding to market signals. Children are put to work by their parents, some of whom might not have the child’s best interests in mind. In the second place, parents in these contexts often lack knowledge of the true costs of taking their children out of school. They may

102

The Ethical Limits to Markets

be motivated only by a short-​term perspective, even though possibly the family would be better off in the long run if children were educated. In the third place, many parents of child laborers are in desperate circumstances, one step away from destitution. In some cases, the family’s survival may depend on the child’s labor. This is quite different from the usual assumption that the participants in a market are able to walk away. Finally, there are significant third-​party effects from the existence of child labor, in addition to the harms suffered by the children themselves. Not only does child labor help produce a group of illiterate adults who cannot contribute to social development or organize for and demand their rights, but it also inhibits the growth of the nation’s economy. One further third-​party effect of child labor is worth noting: when child labor is a widespread practice, this drives down the price of unskilled adult labor. This means that poor families who do not want to put their children to work will have a more difficult time doing so since their own income may be lower than it would have been had child labor been banned.

6.5  Policy Implications: How Should Policy Respond to Morally Problematic Markets? The fact that markets are morally problematic –​either because of the categorical reasons mentioned previously, such as corruption or the undermining of personhood, or because of their harmful consequences for participants or third parties –​does not determine what the best policy response to morally problematic markets should be. It is easy to think that morally problematic markets should be banned, but in some cases, it may be better to tolerate a morally problematic market, such as the market in addictive drugs and alcohol, because the consequences of banning the market are worse. At their best, those who offer categorical reasons against certain markets realize this. Most of those who propose categorical moral objections to various markets leave the policy questions to one side, but they must be faced. In some cases, it may be best to ban the markets even if the costs of doing so are high. Those who argued for the eradication of slavery often recognized that there might be significant costs to doing so, but they believed that social justice or merely minimal decency required all to work toward the immediate end of slavery. In other cases, it might be possible to address the underlying problem that gives rise to the morally problematic market. For example, if the problem that leads to parents’ pulling their children out of school to go to work is the family’s dire poverty, then perhaps social policy should pay parents to keep their children in school. Such a policy, in

6.6  Market Alternatives

103

contrast to banning child labor, does not run the risk of driving child labor into a black market with even worse outcomes for children. If, on the other hand, the problem that leads to so much child labor is poor information about the long-​term benefits of education, then society can direct its efforts to increasing parents’ knowledge of this (Satz, 2010). As these last examples show, in addition to simply permitting markets and banning them, policy can aim to limit or undermine them by, for example, increasing the information that buyers and sellers have about the long-​term consequences of their transactions, or weakening the incentives for the problematic buying and selling. The provision of a generous safety net can be a far more effective response to exploitative labor practices than direct regulation of labor contracts. In the context of thinking about how to respond to morally problematic markets, it is helpful to make use of Hirschman’s (1970) important distinction between “exit” and “voice.” “Exit” occurs when consumers stop buying a firm’s product or when its employees quit their jobs. If enough consumers exit and enough revenue is lost, the firm may modify its product or methods, even in the absence of government regulation. Idealized markets rely on exit to control the behavior of traders. Another way, however, that potential buyers can express their dissatisfaction with a particular product market is by the “voice” option. In this case, consumers directly express their dissatisfaction to the firm. Examples of voice include protests, sit-​ins, and letter writing campaigns. Voice is amplified by the possibility of exit, and exit is made articulate by voice.

6.6  Market Alternatives One final response to morally problematic markets involves creating nonmarket alternatives. Although inadequate, the national kidney allocation system each year provides kidneys for thousands of individuals, who might otherwise seek kidneys on the black market. The inequalities resulting from private health care can be mitigated by government provision or insurance for those who lack resources. Public schools arguably provide for somewhat more equal educational opportunity than market provision of education would achieve. In many cases, allowing multiple forms of provision for a good addresses the problems that might occur if the good were provided in one way alone. While much of the debate in the last two centuries has focused on market versus state provision, there are an array of other options for producing and distributing goods. We have each taught in universities that admit students on the basis of test scores and high-​school performance rather than

104

The Ethical Limits to Markets

ability to pay. Scarce goods are sometimes allocated by lottery, as was the exemption from military service during the Vietnam War; access to tickets to concerts can depend on queuing on a line; friends exchange gifts without worrying about pricing. There are many ways to allocate benefits and burdens, and there is no one way that is best for all goods.

6.7 Conclusions We have argued that moral considerations bear on the acceptability of markets. These considerations provide reasons to regulate and in some cases block certain market exchanges. One worthwhile exercise is to think through how the various approaches we have catalogued in this chapter would apply to two markets that are currently the subject of debate: markets in human kidneys and markets in women’s reproductive labor (as in commercial surrogacy). Whatever you think about these issues, one thing is clear: policy makers and their economic advisers cannot sidestep ethical questions in thinking about markets.

Suggestions for Further Reading A good overview of the issues and the different approaches to the limits of the market is found in Satz (2010). Sandel (2013) and Anderson (1993) discuss the idea that some goods are not appropriately distributed by markets. Walzer (1983) argues that justice requires that goods be distributed according to their social meanings and that maintaining distinct social meanings requires the existence of separate spheres. Radin (1996) criticizes the separate spheres argument and proposes that “human flourishing” should be the touchstone for our decisions about the permissibility of different markets. There is a large literature on the effects of market exchange and in particular monetary payments on motivations. See Frey et  al. (1996) for an overview of the literature. Titmuss (1971) conjectured that markets undermined altruistic behavior with respect to blood donation. Arrow (1972) has a skeptical response.

Questions for Study and Discussion 1. How might economists respond to the idea that the market has “moral” limits? Do markets need moral limits in order to function? What would a society in which everything was for sale be like?

Questions for Study and Discussion

105

2. Do you think that there are moral limits to markets –​why or why not? How can it be morally permissible to prevent competent individuals from engaging in the voluntary exchange of some rights, goods, or services? 3. To what extent might the problems with certain markets that seem problematic be redressed through redistribution? 4. What do you think the most powerful arguments are on behalf of not treating health care as a market good? Do you think those arguments are successful? Why or why not? 5. Do you think there should be a legal market in the following? Why or why not? Would you give the same answer in each case? If not, what explains why your answers to different cases differ? a. Human kidneys b. Sex (prostitution) c. Radar detectors (if their sole purpose is to enable drivers to violate the speed limit with fewer risks of being caught) d. Recreational drugs e. Term papers 6. In Chapter  5, we argued that economic models that treat tax evasion as depending exclusively on the probability of being caught and the severity of the penalties contain the implicit message that moral commitments are not relevant to explaining an individual’s behavior. Does the assumption that markets are always the preferred mechanism for allocating benefits and burdens in a society have similar moral implications? 7. Do you think that, on the whole, living in societies where most goods and services are bought and sold undermines morality or strengthens it? Or does living in a market society instead change the nature of society? 8. One notorious historical market, which helped to provoke the Protestant Reformation, was the sale of “indulgences” by the medieval Catholic Church. By good works or by cash payment, individuals were able to purchase a reduction in the time that, owing to their sins, they would have to spend in purgatory. What, if anything, is objectionable about this market? 9. Sometimes markets have limits simply because they do not “work,” while in other cases markets would function efficiently; yet there is a social choice not to permit them to function. Give examples of cases of these two kinds and explain why the distinction is important.

106

The Ethical Limits to Markets

10. What do you think of the proposal to expand the use of markets by finding ways to lessen people’s repugnance to them? How could we lessen our repugnance to buying and selling kidneys or rearing rights? Could we lessen our repugnance to slavery? Should we? 11. Give examples of markets that corrupt certain relationships or goods and markets that do not do so. What explains why some markets appear to corrupt goods and relationships and why some markets do not do so? 12. Until recently, most children in most societies worked. Affluent societies now have the luxury of allowing children not to work. But is it justified to prohibit child labor, even if the parents believe that their children ought to work or their families need the income their children provide? 13. Banning certain markets has proven to be costly or of limited effectiveness. (Consider, for example, Prohibition or attempts to ban prostitution or recreational drugs.) Should attempts to limit these markets be abandoned? What alternatives are there?

PART T WO

WELFARE AND CONSEQUENCES

Economic outcomes may be better or worse along several dimensions. They may make people better off or worse off, enhance or restrict freedoms, protect or threaten rights, and they may be just or unjust. All of these dimensions are important in evaluating outcomes. As illustrated in the World Bank memorandum, economists generally evaluate outcomes by their contribution to individual welfare. Other things being equal, one outcome A is better than another outcome B if and only if A makes people better off than B. Economists also evaluate institutions and policies, as well as outcomes, in terms of their welfare consequences for individuals. Economists thus focus on only one of the many evaluative questions they could ask about economic institutions, policies, and outcomes:  “How well do they promote welfare?” Since the assessment of outcomes, institutions, and policies rests exclusively on their consequences for individual welfare, the theory of individual welfare is crucial to normative economics. Indeed, normative economics is often called “welfare economics.” Part II focuses largely on welfarism –​the evaluation of outcomes, institutions, actions, and policies in terms of their effect on individual welfare (which we also refer to as well-​being). Chapter 7 discusses consequentialism (which is concerned with outcomes but need not be welfarist), with a special emphasis on one consequentialist theory: utilitarianism. Utilitarianism claims that what is right to do is what maximizes total or average welfare. In focusing assessment entirely on consequences for individual well-​being, utilitarianism resembles welfare economics; but unlike welfare economics, utilitarianism requires a notion of well-​being that can be compared across persons. Otherwise, there would be no way to add up people’s well-​being. Utilitarianism preceded and influenced the development of contemporary economics. It purports to be a complete moral theory –​assessing 107

108

Welfare and Consequences

individual actions and characters as well as economic policies, institutions, and outcomes –​and it is not necessarily wedded to any specific theory of what constitutes a person’s well-​being. The notion of well-​being is the topic of Chapter 8. Chapter 9 concludes Part II with a discussion of standard welfare economics and the way it attempts to capture a part of utilitarianism while rejecting interpersonal welfare comparisons.

SEVEN

Utilitarianism, Consequentialism, and Justice

Consequentialism is the doctrine that one should judge things morally by their intrinsic value and by the value of their consequences. It specifies a particular structure for ethics. In a consequentialist framework one must first decide what is intrinsically valuable. Questions of intrinsic value are not necessarily the most important moral questions, but they must be answered first because everything else depends on their answers. Then one assesses actions, policies, and institutions in terms of their “results” –​that is, their own value and the value of their consequences. Welfare economics presupposes a consequentialist moral theory in which only welfare has intrinsic value. Consequentialist assessment is always comparative: an action, policy, or institution is morally right or permissible if its net results are no worse than the results of any alternative. If the results of a particular policy are better than those of any alternative, then the policy is morally obligatory. Whether a policy or action is right or wrong depends on both what state of affairs will obtain if it is implemented and what the state of affairs will be if any feasible alternative is implemented. Nothing is absolutely impermissible. The right action may have terrible consequences when the consequences of all of the alternatives are even worse, and even a policy with terrific results is impermissible if there is a better alternative. A utilitarian is a consequentialist who says that what is intrinsically good is individual “welfare” or “well-​being.” Is individual welfare a mental state like happiness, the satisfaction of actual preferences, the satisfaction of “rational” or “informed” preferences, or something else not tied to preferences or mental states? Different utilitarian theories disagree about what well-​being is. The fundamental principle of utilitarianism is that one should do whatever maximizes overall welfare, taking everyone into account. This 109

110

Utilitarianism, Consequentialism, and Justice

formulation masks a disagreement between those utilitarians who favor the maximization of total welfare and those who instead support maximizing average welfare. However, both average and total utilitarians agree that the distribution of welfare does not matter: all that matters is its total or average amount. Adding up or averaging well-​being requires measuring the well-​being of different individuals on the same scale. One must be able to compare the increases or decreases in one person’s well-​being to the increases or decreases in the well-​being of others. Notice that utilitarianism is not a selfish doctrine. Morality demands that one maximize total or average welfare, not one’s own welfare. Indeed, an important objection to utilitarianism is that it may require that people make impossibly large sacrifices to benefit others (see Murphy 2000). How does utilitarianism work? Consider how a utilitarian would approach the question of whether there should be a significant estate tax. Since the entire focus of utilitarianism is on the consequences of policies for well-​being, questions such as whether people have a fundamental right to dispose of their wealth as they wish are irrelevant except insofar as they contribute to well-​being. So too are questions of moral desert or the weight of past promises. The only relevant question is whether imposing an estate tax (at some particular tax rate) results in more total or average welfare than the alternatives. So a utilitarian argument for or against a 50 percent tax on estates greater than $2,000,000 will depend on factual issues such as whether the tax enhances or constrains economic growth and the solvency of government. Everyone’s welfare counts in this assessment, including both those upon whom an estate tax has a direct effect and the much larger number who are affected only indirectly. In the terms of what Mill calls “Bentham’s dictum,” “Everybody to count for one, nobody for more than one” (Mill 1863, ch. 5).

7.1  Clarifying Utilitarianism This sketch of utilitarianism leaves many questions unanswered. First, and most crucially, what exactly is welfare? Individual “welfare” has many different meanings. Consider: are happiness and the satisfaction of preferences the same thing? Satisfying an agent’s preferences does not always make the agent happier. (Think of those who seek happiness in a bottle of whiskey.) The major nineteenth-​century utilitarians (especially Bentham, Mill, and Sidgwick) took utility to be a mental state like happiness or pleasure (or, more precisely, to be that property of objects that causes such mental states;

7.1  Clarifying Utilitarianism

111

see Broome 1991a), but few contemporary moral philosophers agree. Most contemporary utilitarians take welfare to be the satisfaction of rational and informed preferences. Chapter 8 will explore views about what constitutes individual welfare. Second, the consequences of actions, policies, and institutions are usually uncertain. When legislators instituted a policy believing mistakenly that its consequences would be better than any of the alternatives, did they do the right thing because the expected consequences were better than the alternatives, or did they do the wrong thing because the consequences of some alternative would have been better? It is awkward to take the first alternative and judge that the policy was “right from their ex ante perspective” but “wrong from an ex post perspective.” It makes more sense to judge that according to the utilitarian standard the legislators made the wrong choice –​no matter how rational, blameless, or even noble they were in making the mistake. In other words, what matter to the utilitarian assessment of an action or policy are its consequences, full stop. When we say to someone who blamelessly made a bad choice, “You did the right thing,” we mean only that, given what the agent knew, she chose as well as she could –​not that there was no alternative that would have been morally better. Although utilitarian appraisals of proposed actions and policies must be made ex ante and thus always depend on beliefs about what their consequences will be, whether the actions or policies are right depends on what the consequences turn out in fact to be. Third, whose welfare counts? Should an agent consider only the welfare of currently living human beings? What about the welfare of those not yet born –​and of those who might, as the result of one policy or another, not even be conceived? (See Parfit 1984, ch. 17.) These issues are particularly pressing with respect to the problem of climate change, because the policies adopted now are likely to have major effects on future generations. Climate change raises questions not only about how to take account of the well-​being of individuals whose very existence may depend on our policies, but also concerning the fairness of our treatment of them. (These issues concerning the interests of future generations are as worrying for standard welfare economics as they are for utilitarianism.) In addition to deciding which human interests count, there is also the question of whether evaluations of alternative policies should consider the welfare –​or at least the pains and pleasures –​of nonhuman animals. Most people believe that it is morally wrong to inflict suffering on animals “needlessly.” Commercial farming of animals for their meat and hides raises moral questions, particularly when animals are confined in tight cages,

112

Utilitarianism, Consequentialism, and Justice

frightened, and badly treated. If utilitarians count the pleasures and pains of all sentient beings, then they can explain quite naturally why the treatment of animals matters morally (Singer 1975). But utilitarians face difficult problems comparing human and nonhuman welfare (how many frogs is a human life worth?) as well as measuring animal welfare. (What is the value to a chicken of being able to roam freely?) Beyond nonhuman animals, there are questions about how to capture the value of trees, rivers, or natural beauty: are these only valuable because of the pleasure they give to sentient beings capable of pain and pleasure? Fourth, should utilitarians be concerned with total welfare or average welfare? Since average welfare is total welfare divided by the size of the population, total and average utilitarianism will coincide when population size is fixed. But if alternative policies have consequences for how many people there will be, then what maximizes total welfare may differ from what maximizes average welfare. Average utilitarianism seems initially more appealing. A  small population whose average well-​being is high seems more attractive than a much larger population with low levels of individual well-​ being, which has more total well-​being. At the same time, how can it be wrong (as average utilitarianism implies) to do something that adds to total well-​being, such as having a child whose life will be happy, although less happy than the average? These theoretical issues bear directly on pressing problems concerning population control (Parfit 1984, pt. 4; Broome 2004). Fifth, how should utilitarianism guide individual and social decision making? Actions and policies are morally obligatory if they maximize utility, but it does not follow that the best way to make moral decisions is to attempt to calculate the welfare consequences of different actions and policies. Not only is calculating the consequences of actions on the welfare of the whole present (and future?) human (and animal?) population of the world fraught with uncertainties (and not much fun), it is also likely to introduce lots of bias because people tend to take a more favorable view of the overall consequences of actions that benefit them personally. Calculation is a time-​ consuming action, and not one that is likely to maximize utility. A person does not usually become happy by directly aiming to maximize his happiness; nor is he likely to do right by directly aiming to maximize total happiness. Furthermore, uncoordinated actions by numerous individuals may lead to bad aggregate consequences that could be prevented by requiring people instead to follow simple rules. People are more likely to perform actions that actually maximize welfare if they do not calculate the welfare consequences of their actions and instead act on general rules, such as “tell the truth,” “keep your promises,” and so forth.

7.1  Clarifying Utilitarianism

113

Policy makers may be in a better position to investigate the consequences of alternatives, but they are also often better advised to adhere to rules, for it is difficult to tell what the consequences of policies will be and what impact they will have on total welfare. Consider, for example, the Bush and Obama administration policies of imprisoning “enemy combatants” indefinitely without any right to an attorney. Since the consequences of a policy such as this one are highly uncertain and could be disastrous, there is a utilitarian case for treating it as off-​limits –​which is precisely the point of constitutional and international law. At this point it might appear that utilitarianism is pulling a disappearing trick, having taken its bow. It seems that the utilitarian is now ceding the stage to defenders of traditional moral principles. But even though the best way for individuals to maximize utility is usually to stick to traditional moral rules, utilitarianism does not endorse all of traditional morality. It seems plausible that rules against lying maximize utility, but it is not obvious, for example, that utilitarianism supports some of traditional morality’s objections to expanding the scope of markets to permit paid adoptions (“baby selling”) or buying and selling of organs for transplantation. People’s sentiments do not always guide them well, and it is crucial that some people think carefully about what sorts of policies will do the most good and advise the rest of us. For example, defenders of “effective altruism” have argued that there are good ways to amplify the impact of charity (MacAskill 2015; Singer 2015). In pointing out that people generally do better following established moral practices rather than attempting to estimate the aggregate welfare consequences of alternative actions, it may appear that we are endorsing “rule utilitarianism” (Hooker 2001; Mason 1998), but this is not the case. According to rule utilitarianism, actions are right if and only if they conform to a set of rules such that the consequences of the general adoption of these rules for total or average welfare are no worse than the consequences of the general adoption of any other set of rules. Hence rule utilitarianism maintains that people ought to adhere to the utility-​maximizing rules even in the unusual case in which they are confident that breaking a rule would result in more total welfare. If the point of morality is to maximize welfare, then in these unusual circumstances, individuals should violate the rules.1 The sixth question concerns interpersonal comparisons of well-​being and deserves a section to itself. Rule utilitarianism may, however, have other justifications. See Hooker (2001, ch. 4) and Parfit (2011). 1

114

Utilitarianism, Consequentialism, and Justice

7.2  Interpersonal Comparisons of Well-​Being If policy A benefits Ira and harms Jill while policy B benefits Jill and harms Ira, then there is no way for policy makers to judge which results in greater welfare unless they can compare how much Ira and Jill are each benefited and harmed by A and B. Since the benefits and burdens of alternative policies typically fall on different individuals, there is no way to compare the total or average welfare resulting from alternative policies without making interpersonal comparisons. Many nonutilitarian ethical systems also require interpersonal comparisons of well-​being. To be rationally benevolent, one must be able to judge where one’s efforts will do the most good. To treat the interests of different people equally, one must be able to compare the net effects of one’s actions on the interests of each. Interpersonal comparisons are also crucial to a classic economic argument supporting the redistribution of wealth and income. Although utilitarians are indifferent to the distribution of welfare, they are not indifferent to the distribution of income, because the amount of welfare that results from a given amount of income depends on how it is distributed. Economists writing near the beginning of the last century, such as A.  C. Pigou (1920), argued that total welfare is maximized by equalizing incomes as much as is consistent with retaining incentives to produce. Citing the diminishing marginal utility of income, they maintained that, for example, an extra thousand dollars contributes less to the well-​being of someone with an income of fifty thousand dollars than to the well-​being of someone with an income of ten thousand dollars. Other things being equal, then, a more equal distribution of income increases total welfare. This argument assumes that one can make interpersonal comparisons of the amount that a thousand dollars contributes to the well-​being of different people with different incomes. If interpersonal comparisons cannot be made, then this argument cannot be made either. So-​called unit comparisons of utility differences are needed to compare the benefits and harms policies might cause to different people. One need only compare how individual utilities change, not their absolute levels. Comparisons of utility levels are needed if, for example, policy makers want to know who is worst off. There are obvious difficulties in comparing how well off different people are. The main way economists and philosophers have attempted to understand interpersonal comparisons is via what Kenneth Arrow (1978) calls “judgments of extended sympathy” (see also Kolm 1972). Suppose we ask people to express preferences not only among ordinary alternatives but also among “extended” alternatives, such as the alternative of being Ira with

7.2  Interpersonal Comparisons of Well-Being

115

some option x and being Jill with y. We might then say that Jill-​with-​y is better off than Ira-​with-​x if people prefer to be Jill-​with-​y to being Ira-​with-​x. This way of understanding interpersonal comparisons faces serious problems. It passes the buck, since the extended preferences agents have would seem to depend on interpersonal comparisons. Moreover, if people disagree about the ranking of being Ira with x and being Jill with y, whose extended preferences should decide? And more precisely, what exactly is the connection between, on the one hand, whether Ira is better off with x than Jill is with y and, on the other, McPherson’s or Satz’s or Hausman’s preferences between the extended alternatives? John Harsanyi (1977b) suggests that, rather than employing one’s own preferences to compare Jill-​with-​y and Ira-​with-​x, one should compare how well off one would be with x if one had Ira’s preferences to how well off one would be with y if one had Jill’s preferences. (How this is supposed to be easier than comparing how well off Ira is with x and Jill is with y is not obvious.) In Harsanyi’s view, there is a single impersonal extended preference ranking to which our empathic abilities are a useful although imperfect guide. Judgments resulting from putting yourself in someone else’s shoes in this way can be used to construct interpersonal comparisons. Alfred MacKay (1986) calls this the “mental shoehorn” tactic (see also Griffin 1986, ch. 7). It is doubtful that extended preferences can provide a basis for interpersonal comparisons. Even if these preferences were unanimous, as is often not the case, they answer the wrong question. Everyone might prefer to be Jill with y –​even though Ira is better off –​simply because they admire Jill more than Ira. One might, for example, prefer to be Keats (who died of consumption at age twenty-​five) rather than Queen Victoria (who lived to eighty-​one and occupied the throne for sixty-​three years) yet nevertheless believe that Victoria enjoyed a higher level of well-​being than Keats. Most economists take the problems of making interpersonal comparisons as a decisive reason to reject utilitarianism. But, however problematic in theory, interpersonal comparisons seem entirely feasible in practice: people make them all the time. For example, as a rough approximation, policy makers can suppose that those whose situations are similar in terms of income, health, and personal relations are equally well off, or that those who have a very low level of income, health, and personal relations are worse off than those who have a very high level of income, health, and personal relations. Utilitarianism remains an extremely influential ethical view, sometimes named explicitly and sometimes not, particularly when one is concerned with issues of public policy, because it makes ethical questions in principle matters of straightforward calculation. Such calculations may be difficult to

116

Utilitarianism, Consequentialism, and Justice

carry out, and in some cases, utilitarianism will not give any definite advice owing to problems in learning the effects of policies and in measuring their welfare consequences. But there will be clear cases, and the reasons for indecision and disagreement can be stated clearly.

7.3  Justifying Utilitarianism Utilitarianism is a tempting ethical theory. In many cases it matches common intuitions; and it is plausible to think that morality is centrally concerned with human well-​being and that the consequences of our actions matter to their assessment. Furthermore, utlitarianism offers guidance in those cases where people’s intuitions conflict: it shows how even difficult moral questions can be decided. Which policy or action a person ought to adopt depends on the consequences of alternative policies. Utilitarians can even cite the imperfections of human knowledge of consequences to explain why ethical questions are so hard to answer. If actual ethical systems are to a considerable extent implicitly utilitarian, then the utilitarian can offer a plausible explanation for why moral codes differ in different societies. Leaving one’s grandparents out to die in the cold may have been morally permissible among those living in the harsh conditions of the Arctic; it may have maximized welfare in those conditions. The same policy is morally impermissible in affluent societies, where it does not maximize welfare. According to utilitarianism, what is morally right or wrong depends on the consequences, which in turn depend on the facts of the case. Utilitarianism is “absolutist” in one sense –​whatever action maximizes welfare is the morally right action to perform –​but it is not absolutist in the sense of supporting inflexible formulas for conduct. Even something as apparently heinous as intentionally killing civilians in wartime might in some circumstances produce more welfare than refraining from doing so and thus turn out to be morally justified. Ultimate questions of justification are very difficult, but utilitarians have two ways to defend their doctrine. The first justification rests on the claim that welfare or well-​being is the only intrinsically good thing. If the basic ethical obligation is to promote good states of affairs, then utilitarianism will follow; this is the justification for utilitarianism offered by Mill and Sidgwick. A second justification relies on the notion of equal respect (Griffin 1986, ch. 9). Interpreting equal respect as giving equal weight to everyone’s interests leads naturally to utilitarianism. The defense of utilitarianism in terms of equal respect can alternatively invoke the idea of a hypothetical agreement on the part of rational agents

7.4  Contemporary Consequentialism

117

concerned to advance their interests. From an impartial point of view, one can argue that these agents will endorse whatever moral principle serves individual interests impartially. John Harsanyi argues that that moral principle is average utilitarianism (1955). Suppose one models an impartial moral perspective as the perspective of an agent who thinks that it is equally probable that he or she could be any of the n members of society. The expected well-​being of such an agent will be (1/​n)U1 + (1/​n)U2 + … + (1/​n)Un, where Ui is the well-​being of the i’th member of society. (1/​n)U1 + (1/​n)U2 + … + (1/​n)Un is, of course, the average welfare in the society. So, Harsanyi argues, impartial agents concerned to advance their interests –​that is, to maximize their expected welfare –​will endorse average utilitarianism. These two justifications point toward differing interpretations of utilitarianism. The first argument  –​in which utilitarianism derives from the attraction of maximizing the good –​leads naturally to maximizing the total utility of all sentient beings (whether human or not), whereas equal respect or contractualist arguments point toward maximizing the average utility only of those sentient beings who are rational. As noted in Section 7.2, these two outlooks have sharply different implications with regard to population policy.

7.4  Contemporary Consequentialism During the decades that preceded the 1970s, utilitarianism appeared to be almost dead. Although it continued to influence policy makers, most philosophers did not take it seriously as a moral philosophy. Most economists had earlier abandoned utilitarianism in the face of the difficulties posed by interpersonal utility comparisons. And to pound the final nails in the proverbial coffin, John Rawls, in A Theory of Justice (1971), offered systematic criticism as well as an alternative theory that was suitable for guiding policy. The resurgence in practical moral philosophizing that was so prominent in the 1970s usually took for granted some sort of reciprocity or rights perspective rather than any sort of consequentialism. Yet by the end of the 1980s utilitarianism and consequentialism were again highly influential in both theoretical and applied moral philosophy. The consequentialists of that era defend very different ethical theories, with James Griffin, for example, developing a sophisticated objective-​list variant of utilitarianism and Amartya Sen developing a nonutilitarian view of consequential evaluation in which rights, capabilities, and functionings play a more central part than well-​being. For instance, Sen (1979) suggests

118

Utilitarianism, Consequentialism, and Justice

that consequentialists can regard rights violations as themselves bad consequences in addition to any welfare losses that may accompany them. This work shares many of the features of modern economic theory and in many cases shows the influence of economic modeling. Consider that most of these consequentialists link ethics to the theory of rationality. John Harsanyi, for example, writes, “Ethics … is a theory of rational behaviour in the service of the common interests of society as a whole” (1977a, p. 43). Most moral theorists do not go this far; but, as Samuel Scheffler (1982) especially has stressed, refusing to make trade-​offs among different objectives and to maximize some objective opens one to charges of irrationality. Consequentialist theorizing –​with its close association between rationality and ethics  –​is thus more congenial to economists and more easily integrated into normative economics than are, for example, rights-​based ethical views that are not easily opened to calculation and trade-​offs. The past generation of utilitarians and consequentialists has been influenced by developments in economics and game theory. For example, Matthew Adler’s Well-​Being and Fair Distribution (2012) owes as much to economists’ discussions of social welfare functions as it does to philosophical discussions of justice. In Russell Hardin’s utilitarianism, human ignorance of consequences and the difficulties of measuring and comparing utilities hold center stage, and concepts from game theory are put to work generating the outlines of a theory of property rights and its limits (Hardin 1988). Of particular interest has been the development of consequentialist moral theories in which the valuable consequences to be maximized are things other than well-​being (see especially Parfit 1984). One important value in many of these theories is the satisfaction of needs; even utilitarians emphasize it, though not as a fundamental and intrinsic good. For example, although James Griffin is sympathetic to “informed preference” utilitarianism, he argues that policy should focus on needs because a government can more easily determine what people need than what will satisfy their informed preferences (1986, ch. 3). Griffin’s emphasis on the empirical tractability of needs is ironic, given how averse economists have been to distinguishing needs from mere preferences. This aversion has arisen not from empirical difficulties but rather from theoretical objections to drawing the distinction. (Here is a case when moral philosophers may be more practical than economists!) In political discussions of economic policy, concern about human needs is already ubiquitous, and if philosophers can provide both a rationale for taking needs seriously in social decision making and a principled way of drawing the distinction between needs and wants

7.5  Is Utilitarianism Plausible?

119

(Braybrooke 1987; Thomson 1987; Dasgupta 1995; Sen and Nussbaum 1993; Nussbaum 2000), then economists can put their modeling tools to work to help devise policies that will satisfy needs. A good example lies in the area of medical research, where difficult choices must always be made about how much to invest in searching for treatments for various diseases. Rational allocation of scarce resources across different fields of research requires assessing the costs of research, the probabilities of success, and the relative urgency of progress in different fields. It is plausible that philosophers can contribute to making such choices well by developing defensible judgments about the value of health and the relative importance of the needs that would be met by successful treatment of different diseases. The deep problems of utilitarianism  –​in particular, those concerning interpersonal comparisons of well-​being  –​do not preclude operationalizing utilitarianism by providing a specification of the utility function to be maximized. One could, for example, stipulate a single utility function that roughly represents everyone’s preferences. Preferences represented by such a common utility function would most plausibly be defined not over all marketed goods and services but instead over fundamental goods such as nutrition, shelter, or clothing that the agent extracts or constructs from marketed goods and services such as apricots, apartments, and aprons (Becker 1976; Kolm 2002). Behind the large differences in people’s manifest preferences there might be agreement in preferences among the underlying goods. For example, Jill may prefer to eat at home while Jack prefers restaurants. The differences in their preferences may result more from differences in the opportunity costs they face in producing good meals at home than from any differences in underlying preferences. It is not absurd to postulate (as an approximation) a common utility function, and in terms of that function there might then be little problem determining the total utility of alternative policies –​apart from the general difficulties of predicting their consequences. But we are in no position to judge how useful the idea of a shared “deep” utility function underlying apparent differences in preferences may be.

7.5  Is Utilitarianism Plausible? The most powerful objection to utilitarianism is that it clashes sharply with many of our moral intuitions. What if a hereditary caste society results in more total happiness than a liberal democracy? A  utilitarian would have to opt for the hereditary caste society. Or suppose that false testimony by

120

Utilitarianism, Consequentialism, and Justice

witnesses to the recent killings in the United States of African Americans by police would have led to indictments and convictions and thereby have prevented some of the severe rioting these incidents have provoked. It would seem that a utilitarian should favor perjury in cases such as these. Even if the convictions were unjust, the overall consequences of perjury would appear to be far better than the consequences of telling the truth. Yet most people believe that perjury is morally impermissible. There are four possible responses to this apparent conflict between utilitarianism and what most people believe is right and wrong. First, utilitarians can argue that the conflict derives from a misapplication of utilitarianism. As John Rawls insisted in his “Two Concepts of Rules” (1955), one should distinguish questions about the design of institutions from questions about enforcement of their norms. Even if perjury would in these cases maximize utility, utilitarians would not favor laws and customs permitting perjury whenever witnesses conscientiously believe that perjuring themselves would maximize utility. So law and custom should condemn perjury, even in a case when committing perjury might maximize utility. This is not an argument for rule utilitarianism. It establishes only the weaker conclusion that perjury should be illegal and socially proscribed. It might still be morally obligatory for witnesses to break the law if they can do so secretly. Second, utilitarians may object that cases such as the last one presuppose knowledge that is unattainable (Hardin 1988). When one takes into account the unavoidable uncertainties, it may turn out that utilitarianism does not in fact recommend perjury. This way of reconciling utilitarianism and moral intuition is not entirely satisfactory, because many people would say that whether perjury is morally right or wrong in this case does not depend on whether it would be detected or on whether it would ensure a conviction. A third response to the apparent conflict between utilitarianism and moral intuition is to challenge the authority of “intuition.” A  utilitarian favors educating people to have strong and unambiguous moral convictions that promote desirable conduct in typical situations. It will be good on the whole if people are strongly moved by such feelings, but their considered moral judgments have no independent evidential force that can help in resolving hard cases (see Hare 1981). In reply, the critic of utilitarianism can question what basis for morality there could be apart from intuition and our considered reflection on it. If people cannot take the intuitions they retain in the face of serious reflection seriously (though not uncritically), then they have no foundation upon which to argue for or against moral principles. So one arrives at a fourth

7.6  Consequentialism and Deontology

121

response to the apparent conflict between intuition and utilitarianism, which is to reject utilitarianism.

7.6  Consequentialism and Deontology It might be thought that the general structure of evaluation we have called consequentialism is not subject to similar intuitive objections. Because consequentialism does not specify what counts as good, one might think that all ethical theories can be regarded as consequentialist. For example, those moralities that stress freedom rather than welfare might be regarded in consequentialist terms as seeking to maximize freedom, while those that stress duties might mandate maximizing conformity with duty. But this defense of consequentialism misunderstands the structure of many moral theories, which do not solely aim at maximizing some final value, such as freedom or duty. Many moral theories also give to people another end for the sake of which the final goal, whether it be freedom or even welfare, is pursued: the persons about whom the agent cares (Anderson 1993). Even if lying to a person would promote his freedom, it is unacceptable to do so on these theories because it is a disrespectful way to treat him. In Scheffler’s terminology, deontological (nonconsequentialist) ethical theories employ both “agent-​centered prerogatives” (they sometimes permit agents to act in a way that does not maximize the good) and “agent-​ centered constraints” (they sometimes prohibit agents from acting so as to maximize the good). Agent-​centered prerogatives and constraints are puzzling. How can it be morally permissible, let alone morally obligatory, to choose the lesser good? Deontological theories not only conflict with consequentialism, they appear to conflict with rationality itself. Consequentialists might claim that if a person is serious about regarding killing as wrong and if she is convinced that murdering one innocent person will prevent the murder of two others, then it is only irrational squeamishness rather than moral principle that prevents her from committing the single murder. But refusing to murder an innocent person does not appear to be mere squeamishness. Acknowledging this, the consequentialist might argue that cases in which one can, with certainty, prevent two murders by committing one almost never occur. Individuals are thus better off following rules that prevent the killing of the innocent. Conflicts between everyday moral principles and maximizing overall welfare are real and not uncommon. Consider, for example, the case of Baby Jessica, a toddler who in 1987 fell into a well in Texas. After a massive human effort costing hundreds of thousands of dollars, she was rescued.

122

Utilitarianism, Consequentialism, and Justice

The resources that were used to rescue her could have prevented the death and injury of hundreds of other American children if they had been devoted instead to better prenatal care. They could have saved the lives of thousands of malnourished infants in less developed countries. The distribution of health-​care resources is shot through with similar examples. Consequentialism implies that resources should go where they do the most good. Nonconsequentialists would counter that we owe more to someone at immediate risk of death than we do to those who face small risks, even if failing to attend to the small risks leads to a larger number of deaths (Kamm 1993). It is worth underscoring that deontological constraints need not be absolute. Whereas it might be morally wrong to kill one person to save two lives, it might be morally permissible to shoot down a hijacked airliner heading for a nuclear power plant that would kill hundreds of thousands. Yet, even with such a qualification, the consequentialist rejects deontological constraints as irrational. In theory, it is possible to bring about a reconciliation between rationality and absolutist or deontological restrictions on intentional killing by distinguishing between x being murdered by someone and my murdering x. If that is the case, then there is no inconsistency in my preferring two others being murdered by some other person to my murdering x. Similarly, there is no inconsistency in preferring the death of a single child, Baby Jessica, to the death of hundreds yet preferring the death of hundreds of children to my standing by and letting Baby Jessica die. In order to make such seeming reconciliations of these two different moral perspectives more than ad hoc gimmicks, defenders of consequentialism must say something about why it is rational to make these distinctions and put so much weight on them. Here again we can see that utility theory presupposes a substantive background theory of “rational requirements of indifference” specifying when it is rationally permissible to distinguish among outcomes (cf. Section 4.3). But there is a new wrinkle. In this case, the background theory, which here determines when it is rational to distinguish among actions and outcomes, depends on substantive moral principles. One proposed reconciliation allows the assessment of the goodness or badness of a set of consequences to vary with the perspective of the person doing the evaluating (Sen 1982b). An onlooker might conclude that it is morally better for one person to be murdered rather than for two to be murdered or for one child to perish rather than hundreds. But from the standpoint of the decision maker who is the acting party, what is at stake is whether I commit a murder or whether I instead stand by and let two children die by other processes. Everyone might agree both that two murders

Suggestions for Further Reading

123

are worse than one and that, from the point of view of an agent facing the prospect of committing a murder, carrying out the murder is morally worse than failing to prevent two murders from happening. It might seem that evaluator relativity is less likely to be a consideration in public policy than in personal morality because policy should be made from an impersonal point of view. But as the Baby Jessica case and other examples from health-​care policy show, it is not obvious that policy should attempt simply to maximize overall well-​being. If Americans were consequentialists and did not rescue the likes of Baby Jessica –​but also did not look the other way when hundreds of thousands of other children needlessly died –​then the United States would in important ways be a better society. But if Americans could ignore Baby Jessica’s plight with no more than a self-​congratulatory recognition of how rational they were, who knows how inhumane they might become?

7.7  Conclusion: Should Economists Embrace Utilitarianism? Welfare economists will find some versions of utilitarianism attractive because, apart from relying on interpersonal utility comparisons, they are so similar to standard welfare economics. Just define the rough-​and-​ ready utility functions that will represent the preferences of the individuals affected and stipulate a way of making the interpersonal comparisons, and the way to utilitarian policy analysis is open. Welfare economists would not have to give up their focus on outcomes or their view of welfare. Of course, it is a bold and controversial step to stipulate what the utility functions should be and how the utilities of different people should be compared, but this step is no bolder than the identification of welfare with the satisfaction of preferences. Taking this further step would enrich normative economics. Yet many of those concerned with policy would take issue with the suggestion that normative economists should become more explicitly utilitarian, precisely because they would like to see economists make a more radical break with current practice. Before considering this critique, let us turn in the next two chapters to a more detailed treatment of current practice.

Suggestions for Further Reading Classic writings on utilitarianism include Bentham (1789), Mill (1863), and Sidgwick (1901). The collection of essays by Plamenatz (1967) provides a number of helpful philosophical treatments of problems of utilitarianism.

124

Utilitarianism, Consequentialism, and Justice

An old but still useful exchange on utilitarianism is Smart and Williams (1973). A helpful volume with emphasis on economic aspects is Sen and Williams (1982). The most important consequentialist works are those of Brandt (1979), Broome (1991b, 2004), Griffin (1986, 1996), Hardin (1988), Hare (1981), Harsanyi (1977a), Kagan (1989, 1997), Parfit (1984, 2011) Railton (1984), Sen (1982b), and Singer (1979). For discussions of rule utilitarianism see Hooker et al. (2000), Hooker (2001), and Parfit (2011). Applications of consequentialist moral philosophy to practical issues abound. See particularly Glover (1990) and Singer (1986, 2002). For discussion of the rationality of deontological moral principles that permit options not to do what maximizes the good or forbid maximizing the good, see Nagel (1986, ch. 9), Scheffler (1982, 1988), Kagan (1989), and Kamm (1993).

Questions for Study and Discussion 1. Utilitarian judgments of outcomes and policies are always comparative. Why? Is the fact that utilitarianism makes no absolute judgments a virtue of the view or an objectionable feature? 2. Do you think that utilitarianism is a reasonable guide to how we should treat nonhuman animals? 3. Do you think that a stronger case can be made for total utilitarianism or average utilitarianism? 4. It is often said that utilitarianism provides a plausible criterion concerning what is right or wrong but not a plausible method of making moral decisions. Why? What is the difference? 5. What do you think of Harsanyi’s argument for utilitarianism in Section 7.3? Are you convinced? Why or why not? 6. Utilitarianism is an objective theory of what is right and wrong. If there is no alternative to P that leads to more total welfare, then P is morally permissible, regardless of what anybody in any society thinks. Yet, as the example of abandoning the elderly to freeze or starve illustrates, utilitarianism is open to the possibility that what is right in one society may be wrong in another. How is this possible? 7. Why are not all moral theories consequentialist? Are nonconsequentialist moral theories irrational?

Questions for Study and Discussion

125

8. In your opinion would it have been better if the resources devoted to saving Baby Jessica had been used instead to save many lives in less developed countries? 9. Suppose some people are much better at converting dollars into satisfactions than others. Maximizing society’s utility would require giving these people disproportionately larger amounts of income than others. Is this fair? Is it right? 10. As we noted in Chapter 1, inequalities in the distribution of wealth and income in the United States are large and growing rapidly. According to Saez and Zucman, “The 16,000 families making up the richest 0.01%, with an average net worth of $371 million, now control 11.2% of total wealth –​back to the 1916 share, which is the highest on record.” Do you think that utilitarians should favor policies to lessen inequalities in income and wealth in the United States today? Why or why not?

EIGHT

Welfare

When people in modern Western cultures think about morality, they tend to focus on what is morally permissible or impermissible, right or wrong. But there are other important matters of moral concern: questions about what is good or bad and, more specifically, about what is good or bad for people. What is good for a particular agent, Scrooge, will depend on factors such as Scrooge’s character, ability, and circumstances, and what is good for Scrooge may differ from what is good for Marley. Most of the differences will likely concern instrumental goods –​things that are good because they are means to something else that is good. If one focuses on intrinsic goods –​ things that are good in themselves, independently of their consequences –​ then there will tend to be less variation from individual to individual. Size 7 shoes are good for Scrooge while size 12 shoes are good for Marley, but both pairs serve the same end of walking. If nothing were good in itself, then nothing could be good as a means to some other end. There must be intrinsic goods in order for there to be instrumental goods. In addition to its concern with right and wrong, moral philosophy offers theories about which things are intrinsically good for human beings. All plausible moral theories assign an important place to conceptions of such a good. This is obviously true of utilitarianism. However, even nonutilitarian views that emphasize rights, fairness, and justice need a conception of human well-​being, if for no other reason than that these theories recognize the importance of benevolence, which requires some view about what makes people better off. Even the core notions of nonutilitarian theories often make reference to well-​being. For example, most theories of justice are concerned with how well individuals’ basic interests are met, and theories of right action often involve avoiding harm to people. Notions of harms and interests are plainly connected to notions of well-​being. 126

8.2  Welfare in Economics

127

8.1  Theories of Well-​Being In chatting with one’s neighbors, as in studying moral philosophy, one finds many different theories of well-​being. According to some religious views, the ultimate good lies in a specific relationship with God, while in others one’s relationship with other humans is first and foremost. Some philosophers believe that only mental states are intrinsically good, but there is less agreement here than it seems because there are so many different views concerning which mental states are intrinsically good. Jeremy Bentham holds that the only intrinsic good is pleasure, while John Stuart Mill holds that it consists of “happiness,” which involves judgments and satisfactions of a higher quality than the pleasures available to nonhuman animals. Mystics find the good in contemplative states of mind. Henry Sidgwick argued for a hybrid view that the good is any mental state (such as happiness, pleasure, hope, or love) that is intrinsically desirable. Other philosophers endorse as intrinsic goods a potpourri of nonmental states ranging from human health and intimate relationships to achievements. The theory of well-​being is a messy area of philosophy, and all the various philosophical theories face serious difficulties. These controversies are enough to send economists running back to their graphs. But economists cannot avoid thinking about well-​being if they want to be able to judge when welfare increases or decreases. Theories of well-​being or welfare can be classified as formal or substantive. A  substantive theory of well-​being explicitly delineates what things are intrinsically good for people. Hedonism is an example of a substantive theory of well-​being; its adherents hold that well-​being is pleasure. Formal theories of well-​being specify how to determine what things are intrinsically good for people, without identifying what those things are. To maintain that welfare is the satisfaction of preferences is to offer a formal theory of well-​being. This theory does not entail anything about what things are good for individuals. A formal theory can be compatible with a substantive theory. For example, if happiness is the ultimate object of preference, then it is both the case that well-​being is the satisfaction of preference and that well-​being is happiness.

8.2  Welfare in Economics Most economists do not directly address questions concerning welfare. Among those economists who do write about welfare, most are attracted to a formal theory, no doubt because formal theories appear to involve less controversial philosophical commitments than substantive theories.

128

Welfare

By leaving the substantive question of what is good up to the individual, normative economists aim to show their philosophical modesty. Measuring well-​being by how well satisfied an agent’s preferences are also appeals to the antipaternalist view that individuals are the best judges of their own well-​being. But, as we shall argue in this chapter, welfare is not best understood in terms of preference satisfaction. Given mainstream economists’ commitments to utility theory in explaining human choices, it is natural that they would look to levels of utility  –​that is, to the extent to which states of affairs conform to an agent’s preferences  –​as the fundamental measure of human well-​ being for evaluative purposes as well. If individuals are exclusively self-​ interested, then they will prefer x to y if and only if they believe that x is better for them than y is. If they are perfectly well informed concerning the facts and are good judges, then it seems that what they prefer will coincide with what is better for them. It follows that if a person, Jill, is a competent evaluator, self-​interested, and well informed, then Jill prefers x to y if and only if x is actually better for her than y. Assuming that individuals are generally rational and competent judges and that they are self-​interested and well informed, economists find it natural to identify welfare and preference satisfaction. In their applied work, economists often rely on more objective measures of “real income” rather than utility measures, but most economists view this as a compromise with data limitations. They regard real income as an imperfect proxy for preference satisfaction. There are two possible explanations for the coincidence of preference satisfaction and welfare. One possibility, which we call “the constitutive view,” is that Jill’s preferring x to y when Jill is self-​interested and Jill’s factual beliefs are correct makes it the case that x is better for her than y. On this interpretation, normative economists are committed to the philosophical thesis that well-​being consists in the satisfaction of well-​informed self-​ interested preferences. Although what people actually prefer may not be good for them because their preferences are not directed toward themselves or because they hold false beliefs, actual preferences will typically coincide with well-​informed self-​interested preferences and will determine what is good for individuals. We call the second explanation of why the satisfaction of self-​interested and well-​informed preferences is a guide to well-​being “the evidential view.” It is much more modest philosophically. If economists assume that agents are competent evaluators –​that is, that when their beliefs about the consequences and properties of alternatives are true, they are generally good judges

8.2  Welfare in Economics

129

of what is good for them –​then when Jill’s beliefs are true, Jill’s judgment that x is better for her than is y is reliable evidence that x is in fact better for her than is y. On the evidential view, economists need not accept any substantive theory of well-​being. Nevertheless, in order for economists to figure out when the evidence is flawed and when it is accurate, the notion of well-​being cannot be completely empty. If economists suppose that people are generally self-​interested and evaluatively competent, how do they determine which are the contexts in which this general supposition is unjustified? We think that platitudes that are part of the everyday view of well-​being provide some guidance. In those contexts in which people prefer the sorts of things that we think are generally better for people, such as greater health, closer friends, greater financial security, and so forth, economists have grounds to believe that people are more or less self-​interested and competent evaluators. In contexts where people’s preferences do not generally conform to platitudes such as these, one should be cautious of drawing any inferences concerning well-​being from people’s preferences. There is no need to endorse any special philosophical theory in addition. Both the constitutive view and the evidential view justify measuring well-​being by the satisfaction of preferences, although both recognize that what people prefer and what is good for them will differ when their preferences are not self-​interested or when they depend on false beliefs. Welfare economists seldom mention these two important qualifications, and in practice economists sometimes operate with the mistaken view that everyone is self-​interested. But, of course, people are not always self-​interested. Only sociopaths are unable to distinguish what they believe to be best, all things considered, and what they believe to be best only for themselves. For most people, the two do not always coincide. Someone who is taking steps to prevent environmental disaster two centuries from now is not satisfying her self-​interested preferences. Nor do people always have good information. Satisfying people’s false beliefs about the properties and consequences of alternatives is not necessarily best for them. Drinking poisoned water that one believes to be safe does not make one better off. There are circumstances in which it is a reasonable approximation to suppose that people are competent evaluators and that they are self-​interested and have true beliefs, and in those circumstances, both the constitutive and the evidential view imply that well-​being coincides with the satisfaction of people’s actual preferences. One reason why the coincidence of well-​being with preference satisfaction appeals to economists is that most economists dislike paternalism –​that is, coercion that aims to benefit the person who is coerced. If what people want coincides with what is good for them, then

130

Welfare

no coercive interference with their actions could possibly benefit them. However, this is not a good reason to take welfare to coincide with the satisfaction of actual preferences. Paternalism is sometimes desirable. Grabbing someone to prevent her from stepping into an open manhole that she did not see is one simple example of a justified paternalistic action; so arguably are seatbelt laws. There are serious objections to the view that one can measure welfare by an agent’s self-​interested preference satisfaction, whether that preference satisfaction is interpreted as evidentially or as constitutively relevant to well-​being. Two are particularly obvious. The first is that, unlike the purely self-​interested creatures who populate a good deal of economic theory, people care about more than their own well-​being. People sometimes sacrifice their own well-​being in order to benefit others they love or to do harm to those they hate. And while almost no one’s welfare is affected by whether the continuum hypothesis turns out to be true, it is easy to see how mathematicians could have a preference about that. Many people prefer that their unborn great great grandchildren have good lives, even though that will have no bearing on their own well-​being. The second objection arises from the fact that people are ignorant of many things. Consequentially, people may prefer something that is bad for them because they mistakenly believe it is beneficial. It is not true that x is better for A than y if and only if A prefers x to y. Indeed, many people live in circumstances in which their governments, their poverty, or their lack of education makes it almost impossible for them to make informed choices.

8.3  Against the Constitutive View: Well-​Being Is Not Preference Satisfaction It is not clear whether economists are committed to the constitutive view of the relationship between preference satisfaction and well-​being. It is more charitable to interpret the connection between preference satisfaction and well-​being that they rely on as reflecting only an evidential connection. The evidential view is not without its problems, but these problems are small compared to the philosophical difficulties attaching to the constitutive view. Before presenting these difficulties, we should point out that our rejection of the view that well-​being is the satisfaction of informed and self-​ interested preferences is opposed to the philosophical consensus, which judges these theories to be among the most promising of theories of well-​ being. The views of philosophers are, however, somewhat different from those of economists. While economists (on our interpretation) leave

8.3  Against the Constitutive View: Well-Being Is Not Preference

131

implicit the restriction to self-​regarding preferences that are undistorted by false beliefs, philosophers such as Gauthier (1986, ch. 2); Goodin (1986); and Griffin (1986) develop with care restrictions on which preferences should be considered. They defend a theory of well-​being as the satisfaction of well-​informed, self-​regarding, and “rational” preferences. Notice that taking well-​being to be the satisfaction of rational preferences rather than actual preferences free of mistaken beliefs shifts the emphasis from what people in fact prefer to what is rational for them to prefer. If it could be shown, for example, that it is rational for everyone to prefer happiness to unhappiness or virtue to vice, then a view of well-​being as the satisfaction of rational preferences would lead to a substantive view of well-​being as involving happiness or virtue. A number of considerations have made a constitutive theory of well being as preference satisfaction seem attractive. It offers an immediate explanation for the fact that different things contribute to the well-​being of different people. It can be deployed without taking sides in the different substantive ends that people pursue. Indeed, among the many things that apparently make life go well are successes in the worthwhile projects that individuals have chosen. It is only because individuals prefer to pursue those projects that those successes contribute to their lives. If Henry V did not actually want to vanquish the French at Agincourt, then it would have meant little for him to succeed at it. Moreover, because what satisfies people’s preferences (at least if they are well informed and self-​interested) typically makes them happy, the preference satisfaction view of well-​being can explain why hedonism (the view that well-​being is pleasure) has been so influential. In addition, because well-​being seems to involve satisfaction, it might seem plausible to associate well-​being with preference satisfaction. But this thought rests on an equivocation on the word “satisfaction.” The satisfaction of preferences is like the satisfaction of degree requirements. It has no necessary connection to any feelings of satisfaction. To satisfy a preference for state of affairs x over y is for x to obtain rather than y. Knowing that x obtains may lead one to feel satisfied. But otherwise there is only a contingent connection between the satisfaction of a preference and any feeling of satisfaction. Satisfying the preferences of those who, concerned about climate change, want the Earth to remain habitable three hundred years from now will give them no feeling of satisfaction, because they will be long dead. Despite the considerations that may make preference satisfaction theories of well-​being seem plausible, these theories should be rejected. Recall

132

Welfare

that we have described them as maintaining that the satisfaction of well-​ informed and self-​interested preferences constitutes well-​being. As careful readers may already have noticed, what can it mean to say that some preferences are “self-​interested”? In ordinary language, Satz’s preferences are self-​interested if and only if they are directed toward Satz’s own well-​being. To maintain that Satz’s well-​being consists in the satisfaction of those preferences of Satz that are directed toward Satz’s well-​being presupposes some other theory of well-​being than preference satisfaction. Philosophers have, unsurprisingly, noticed the problem. At the same time, some restriction on preferences is needed, lest one is forced to conclude that the satisfaction of preferences that have nothing to do with people’s own lives nevertheless makes them better off. Consider the following example of Derek Parfit’s: Suppose that I meet a stranger who has what is believed to be a fatal disease. My sympathy is aroused, and I strongly want this stranger to be cured. We never meet again. Later, unknown to me, this stranger is cured. On the Unrestricted Desire Fulfillment Theory, this event is good for me, and makes my life go better. This is not plausible. We should reject this theory. (1984: 494)

To restrict the class of preferences whose satisfaction constitutes well-​being without surreptitiously invoking some other notion of well-​being to distinguish self-​interested preferences, philosophers have attempted to distinguish those preferences that are “about” the agent from those that are not. For example, Mark Overvold (1984) proposes that an agent’s well-​being should be regarded as the satisfaction of those among the agent’s preferences that concern states of affairs that entail the agent’s existence. A different way to respond to the problem that the satisfaction of Parfit’s preference that the stranger be cured seems not to make Parfit any better off is to conclude that whether or not Parfit prefers x to y has by itself no effect on whether x is better for Parfit than is y. Why should anyone have ever thought that some state of affairs x that would otherwise be no better for Jack than is some other state of affairs y becomes better for him simply because he comes to prefer x to y? Suppose Jack has a preference concerning some aspect of his body that is hard to observe and irrelevant to his health: perhaps he prefers to have a small benign tumor on his spleen. This is clearly a “self-​regarding” preference; it concern states of affairs that entail his existence. Is it plausible to believe that, owing to his preferences, Jack’s possessing a small benign tumor on his spleen makes him better off? Defenders of a preference satisfaction view might respond that all of Jack’s preferences must be based on some advantage he sees for himself. Unless that advantage depends on a false belief, it shows that what Jack

8.3  Against the Constitutive View: Well-Being Is Not Preference

133

prefers is, after all, better for him. But such a defense implicitly denies that well-​being is the satisfaction of preferences. If a small tumor on his spleen is better for Jack, it is because of the advantages that led him to form his preference, rather than the preference itself. There is a gap between what people prefer and what is actually good for them. It is important not to confuse the satisfaction of a preference with coming to know that a preference is satisfied. If Jack learns he has a benign spleen tumor, he may be pleased and in this way he is better off. But to point to people’s pleasure at learning that their preferences are satisfied is to take well-​being to be pleasure, not preference satisfaction. The defender of the constitutive theory might reply that even if preferences cannot conjure value out of nowhere, they are necessary in order for an agent’s activities to promote her well-​being. Suppose that Jill is very serious about bowling. She works hard at improving her skills and concentration and her skills improve. Her improvement gives her pleasure, but the contribution that Jill’s bowling makes to the value of her life cannot be captured by the pleasure she takes in the activity. Her life is better for the successes she achieves in this activity, and if it were not for her preference for this activity over others, she would not have had nearly so much pleasure, and the successes she achieves would be unlikely and would not contribute as much value to her life. Our response is that although it is crucial that Jill prefer bowling to other activities, what actually makes her better off is her achievement and the pleasure it yields, not the fact that it satisfies her preferences. If economists want to be philosophically modest, then they should not accept the constitutive view, which commits them to a controversial philosophical theory of what constitutes well-​being. So far, we have not mentioned the problems that arise when people prefer things that are, by some criteria, objectively bad for them. Some of these preferences may reflect poor information; it might be reasonable to believe that had the people had better information, they would not have prefered things that make their life go worse. But in other cases, lack of knowledge does not seem to be the relevant issue: would sadists who have access to full information about the harms they cause automatically change their preferences? This is doubtful. The evidential view avoids these philosophical difficulties, and it serves the practical purposes of economists just as well as the constitutive view does. Measurements of well-​being have many practical purposes: to establish benchmarks for equality, to assess progress over time, to compare living standards across communities, to weigh the comparative claims that different persons may justifiably make on social resources, and so on. What most

134

Welfare

welfare economists care about is whether they can measure well-​being by people’s preferences, which the evidential view of the connection between preference satisfaction and well-​being permits. In addition, as Sections 8.4 and 8.5 argue, the evidential view avoids a number of additional objections to accounts that define welfare in terms of preference satisfaction.

8.4  Conflicting Preferences and Well-​Being Measuring welfare by preference satisfaction leads to complications when preferences change. As Brandt (1979, ch. 13) and Parfit (1984, ch. 8) point out, if an individual’s preference ranking changes, then it is unclear whether the individual is made better off by satisfying or frustrating the original preferences. Such cases seem to give rise to something akin to the problem of interpersonal utility comparisons: how does Jack’s present self ’s satisfaction with y compare with his past self ’s satisfaction with z (Gibbard 1986)? Should one care about satisfying preferences that people no longer have? Is there any reason now to satisfy Hausman’s childhood desire to be a garbage collector? At this stage in his career, Hausman thinks not. Why not? Why should his current desires take priority over his past desires? One answer might be that his current desires not to be a garbage collector are stronger than his childhood desires. But as a six-​year-​old he longed intensely to drive a garbage truck. A second reason not to care about satisfying preferences that people no longer have is that giving people what they no longer want gives them no feeling of satisfaction. But this reason is not available on the constitutive view of the connection between welfare and preference satisfaction. In giving this reason one has shifted from a preference satisfaction theory to a mental-​state theory of welfare. The evidential view also enables us to give a third reason why satisfying Hausman’s former preference will not benefit him: Hausman now has a much better understanding of what will make his life go well or badly than he had when he was six. He is (he hopes) a more competent evaluator. Of course, not all instances of preference change can be dealt with so easily. When people’s preferences change, their later preferences do not always reflect better judgment than their earlier preferences. This theoretical problem of preference change is linked to a practical problem: policies and institutions change people’s preferences. (It is hard to believe that the billions spent on propaganda and advertising have no effect on preferences.) Assessments of policy must then depend in part on one’s views concerning which institutions provide a suitable framework for developing desirable preferences (McPherson 1982, 1983b). Additionally,

8.5  Assessing Preferences

135

how concerned should one be about satisfying current preferences if one judges that they are likely to change? Should one aim to modify preferences so that they will be easier to satisfy? How should one choose between either satisfying existing preferences or modifying preferences first and then satisfying the modified preferences? On a constitutive view, these questions have no good answer. On an evidential view, the answer presupposes a substantive theory of well-​being that enables economists to judge which preferences offer better or worse evidence concerning well-​being. Economists do not have such a theory, and most of them do not want it. But, given preference change, there is a significant ambiguity about what satisfies preferences.

8.5  Assessing Preferences On the constitutive view, provided that McPherson is well informed, one makes him better off by satisfying his self-​regarding preferences, regardless of how idiosyncratic or obnoxious they are and regardless of how they were formed. But some of the real world uses of the theory of welfare demand that policy makers discriminate among preferences. For example, as Thomas Scanlon has pointed out, both policy makers and public discourse rely on a relatively objective standard of “urgency” when weighing the strength of competing claims for social provision. The urgency of people’s claims does not correspond to the strength of their preferences. For example, even if members of a destitute religious group prefer subsidies to build religious monuments over receiving food and shelter, their beneficent fellow human beings (whether fellow citizens or foreign donors) might acknowledge a moral obligation only to provide food and shelter (Scanlon 1975; see also Sagoff 1986). Without pretending to settle all questions about what makes people better off or worse off, and without taking any stand concerning what things are of ultimate value, a good deal of social policy rests on a settled agreement of what beneficence demands. What we owe to others is understood in terms of objective factors such as relieving hunger or homelessness, not in terms of the subjective preferences that agents happen to have. One way of defending discrimination among preferences (which is compatible with an evidential but not a constitutive view of the connection between preferences and welfare) would be to argue that members of the religious group are mistaken about what promotes their well-​being. An alternative interpretation  –​that perhaps accords better with the spirit of liberalism –​is to insist that the policy decisions in a liberal state depend

136

Welfare

on different standards than the decisions of individuals concerning what goods to pursue. The state’s job is to provide for basic needs and to make available to individuals a wide range of activities, not to promote the subjectively valuable projects of individuals. Here one can see, by the way, one reason for preferring “in-​kind” provision to transfer payments. If donors provide cash benefits, destitute members of this religious group will not use the funds to alleviate the needs that gave rise to our obligation to assist them. Those who defend the view that policy should aim to promote welfare, where welfare is measured by preference satisfaction, might claim that our resistance to honoring idiosyncratic preferences is purely practical: we would open ourselves to manipulation and misrepresentation if we let people’s reports of their subjective needs govern public distribution of benefits. One can also argue that the state has no business partnering with individuals in their own projects. Should public benevolence or justice be sensitive to what people prefer? What moral pull should satisfying Hausman’s preferences have on McPherson or Satz? Thomas Nagel has argued that if something is valuable to people only because they want it, then their getting it has no direct moral importance for others (1986, ch. 9). Others have no reason to satisfy Hausman’s preferences unless they can make sense of why what he wants is worth wanting, or why his life will be better in some substantive way if he gets what he wants. On the constitutive view, there is nothing to be said about whether what people prefer is worth preferring, because to maintain that some things are not worth pursuing assumes that there is some source of value other than the satisfaction of preferences. On the evidential view, in contrast, there is nothing odd about asking whether something is worth pursuing  –​although it may be difficult to reach agreement about this in many cases. Consider that many Boston residents desperately wanted the Red Sox to win the World Series in 2004. Their happiness when the Red Sox won gave others, even Yankees fans, some reason to want their preferences to be satisfied. But Nagel maintains –​very plausibly, we believe –​that the mere preferences of Red Sox fans (as opposed to their happiness or unhappiness or other grounds for their preferences) should be of no moral importance to others. This line of thought implies that social policy should not be concerned with satisfying preferences except insofar as doing so coincides, as the evidential view maintains, with other social objectives. Consider those who have expensive tastes. An even-​handed concern to satisfy their preferences appears to be unfair to those with more modest tastes. If welfare is preference satisfaction (or indeed happiness), then

8.5  Assessing Preferences

137

a person who has cultivated a taste for “prephylloxera claret and plover’s eggs” (Arrow 1973, p. 254) without an income that makes them affordable is worse off than someone with a similar income who wants only affordable beans and franks (see also Dworkin 1981a). But should social policy be responsive to expensive preferences? What is at issue may be fairness rather than well-​being: the person with the taste for plover’s eggs may well be worse off without the plover’s eggs, but policy makers distributing scarce resources do not have to care. Should the unsatisfied preferences of those with expensive tastes not count at all? Perhaps their unhappiness should count in some way, but why should their inability to procure the expensive things they want be a matter of social concern? Those who defend more objective views of well-​being maintain that, except insofar as they are unhappy, the fancy eaters are not at all worse off. If one refuses to be disturbed about failures to satisfy expensive preferences, then either one’s benevolence is limited or one does not accept the view that preference satisfaction constitutes well-​being. Racist, sadistic, and other antisocial preferences raise related problems for the constitutive view. Some of these may be based on false beliefs. But that is not always the case. One reason why it is bad to satisfy them is that doing so frustrates other (and characteristically stronger) preferences. Should policy makers balance the harms of these preferences against the “benefits” of satisfying them? Should antisocial preferences count at all (Harsanyi 1977a, p. 56)? Given either a hedonistic or a preference satisfaction theory of well-​ being, it seems that antisocial preferences ought to count and that a benevolent person should, other things being equal, strive to satisfy them. More objective views of well-​being, on the other hand, such as the evidential view of the relations between well-​being and preference satisfaction, permit one to deny that individuals are made better off when their antisocial preferences are satisfied. This is certainly a mark in their favor. Defenders of a preference satisfaction view of welfare have some ways of deflecting the claims of those with antisocial and expensive preferences without questioning the identification of well-​being and the satisfaction of preferences. First, they can claim that some additional moral consideration –​such as justice, which is distinct from benevolence  –​is involved. Racist preferences do not count because they deny some people’s rights. Second, they can note that preferences are malleable and that expressions of preferences respond to incentives. Frustrating expensive and antisocial preferences discourages their expression and may in the long run enable more preferences overall to be satisfied. These points are plausible, but they leave defenders of a constitutive view of the relationship between preference

138

Welfare

satisfaction and well-​being with the unsatisfactory implication that, other things being equal, there is reason to satisfy antisocial and expensive ­preferences –​as if the extent to which racism limits well-​being depends on how preference satisfactions turn out. A final difficulty with the view that well-​being coincides with preference satisfaction concerns preference formation. Some preferences derive from previous injustice, coercion, or manipulation, and people may form preferences as the result of problematic psychological mechanisms. Some people want things precisely because they cannot have them (“The grass is always greener on the other side of the fence”), while others spurn what is beyond their reach, like the fox who judged the unobtainable grapes to be sour (Elster 1983; Sen 1987b, 1990a). As Sen has pointed out, women who are systematically denied roles in public life or equal shares of consumption goods may learn not to want these things. Women who have never known freedom may not know how to value it. Women who have never been paid wages comparable to men’s wages and have been subject to regular abuse may not demand fair wages or even physical security. But liberties, high wages, and protection from domestic violence will more plausibly make these women better-​off than giving them what they may prefer. Satisfying preferences that result from coercion, manipulation, or “perverse” preference formation mechanisms does not always make people better off. Suspect preferences sometimes reflect false beliefs, and some of the problems here are versions of problems we have already discussed. But coercion and manipulation may also distort the psychological mechanisms that lead people to value some things more than others. For example, there is evidence that employers keep their bonded laborers separate from other wage workers so they cannot have access to information about leaving bondage; the employers seek to cultivate servility (Schaffner 1995). When the preferences of oppressed people derive from the circumstances of their oppression, one cannot measure their welfare by considering how well their preferences are satisfied. The constitutive view fails to register this point.

8.6  The New Hedonist Welfare Economics Many nineteenth-​and early twentieth-​century economists were utilitarians who regarded well-​being as a mental state such as pleasure or happiness. In contrast, modern economic theory, as developed in the 1930s, abandoned hedonism. Because economists found that the basic propositions of demand theory and consumer behavior require only that people have stable preference rankings with certain properties, most took well-​being to be the

8.6  The New Hedonist Welfare Economics

139

satisfaction of preferences. Even if many economists informally continue to associate welfare and happiness, standard welfare economics for the better part of a century has measured welfare by preference satisfaction. Over the last two decades, a number of prominent economists, impressed with improved techniques for measuring subjective states, have adopted more sympathetic attitudes toward hedonism (Kahneman and Sugden 2005; Kahneman and Krueger 2006; Kahneman and Thaler 2006). They assume that people’s preferences depend on their own estimates of the happiness alternatives will produce, and that direct measures of pain and pleasure would be more accurate measures of well-​being. For example, Kahneman and Krueger (2006) define a misery index that they call the “U-​index,” which measures what proportion of their time people spend in unpleasant states. An espisode counts as unpleasant if the most intense feeling people report concerning it is negative. Kahneman and Krueger then measure people’s subjective well-​being by the percentage of their time that people spend in unpleasant episodes. To lessen the unhappiness they measure, Kahneman and Krueger suggest mental health interventions and policies to reduce the amount of time spent engaged in unpleasant activities such as commuting to work or doing housework. The proposal that welfare economics should focus on happiness (“experienced utility”) rather than preferences (“decision utility”) has already had a practical impact. Both the Organization for Economic Development and Cooperation) (OECD) and the British government have begun to collect systematic data concerning subjective well-​being and have sought to appraise policies by their consequences for people’s subjective states. The measures of subjective well-​being rely on survey questions concerning people’s affect, their “life evaluation,” and their psychological flourishing. Something of the same ambiguity we saw between a constitutive and an evidential view of the relationship between preference satisfaction and well-​being may arise here. It is ambiguous whether these measures assume that people’s subjective states as measured by these surveys constitute their well-​being or whether they provide evidence of their well-​being, which need not be a subjective state. Hedonism has obvious attractions. Pleasure strikes many people as intrinsically good, and pleasures seem to be measurable. Many different things give people pleasure, and so hedonism recognizes that different activities and outcomes are good for different people. But hedonism also faces serious objections. Pleasures and pains are diverse. The satisfaction of solving a math problem is a different kind of mental state from the delight of hearing some wonderful song or the comfort of a warm bath after a day outdoors in the cold. Unpleasant states are at least as diverse. How does one

140

Welfare

put vertigo, nausea, and toothaches on a single scale with the pleasures of achievements, music, or baths? Although contemporary psychologists have made progress on the measurement difficulties, there is a fundamental objection to hedonism and indeed to all mental state theories of well-​being. Those who maintain that well-​being consists in mental states are committed to the view that two individuals who are in the same mental state are equally well off, even if their objective circumstances differ. So someone whose life is going splendidly may be no better off than someone else whose life is going miserably but who mistakenly believes that it is going well. Robert Nozick (1974, p. 41) proposed a hypothetical case that illustrates this difficulty. Suppose there were an “experience machine” that could give people the highest quality experiences possible. These terrific experiences might be intense physical pleasures or they might be experiences of climbing Everest or composing a symphony that surpasses Beethoven’s best efforts. (Once attached to the machine, people no longer know that they are attached; the machine takes over their consciousness and memories completely.) Suppose people attached to this machine experience whatever states of consciousness that mental-​state theorists of well-​being claim to constitute what is ultimately and intrinsically good. The mental-​state theorist would then have to say that all people would be better off permanently hooked up to a reliable experience machine rather than living their own lives and experiencing the decidedly mixed mental states that will be part of any human life. Some people, whose lives are ones of intense and unremitting suffering, would arguably be better off on the machine. But if one believes that those who are hooked up to the experience machine are missing some of the intrinsically good things in life –​even though they are not, by assumption, missing the best mental states –​then one cannot accept a mental-​state view of well-​being. We conclude that mental state theories of well-​being, including hedonism, are mistaken. But just as preference satisfaction may provide evidence concerning well-​being even if it does not constitute well-​being, so subjective experience may indicate well-​being without constituting it. The fact that hedonism is untenable as a theory of well-​being does not rule out the proposal to measure well-​being by people’s subjective states. Nevertheless, we are skeptical. We have already raised some doubts about whether subjective experience is a good indicator of well being. We also have doubts about the reliability of the measures of subjective experience. There is, as yet, no canonical way of measuring subjective experience, and the different measures give different results. Furthermore, when economists evaluate policies by subjective experience rather than by preference

8.7  Other Theories of Well-Being

141

satisfaction, they cannot piggyback on individuals’ judgments of the ways in which policies bear on more distant ends. For example, some surveys of people’s moods have shown that people find taking care of their children as unpleasant as housework (Kahneman et al. 2004). This finding does not justify policies designed to reduce the amount of time people spend rearing their children, because child rearing has many important and less immediate consequences. Standard welfare economics relies on the trade-​ offs people make between immediate affect and longer-​term consequences, which are reflected in their preferences for child rearing over dishwashing. In contrast, economists who measure welfare by subjective experience need to rely on their own judgments of how the values of the longer-​term consequences of policies should be balanced against their immediate pleasures or pains.

8.7  Other Theories of Well-​Being Substantive theories of welfare purport to say which things are intrinsically good. Hedonism is a substantive theory, as are “perfectionist” views (Griffin 1986, ch. 4; Raz 1984, ch. 12)  and what Parfit calls “objective list” views. Substantive views are objective in the sense that what is good for people is not determined by whether people believe it is good for them. Even as they take well-​being to coincide with the satisfaction of preference, many economists are also committed to a substantive view of well-​being as material self-​interest. These views are compatible if and only if people’s preferences reflect their material self-​interest. As the example of hedonism shows, the objectivity of substantive views of well-​being does not imply that subjective states do not matter, and most substantive theories give some weight to mental states. While objective views are controversial, they should nevertheless be tempting to economists insofar as they make well-​being more readily measurable than are mental states. In addition to material self-​interest, one objective view of what matters for economic policy, which may be relevant to normative economists, is John Rawls’s account of “primary goods.” This is closer to a generalized notion of resources than to an alternative theory of well-​being. In A Theory of Justice (1971) Rawls takes well-​being to be the satisfaction of rational preference, but he does not think that justice should focus on well-​being. One reason is that people’s well-​being depends in large part on their own efforts. Amartya Sen defends a theory of well-​being in terms of the capabilities and the functionings an individual attains. A person’s functionings consist in

142

Welfare

those things that the person does and experiences. Walking, playing the piano, being well nourished, loving one’s friends, understanding Chinese, and appreciating cubism are all functionings. But Sen does not think that social policy should concern itself directly with the functionings that people achieve. Social policy, insofar as it is motivated by a concern for welfare, should instead focus on capabilities (Sen 1987c, 1992a). A capability is the ability to achieve a certain sort of functioning. For example, literacy is a capability while reading is a functioning. People may value capabilities for their own sake as well as for the functionings they permit –​you are glad to know you can walk around even if you are inclined to stay put. In Sen’s view, social policy should be less concerned with functionings –​ with what people make of their capabilities –​than with capabilities, because functionings depend heavily on individual choices. For example, a shortfall in functioning such as malnutrition might stem from an individual’s decision to embark on a religious fast rather than from any unjust deprivation. In contrast to Rawls, Sen maintains that social policy should not focus on resources or primary goods, because people with equal primary goods may lead very different lives owing to traits that are internal to the individual such as disabilities. For example, as a result of a digestive disorder, someone may be malnourished despite having a normal diet. In Sen’s view, the focus of policy should be on ensuring that people have a decent set of capabilities, broadly understood. Martha Nussbaum has also conceptualized human flourishing in terms of capabilities (2000, 2001). Her view of capabilities differs from Sen’s in two ways. First, its roots are in Aristotle’s moral philosophy rather than in the inadequacies of conventional welfare economics. Second, Nussbaum has formulated a specific list of central human capabilities.1 She recognizes that some items on the list might be realized differently in different societies and that some items are more firmly fixed than others. Nonetheless, she argues that the list has survived considerable cross-​cultural scrutiny. While Sen has also been concerned to make capabilities measurable and practical, he has been more cautious about specifying the most important human capabilities. Nussbaum’s and Sen’s approaches complicate the measurement of well-​ being. People vary in the capabilities that are important to them, and they Nussbaum’s list of central human capabilities (2001, pp. 416–​418) consists of the following 10: (1) Life; (2) Bodily Health; (3) Bodily Integrity; (4) Senses, Imagination, and Thought; (5) Emotions; (6) Practical Reason; (7) Affiliation; (8) Other Species; (9) Play; and (10) Control over One’s Environment, both Political and Material. 1

Suggestions for Further Reading

143

may disagree about the rankings of capabilities. Do preferences reenter to rank capabilities? Furthermore, figuring out what capabilities a person has is a fraught enterprise. Does Satz still have the capability to be a professional musician, or did she lose that when she gave up the piano at age thirteen? And did she have that capability at age thirteen? Even if the problems of measuring the components of well-​being and assigning weights to them prove to be greater for more objective approaches than are the problems of measurement in standard welfare economics, it can still be argued, following Sen (paraphrasing Carveth Read and John Maynard Keynes), that “it is better to be vaguely right than precisely wrong” (1987a, p.  6). In particular, objective approaches to well-​being may lead research in directions that link up naturally to the normative terms characteristic of policy debate. For example, an objective index is much more helpful as a measure of the extent of deprivation in less developed countries than is utility theory. The index that is currently most used, the human development index (which is, roughly, an average of life expectancy, literacy, and per capita GDP), owes a great deal to Sen’s views (and to Sen himself). A much more detailed set of indicators can be found in Anand et al. (2009).

8.8  Conclusions Measuring well-​being by the satisfaction of preferences is problematic. It mistakenly suggests that social policy should attend to all preferences –​even if they are expensive, antisocial, or the results of false beliefs, manipulation, or problematic psychological processes. The focus on preference satisfaction does not cleanly link up with the normative terms of policy debate, and it leads to difficulties when preferences change and conflict. Yet as we shall see in the following chapter, economists have found ways of measuring preference satisfaction in order to inform policy, and, as we have noted, identifying well-​being with the satisfaction of preferences ties together positive and normative economics. Whether the link is of value depends on how well economics succeeds in addressing problems of human welfare.

Suggestions for Further Reading Defenses of an informed preference satisfaction view of individual well-​ being can be found in Gauthier (1986, ch. 2), Goodin (1986), Griffin (1986), and Arneson (1990). The introduction to Utilitarianism and Beyond (Sen and Williams 1982) summarizes arguments against preference-​based

144

Welfare

approaches. The central argument here derives from Kraut (2007) and Hausman (2012). See also Mill (1863), Rawls (1982), Sen et al. (1987), and Scanlon (1998). Harris and Olewiler (1979), Hammond (1983), Machina (1987), Broome (1991b, ch. 10), and Hausman and McPherson (1994) discuss the problems that arise in respecting preferences when there are uncertainties. For discussions of the new hedonist welfare economics see Kahneman (2000a, b), Kahneman et  al. (2004), Kahneman and Sugden (2005), Kahneman and Kreuger (2006), Kahneman and Thaler (2006), Layard (2006), OECD (2013), and Vizard and Rusgys (2013). Sen’s views on capabilities and functionings are developed mainly in two of his essays (1985a and 1987c) and are discussed at length in Nussbaum and Sen (1993). Nussbaum (2000, 2001) attempts to recast political theory in terms of fundamental capabilities. For more on the human development index, see http://​hdr.undp.org/​en/​content/​human-​development-​index-​hdi?

Questions for Study and Discussion 1. How does the theory that well-​being is happiness differ from the theory that well-​being is the satisfaction of preferences? Which theory do you think is more plausible? 2. If you were convinced that there were an absolutely reliable experience machine that would give you the subjective experience of what you believe would be an absolutely marvelous life for you, full of all the things that you value the most, would you choose to be connected? Why or why not? 3. What exactly is the difference between the evidential and the constitutive views of the relations between preferences and well-​being? Under what circumstances are both of these views false? If there are many circumstances in which what is best for people does not coincide with what they want, is it not a mistake for economists to measure well-​being by prefrence satisfaction? 4. What are the main objections to the constitutive view of the relationship between well-​being and the satisfaction of preferences? Do you think that these objections are decisive? 5. It seems plausible to maintain that satisfying one of Jill’s preferences makes her better off only if she has the preference at the time it is satisfied. Satisfying preferences that agents no longer have or do not yet have does not count. How would you explain this?

Questions for Study and Discussion

145

6. Some people have modest desires, while others are unsatisfied unless they have very expensive things. If public policy aims impartially to help people to satisfy their preferences, then those who want expensive things have greater claims on public resources than those who are more easily satisfied. Do you accept this implication? Why or why not? 7. What do you think makes for a good human life and what philosophical theory of well-​being seems most plausible to you? 8. Do you think that welfare economics would be more useful if economists evaluated policies by the implications for happiness rather than by their implications for preference satisfaction? Why or why not? 9. Rawls’s theory suggests that policies can be evaluated by their consequences for people’s holdings of primary goods, while Sen and Nussbaum urge instead that policies be evaluated by their consequences for capabilities and functionings. What advantages do these proposals have over focusing on preference satisfaction, as economists do? What disadvantages? 10. Should policy makers count people’s antisocial preferences in policy decisions? Why or why not?

NINE

Welfare Economics

Mainstream economists evaluate economic institutions, policies, and outcomes by asking whether they make people better off, and, as we have seen in Chapter 8, they measure whether people are better off by how well their preferences are satisfied. Although focused on welfare and consequences, few economists are utilitarians, because most deny that evaluations should rely on comparisons of how well satisfied are the preferences of different individuals. Following Robbins (1935, ch. 6), many mainstream economists have regarded interpersonal comparisons as untestable subjective value judgments, although this is changing. We grant that making interpersonal comparisons of the extent to which preferences are satisfied is difficult. On the other hand, appraising economic outcomes, institutions, and policies exclusively by their consequences for welfare without interpersonal comparisons would tie normative economists’ hands behind their backs. They need to be able to compare the consequences of policies for winners and losers in a nonarbitrary way, and they must make room for other important normative concerns, such as justice. This chapter examines how economists tackle these challenges, including the ethical assumptions they make along the way, and it discusses whether the form of economic appraisal that they construct –​that is, welfare economics –​succeeds. Without the possibility of interpersonal comparisons, it seems impossible adequately to encompass moral considerations that should influence policy such as justice, rights, and equality. Accordingly, traditionally normative economists have favored a division of labor whereby welfare economics confines itself to addressing only the welfare consequences of economic institutions and policies and leaves the appraisal of policies with respect to other normative criteria to others. This exclusive focus on welfare is by no means universal among economists, and in later chapters, we will review 146

9.1  Preference Satisfaction, Pareto Efficiency, and Competitive

147

some of the work that economists have done developing other normative dimensions of assessment. In addition, as we noted in the Chapter 8, there is increasing interest among economists in measuring welfare in terms of happiness or life satisfaction. But in this chapter our focus is on standard welfare economics, which assesses outcomes, policies, and institutions exclusively by how much they enhance or diminish welfare, as measured by the extent to which preferences are satisfied. Welfare economics depends both on a specific view of how to measure welfare and on the view that inquiries into welfare can be separated from inquiries into freedom, rights, equality, and justice. In one way, this separation limits economists. They appraise policies only along one dimension. At the same time, it frees them from having to be concerned with anything but welfare. Having bracketed everything except preference satisfaction, welfare economists have sometimes (as in the World Bank memorandum) felt themselves free to ignore all other moral questions, and they run the risk of overstating the significance of their own partial mode of evaluation.

9.1  Preference Satisfaction, Pareto Efficiency, and Competitive Equilibrium Once economists measure well-​being by preferences and reject interpersonal welfare comparisons, the central features of standard normative economics follow naturally. All economists need is one innocuous moral principle of minimal benevolence: other things being equal, it is a morally good thing if people are better off. Because of the “other things being equal” clause, minimal benevolence is uncontroversial. It does not, for example, say whether it is better or worse to make some people better off at the cost of increasing inequality. It only says that if all other morally relevant considerations (such as justice) are not at issue, then it is a morally good thing to make people better off. Those who accept minimal benevolence and identify an individual’s welfare with the satisfaction of that person’s preferences will judge that, other things being equal, it is a morally good thing to satisfy an individual’s preferences. The main issue in standard normative economics is accordingly to what extent economies enable individuals to satisfy their preferences. From this follows the importance of the notions of “Pareto improvements” and “Pareto optimality.” R is a Pareto improvement over S if nobody prefers S to R and somebody prefers R to S. R is a Pareto optimum (which is also called a “Pareto efficient” allocation) if and only if there are no Pareto improvements over R.

148

Welfare Economics

Once economists measure well-​being by preference satisfaction, they can describe a Pareto improvement as a change that makes some people better off without making anyone worse off, and they can say that an allocation is Pareto efficient if it is impossible to make someone better off without making someone else worse off. In the most general sense, “An allocation of resources is efficient if it is impossible to move towards the attainment of one social objective without moving away from the attainment of another objective” (Le Grand 1991, p. 27). Efficiency in economic theory is, however, typically Pareto efficiency –​that is, efficiency with respect to the satisfaction of preferences. There will typically be many Pareto optima, and many of these will be ethically unattractive because of their inequality. If one accepts minimal benevolence then, other things being equal, Pareto improvements are moral improvements, and states of affairs that are not Pareto efficient are morally undesirable. With the help of the notion of a Pareto optimum, we can state two formal theorems of welfare economics. The first welfare theorem establishes that the allocation resulting from any perfectly competitive equilibrium is Pareto optimal. Perfect competition obtains when there are (1)  no interdependencies among people’s utility functions, (2) markets for all goods and services (and thus no externalities), (3) no barriers to entry or exit from any market, and (4) so many traders in every market that no traders can influence prices. A perfectly competitive equilibrium obtains when there is perfect competition and there is no excess demand on any market. The first welfare theorem (under highly idealized circumstances) resembles Adam Smith’s claim (1776, book IV, ch. 2) that a self-​interested agent who “intends only his own gain” is “led by an invisible hand to promote an end [the welfare of society] which was no part of his intention.” Given this first welfare theorem, economists can conclude that, ceteris paribus, perfectly competitive equilibria are morally desirable, and market imperfections that interfere with the achievement of competitive equilibrium are morally undesirable. This is a theoretical defense of perfect competition, not a defense of actual markets or of laissez-​faire policy; and it is only a defense of perfect competition other things being equal. Given these extensive qualifications, it is best not to exaggerate the virtues of Pareto optimality. A Pareto optimal state of affairs, and thus (according to the first welfare theorem) a perfectly competitive equilibrium, escape one sort of moral criticism –​one cannot say of a Pareto optimal state of affairs that it is possible to satisfy anybody’s preferences better without cost to anyone else. This negative formulation captures the full –​and quite limited –​attractions of Pareto optimality. Few, if any, economists would deny that there are other moral considerations

9.1  Preference Satisfaction, Pareto Efficiency, and Competitive

149

that should influence the choice among policies in addition to how well they satisfy preferences. Consider a state of affairs R in which millions of people are starving. If there is no way to make these people better off without making a billionaire slightly worse off, then R is a Pareto optimum. Given the importance of other moral considerations, one might question whether welfare economics is useful. In response to this worry, economists can invoke the second welfare theorem. This theorem has been called the “rebel’s catechism,” because it purports to show rebels that addressing distributional concerns does not require limiting markets. It shows that all Pareto efficient allocations can be obtained as competitive equilibria from some initial distribution of endowments. In other words, one can achieve a Pareto optimal outcome that satisfies other moral constraints, such as the demands of justice, by first redistributing in a lump sum the resources that individuals bring to their market interactions and then letting individuals trade freely under the highly idealized conditions of perfect competition. Economists use this second welfare theorem to argue that market interactions will promote efficiency, and that other moral concerns can be met by adjusting the initial entitlements and resources that people have. Whether they are defenders of laissez faire or of extensive government intervention to address market failures, most economists share a moral commitment to perfect competition as an ideal. It is this commitment that gives point to the analysis of market failures. (Why should correcting market failures matter if market successes are not a good thing?) Yet this commitment seems paradoxical: as we have stressed, economists attempt to steer clear of controversial ethical commitments when doing theoretical welfare economics. (Indeed, some economists have supposed that theoretical welfare economics is independent of all value judgments.) Yet when welfare economists address policy questions, they purport to know how to make life better for people. What explains how economists with these views can feel themselves possessed of moral authority without the trouble of doing moral philosophy? The answer is that welfare economists do not usually regard their focus on preference satisfaction as having any ethical content; it seems to be just a part of the standard view of rationality. Once they measure well-​being by preference satisfaction, they need only add the uncontroversial principle of minimal benevolence to derive the argument for perfect competition sketched here. Unfortunately, the argument that perfectly competitive equilibria can encompass simultaneously all moral desiderata is not sound. It is not always possible to achieve a Pareto optimal outcome while adhering to all other

150

Welfare Economics

moral constraints. As we have seen, there may be moral grounds that prohibit the buying and selling of certain goods or services such as body parts or sexual services. There are objections to the second welfare theorem as well. Equity and efficiency may not be so easy to separate if inequality itself has incentive effects so that a system is efficient only given a distribution of wealth and a different distribution would be more efficient. Additionally, in practice, the second welfare theorem has little significance, because redistributing initial endowments so as to achieve an ethically satisfactory competitive equilibrium poses political and informational difficulties and has huge costs that impair efficiency. The notions of Pareto improvements and Pareto efficiency do not help with the problem of comparing states of affairs that improve some people’s levels of preference satisfaction while harming others. For example, a new road helps some businesses but hurts others located along the path of the previous road. Because benefits and harms fall on different people, the range of social policies on which the Pareto notions can find purchase is severely limited. Normative economics will not be very helpful unless there are other ways of comparing alternative policies.

9.2  Pareto Efficiency, Pareto Improvements, and Benevolence If welfare is measured by preference satisfaction and if the extent to which the preferences of different individuals are satisfied cannot be compared, then how can economists say anything substantial about economic welfare? Minimal benevolence permits economists to judge that Pareto improvements are moral improvements, other things being equal. Most welfare economists endorse an apparently stronger moral claim, which is commonly called The Pareto principle: If X is a Pareto improvement over Y – that is if someone prefers x to Y, and no one prefers y to X, then X is morally better than Y.

The Pareto principle says nothing about how states of affairs X and Y compare when some people prefer X to Y and others have the opposite preferences. If one measures well-​being by preferences, then the Pareto principle implies what John Broome calls “the principle of the personal good”:  “if two prospects are equally good for everyone, they are equally good, and if one of two prospects is better for someone than the other and at least as good for everyone, then it is better” (1989, p. 11). Conversely, the principle of the personal good and the link between preference satisfaction and welfare entail the Pareto principle. Broome’s principle of the personal good also

9.2  Pareto Efficiency, Pareto Improvements, and Benevolence

151

implies the weaker principle of minimal benevolence (that, other things being equal, it is a morally good thing if people are better off). Unlike the principle of personal good, which implies evaluation can sometimes ignore anything except for welfare, minimal benevolence maintains only that in those cases, if any, where there are no other relevant moral considerations, it is good to make individuals better off. To help keep these claims straight, let us restate them slightly as follows: • (Pareto principle) If someone prefers X to Y, and no one prefers Y to X, then X is morally better than Y. • (Welfare is preference satisfaction) X is better for someone than Y if and only if that person prefers X to Y. • (Principle of personal good) If X is better for someone than Y, and Y is not better for anyone than X, then X is morally better than Y. • (Minimal benevolence) If X is better for someone than Y, and Y is not better for anyone than X, then, other things being equal, X is morally better than Y. Given the identification of welfare with preference satisfaction, the principle of personal good and the Pareto principle are equivalent, and both imply minimal benevolence, which is strictly weaker. On the other hand, the policy recommendations made by welfare economics often contain implicitly an “other things being equal clause,” in which case the differences among the Pareto principle, the principle of personal good, and minimal benevolence are moot. There are problems with endorsing all Pareto improvements (as the Pareto principle does) and further problems with endorsing only Pareto improvements (which goes beyond what the Pareto principle advocates). For example, supporting a Pareto improvement may be unappealing when it involves honoring a person’s “nosy” preferences regarding others’ harmless conduct. There are also technical problems combining the Pareto principle with other principles that economists often accept. It has been proven that one cannot consistently endorse Pareto improvements in circumstances of uncertainty if one also holds that individual and social preferences satisfy the axioms of expected utility theory (see Hammond 1983; Seidenfeld, Kadane, and Schervish 1989; Mongin 1995). Unanimity in individual preference rankings may rest on offsetting disagreements both in subjective probability judgments and in preferences among options involving no uncertainty. For example, suppose that unions and large corporations both support some trade agreement, although they have opposing preferences and form different beliefs concerning the likely consequences. The preference and belief

152

Welfare Economics

systems of the two groups satisfy the axioms of expected utility, but there is no coherent set of social preferences and expectations; and the two classes’ agreement with respect to this one preference does not establish that the outcome of the trade agreement will satisfy preferences better than some alternative (see also Levi 1990). Because one of the groups favors the policy only because it has mistaken beliefs about its outcome, there is no reason to conclude that the outcome will be a Pareto improvement. Most importantly, economists need to be able to make comparisons between states of affairs where neither alternative is Pareto superior to the other. Pareto improvements are rare. Economic changes usually involve winners and losers, and it is not a matter of social indifference who wins and who loses. Suppose that there is a single consumption good (bread, say) in some fixed quantity and that everybody prefers more rather than less of it. Then every distribution of bread that exhausts the bread supply is Pareto efficient, and none of these distributions is a Pareto improvement over any other. Moreover, R may be Pareto optimal and S may be suboptimal without R being a Pareto improvement over S. Suppose there are only two people, A and B, and 10 loaves of bread to distribute. A Pareto efficient allocation that gives 7 loaves to A and 3 to B is not a Pareto improvement over the (suboptimal) allocation that wastes 2 loaves and gives 4 to both A and B, and the Pareto principle says nothing about which is better. Very few economic states can be ranked in terms of the relationship of a Pareto improvement. If R is a Pareto improvement over S, then nobody can prefer S to R. If all the individuals in R except for Donald had triple the income in R that they would have in S and Donald had one penny less, then (assuming that Donald for reason of that penny prefers S) R is not a Pareto improvement over S. Notice that with just a tiny bit of redistribution R could be transformed into another state R′ that would be a Pareto improvement over S. This point, which is crucial to cost–​benefit analysis, will be discussed later in this chapter. The Pareto principle and the principle of personal good nevertheless have some real ethical appeal because satisfying preferences arguably has something to do with promoting well-​being and promoting well-​being is a good thing. Moreover, these principles have some important consequences. For example, consider the following potential exchange situation between individuals A and B:  A has a Toyota and B has a Ford. Both A and B value their own vehicle but would prefer to exchange. Exchanging is the “cooperative” solution, and it is a Pareto improvement over the initial allocation. Yet unless both parties can count on being protected from non-​ Pareto-​improving outcomes (such as having their car stolen by the other),

9.2  Pareto Efficiency, Pareto Improvements, and Benevolence

153

they may hesitate to interact at all. For if A could have both the Toyota and the Ford, then that would be best for A and worst for B, and it would be worst for A if B could have both. Property rights prevent either from forcibly appropriating what belongs to the other, thus enabling the exchange to arise voluntarily (Hardin 1988, ch. 3). The Pareto notions help one to understand the contribution that property rights and exchange play, and they permit one to see the contribution property rights make to bringing about and sustaining market outcomes. The outcomes of interactions on perfectly competitive markets are Pareto efficient. But real market outcomes often fail to be efficient, particularly in the case of externalities. As we saw in the introductory discussion of the World Bank memorandum and in the discussion of the limits to markets, sometimes the costs and benefits of an agent’s actions do not fully register as costs or benefits for that agent or for the other exchanging party. Those who pollute the air or deplete the fish population in a lake often need not take the costs imposed on others into account in their private calculations of economic costs and benefits. Nor can someone who builds a lighthouse collect from all those who benefit from it. The benefits and harms that one’s actions cause for which there is no market are externalities.1 Market interactions in the face of externalities are not necessarily Pareto improving. One solution to the problems posed by externalities canvassed by Coase (1960) is a more refined assignment of property rights so that either (a) polluters will have to compensate those with a right to clean air or (b) those who want clean air will have to pay polluters not to pollute. Since transaction costs –​that is, the costs of finding the parties one needs to bargain with and striking and enforcing these bargains –​are often prohibitive, clarifying the assignment of rights does not solve all the problems. Government provision of collective goods (such as lighthouses), government restrictions (such as hunting and fishing limits or limits on pollution), and government taxes or subsidies can mitigate suboptimal outcomes that may result when there are externalities. Most economists regard taxes and subsidies as preferable to restrictions or mandates, because taxes and subsidies are arguably Pareto superior to less 1 More precisely, the externalities we discuss are “nonpecuniary”: the benefits or costs that an agent, Jack, imposes on parties with whom he does not interact are not conveyed by the market. Suppose, in contrast, I figure out how to make stockings more cheaply and sell them to the public at a lower cost; then my action has an effect on parties with whom I have no direct interaction, namely, other stocking manufacturers, and the costs of my actions to them do not show up as costs to me. These latter “pecuniary externalities” do not threaten efficiency.

154

Welfare Economics

flexible requirements and they permit a greater range of individual choice. However, taxation expresses a different attitude toward behavior than does legal prohibition. Regardless of considerations of efficiency, no one would propose taxing murderers rather than prohibiting murders. To tax rather than to fine pollution is to treat pollution as socially and morally acceptable, as it sometimes is. So long as people need power generated by fossil fuels, those who own and operate power plants cannot avoid discharging carbon dioxide (the major culprit in global warming) into the air. Indeed, all of us who have not given up breathing join them in this pollution. Prohibition is out of the question, and taxes on greenhouse emissions or markets in emission rights may be the most sensible way to reduce them. (The remarkable papal encyclical (2015) on climate change is unfortunately blind to the positive role that markets can play toward reducing human damage to the environment.) Other sorts of pollution are not socially acceptable, and some pollution is criminal. Those who dump known poisons into people’s drinking water should be punished rather than taxed (Kelman 1981). Efficiency does not trump all other moral considerations.

9.3  How Welfare Economics Narrows Normative Questions An example clarifies how the Pareto notions may be put to work, and at the same time shows how a standard economic treatment of policy questions narrows the morally relevant issues. Consider the question of whether government should provide health insurance to its poorer citizens directly (“in kind”) or whether it should give them cash that they can use to purchase health care or health insurance. Many economists would offer a simple argument for why the government should supply cash: consider the graph of Figure 9.1. The vertical axis represents quantity of health insurance (h) while the horizontal axis represents the quantity of some composite consumption commodity –​that is, of everything else consumed (ee). Points in the plane represent possible bundles of goods consumed by some individual, whom we call “Alice.” These bundles contain quantities of health insurance qh and the “everything else” commodity, qee. As one moves northeast, the amount of both health insurance and of ee (everything else that is consumed) increases. If Alice prefers having more rather than less of both health insurance and ee, then northeast is also the direction of preference. The curving lines I1, I2, and I3 are “indifference curves,” which are drawn through points (bundles of health insurance and everything else) among which Alice is indifferent. For example, Alice does not care whether she consumes

9.3  How Welfare Economics Narrows Normative Questions

155

qh B2

F

C2

B1

C3 I3

I2 C1

G

I1

B3 qee

Figure 9.1.  Cash versus in-​kind benefits.

bundle F, which contains lots of health insurance and comparatively little of everything else or bundle G, which contains little health insurance and lots of everything else. Fundamental principles of the standard theory of consumer choice imply that indifference curves have the shape shown here and that they do not cross. Since some points on I2 are northeast of points on I1, it follows that Alice prefers every bundle on I2 to every bundle on I1. If Alice is able to change her consumption from any point on I1 to any point on I2, then her preferences are better satisfied and she is better off. The straight lines sloping downward to the right represent three different budget constraints on Alice. Given her income and the prices of health insurance and ee, Alice can afford only those consumption bundles on or below the line B1. Since points above B1 are unaffordable and since, for every point below B1, Alice prefers some point on B1, Alice will consume a bundle on her budget line –​and indeed (apart from some technical complications) she will consume C1, which is the point on the indifference curve that is tangent to her budget line. Suppose initially that Alice’s consumption is C1 and that the government decides to subsidize health insurance. This does not change the amount of money Alice has available to spend, but because it makes health insurance cheaper, her budget constraint will now look something like B2, and her consumption will be C2 on indifference curve I2. Since points on I2 are northeast of points on I1, Alice prefers to be at C2, where she consumes

156

Welfare Economics

more health insurance and slightly more of everything else than she did at C1. (In this case the health-​insurance subsidy led Alice to consume more of everything else, but this result depends on the shape of the indifference curves and does not hold in general. A health insurance subsidy could lead Alice to consume less of ee.) Now suppose that instead of subsidizing health insurance the government gave Alice enough cash that she could, if she wished, purchase C2. If we ignore the effect that an increased demand for health insurance may have on health-​insurance prices, a cash allowance to Alice does not change the relative prices of health insurance and ee. So it will shift her budget constraint to B3, which is parallel to her original budget line B1 and passes through C2. Given her preferences, Alice will not, however, consume C2; she will instead consume C3 on indifference curve I3. The cash payment must satisfy Alice’s preferences at least as well as the in-​kind health-​insurance benefit, and it will typically satisfy her preferences better. Giving individuals cash rather than health insurance is apparently a Pareto improvement over the in-​kind benefit: no one will prefer the in-​kind provision, and many will prefer the cash payment. The apparent advantages of the argument for cash payment do not stop there. It is generally cheaper to give people cash than to give them health-​ insurance benefits, because no government bureaucracy is needed to process claim forms or to investigate medical frauds. (Private health insurance companies have their own bureaucracies, but that is a subject for another occasion.) Since cash benefits are cheaper to administer, the government can satisfy the preferences of recipients at least as well with cash benefits as with in-​kind provision and at a lower cost to taxpayers, or the government can provide more of what people prefer. This argument would not show that cash makes Alice better off than the in-​kind provision of health insurance if her preferences are not a good guide to her welfare. In representing Alice’s preferences by indifference curves, economists have already assumed that her preferences are complete and transitive. But Alice’s preferences need not be self-​interested. Perhaps a desire to help refugees from wars is a prominent part of the other things Alice prefers. In that case, although cash will enable Alice to satisfy her preferences better, it may leave her worse off than benefits in kind. Alice may also have false beliefs about the probability that she will become ill and for that reason purchase too little health insurance, or she may be improvident and fail to take seriously her need for health insurance. The latter case is particularly troublesome, since most people are not prepared to leave Alice to suffer and die if she fails to purchase insurance. This benevolent refusal

9.3  How Welfare Economics Narrows Normative Questions

157

to abandon those who are suffering or at risk of death means that those who do not purchase health insurance receive it anyway. There are a number of other problems with the argument for cash rather than in-​kind benefits. Taxpayers may have preferences for certain kinds of consumption patterns among the recipients of assistance. Consequently, the welfare of taxpayers may be increased by resorting to in-​kind provision, even if diminishes the welfare of the recipients. Many would object, on grounds of liberty, to constraining the consumption decisions of the needy in order to please taxpayers. However, this line of argument departs from the framework of welfare economics by appealing to the value of freedom. A second criticism of cash provision that can be cast within the standard welfare economics framework appeals to the difficulty of monitoring claims for assistance. Cash benefits create incentives to feign illness, whereas waiting in line at a medical clinic is not very enticing. However, the central arguments in favor of in-​kind provision cannot be expressed explicitly in a framework that focuses exclusively on welfare as the satisfaction of preferences. After all, health care possesses an “urgency” (Scanlon 1975) that may be lacking in a person’s other objects of desire. Other arguments are paternalistic. Regardless of whether the poor want more health insurance, policy makers believe that more health insurance would be good for them. The paternalist case for in-​kind benefits cannot be dismissed by supposing that what people most prefer is automatically what is best for them. This is especially the case when we are thinking about health care for children. Whether or not one is inclined to support some paternalistic policies (such as mandatory seat belt laws or prohibitions on recreational drugs), it is important to recognize the morally significant differences between judging that someone is acting against her interest, persuading her to act differently perhaps through the use of incentives, and coercing her. Paternalism takes many different forms, not all equally objectionable. It may sometimes be good social policy to give people cash with no strings attached,2 although we doubt that this is a good way to address health-​care 2 There is currently an active debate about the merits of replacing foreign aid in the form of development projects or specific consumption goods with direct cash gifts to the poor in less developed countries. See, for example, www.givedirectly.org​. The suspicion of the more privileged that the poor will not use cash windfalls wisely might be allayed by attending to the following exchange involving Alfred P. Doolittle, Colonel Pickering, and Henry Higgins, in George Bernard Shaw’s Pygmalion and duplicated in the musical My Fair Lady. Doolittle is importuning Pickering and Higgins for money. HIGGINS. I suppose we must give him a fiver. PICKERING. He’ll make a bad use of it, I’m afraid.

158

Welfare Economics

needs. Whether there should be any strings attached to benefits needs to be considered rather than assumed away by supposing that whatever people prefer is automatically best for them. These aspects of the discussion are poorly handled when the analysis is confined to Paretian welfare concerns. Apart from a nod toward freedom, the argument in favor of cash benefits remains within the terms set by orthodox welfare economics. There is no mention of needs, of the presuppositions of individual dignity, of opportunity, of rights, or of fairness. There is no concern with the moral reasons that make individuals willing to pay taxes to provide such benefits. Are people motivated by a general concern to satisfy the preferences of others, or do they instead see themselves as obligated to help others in need? Might they regard people as having rights to food or medical care, which justify taxing others? These questions are hard to pose within the framework that mainstream economists typically employ.

9.4  Cost–​Benefit Analysis In Chapter 8, we pointed out the limitations of measuring welfare by preference satisfaction. In this chapter we have criticized economists for relying so heavily on the Pareto concepts mainly on the grounds that focusing only on welfare unreasonably narrows the terms in which institutions and policies can be evaluated. What has bothered normative economists, however, is the fact that the scope of the Pareto criteria is so limited. There is almost always some person who prefers that the government not intervene to build a dam, save an endangered species, or clean up a river. Even though externalities often lead to unfortunate outcomes, the feasible remedies are rarely true Pareto improvements. Theorists have struggled to increase the range and force of welfare economics in various ways. The first way relies on the two theorems of welfare economics. These purportedly show that perfect competition is ethically attractive and that other moral concerns can be addressed separately. But the fact that perfect competition guarantees Pareto efficiency is, as we have DOOLITTLE. Not me, Governor, so help me I won’t. Don’t you be afraid that I’ll save it and spare it and live idle on it. There won’t be a penny of it left by Monday: I’ll have to go to work same as if I’d never had it. It won’t pauperize me, you bet. Just one good spree for myself and the missus, giving pleasure to ourselves and employment to others, and satisfaction to you to think it’s not been throwed away. You couldn’t spend it better. HIGGINS [taking out his pocket book and coming between Doolittle and the piano] This is irresistible. Let’s give him ten. [He offers two notes to the dustman.] DOOLITTLE. No, Governor. She wouldn’t have the heart to spend ten; and perhaps I shouldn’t neither. Ten pounds is a lot of money: it makes a man feel prudent like; and then goodbye to happiness.

9.4  Cost–Benefit Analysis

159

Table 9.1.  Potential Pareto improvements

Jack Jill

Status quo

Proposed reform

Reform with compensation

10 coins, 10 food units 15 coins, 8 food units

20 coins, 7 food units 10 coins, 9 food units

18 coins, 7 food units 12 coins, 9 food units

seen, of little moral importance. A slave system can be efficient or inefficient; so can an egalitarian society. Another reason not to be overly impressed with the welfare theorems is that perfect competition is unattainable. Real world markets are rife with asymmetric information, monopoly power, and other imperfections. Furthermore, the efficiency of perfect competition does not imply that eliminating particular impediments to actual competition increases efficiency. Removing one market imperfection while leaving others intact may make a given state of affairs worse rather than better (Lipsey and Lancaster 1956). A more practical way of extending welfare economics –​one that is central to regulatory review in the United States (and to some extent other countries) –​is cost–​benefit analysis. Cost–​benefit analysis derives from essays by John Hicks (1939) and Nicolas Kaldor (1939), although the idea goes back to Pareto and Enrico Barone at the beginning of the twentieth century. Consider two economic outcomes X and Y. There are many different moral comparisons people might make of them. One morally significant difference between X and Y might be distributional, as in the case of the ten units of bread to be divided between two people. Another way that economic outcomes may differ is in the quantity of economic benefits to be distributed –​that is, in their capacity to satisfy preferences. To understand the idea, consider Table 9.1 describing two possible outcomes in a two-​person economy in which there are only two commodities, units of food that cannot be transferred from one person to another and gold coins. Suppose Jack cares a good deal more about money than food and hence prefers the reform to the status quo, while Jill prefers the status quo. Without making interpersonal comparisons of preference satisfaction there is no way to know whether the new policy increases welfare. Suppose, however, that both individuals prefer the reform with compensation to the status quo. If welfare is preference satisfaction, then since both prefer the reform with compensation to the status quo, there must be more total welfare in the compensated reform than in the status quo. The uncompensated reform does not necessarily contain more welfare, but there is, Hicks (1939) and Kaldor (1939) argue, an unambiguous increase in economic benefits  –​a

160

Welfare Economics

“net benefit” in carrying out the uncompensated reform. Jack, the “winner” in the reform, would still prefer the reform if he had to pay Jill enough (namely, two coins) that she would no longer object to the reform. Kaldor and Hicks interpret these facts as showing that the reform has a greater capacity to satisfy preferences. With nothing other than a redistribution of the coins, preferences could be unambiguously better satisfied. Just comparing the reform to the status quo, there is no way to say whether there would be a net benefit in carrying it out, but the compensation test appears to decide the question. In the view of Hicks and Kaldor, welfare economists are in no position to pass moral judgments on economic distribution –​but they do not have to! The increase in efficiency, here understood as the capacity to satisfy preferences, is independent of distribution. Economists should be concerned to enlarge the pie and then leave the ethics of its division to politicians and moral philosophers. Hicks and Kaldor envision a separate dimension of purely economic evaluation. According to Hicks and Kaldor, in order to determine whether X is more economically efficient than Y –​that is, whether X has a greater capacity to satisfy preferences than Y –​one need only determine whether X is a “potential Pareto improvement” over Y. Policy X is a potential Pareto improvement over Y if there is some way of redistributing the goods available in X that makes X an actual Pareto improvement over Y. So, in the simple two-​commodity example the distribution of 20 coins and 7 units of food to Jack and 10 coins and 9 units of food to Jill is a potential Pareto improvement over the distribution of 10 coins and 10 units of food to Jack and 15 coins and 8 unit of food to Jill, because it would be possible to redistribute the coins Jack and Jill have in the reform so as to achieve an actual Pareto improvement. One can also describe a potential Pareto improvement in terms of the possibility of compensation: if X is a potential Pareto improvement over Y, then it is possible for the winners in a change from Y to X to compensate the losers. Whether the winners can compensate the losers needs to be operationalized. One overly simple way (and the only way we shall discuss here) is in terms of willingness to pay. If the amount that winners would be willing to pay to bring about a policy is larger than the amount that losers would need to be compensated to accept the policy, then one has reason to judge that the policy is a potential Pareto improvement over the status quo. Alternatively (and not necessarily equivalently), one could ask losers how much they would pay to retain the status quo and how much winners would require in compensation to forgo the change. These tests

9.5  Objections to Cost–Benefit Analysis

161

purportedly determine which alternative possesses a greater capacity to satisfy preferences. The notion of a potential Pareto improvement thereby provides one possible justification for cost–​benefit analysis (Mishan 1971, 1981). The difference between what the winners in a policy would pay to institute the policy and what the losers require in compensation is a measure of the net benefit of the policy; and the policy with the largest net benefit is best, other things being equal. Dumping toxic wastes in LDCs without paying compensation could be a potential Pareto improvement. Whether or not some potential Pareto improvements might still turn out to be bad as a result of their distributional consequences can be left for others to consider. Furthermore, if the problems with a proposed policy are distributional, then, or so it would seem, so are the solutions. As Hicks argues, judgment concerning economic efficiency should be independent of distributional concerns: Whether or not compensation should be given in any particular case is a question of distribution, upon which there cannot be any generally acceptable principle… If measures making for efficiency are to have a fair chance, it is extremely desirable that they should be freed from distributive complications as much as possible. (Hicks 1939, pp. 711–​712; see also Kaldor 1939, p. 550)

Notice that the notion of efficiency has shifted. Potential Pareto improvements are not actual Pareto improvements, and they do not necessarily increase welfare as measured by preference satisfaction. Potential Pareto improvements purport only to increase an economy’s capacity to satisfy preferences. But, as we shall argue, they may not even do this, if the needed distribution does not take place.

9.5  Objections to Cost–​Benefit Analysis The progenitors of cost–​benefit analysis saw potential Pareto improvements as improvements in productivity, efficiency, or real income, and they hoped to separate the study of these uncontroversial economic benefits from controversial questions concerning distribution. Unfortunately, there is no way to separate questions of efficiency in satisfying preferences from questions concerning distribution. Indeed, as Paul Samuelson pointed out more than sixty years ago, even a preference for policies that yield actual Pareto improvements is not, in general, free of distributional commitments. One objection to the simple form of cost–​benefit analysis we have presented is that the existence of a net benefit does not establish that there has been a real increase in the capacity to satisfy preferences. A net benefit does

Welfare Economics

162 UR

T1 S′

Z Z′

S T2

UP

Figure 9.2.  Interdependence of equity and efficiency.

not imply a bigger “pie.” To see why, consider Figure 9.2, which we borrow with some modifications from Samuelson (1950). The utilities of two representative individuals, Rachel (R) and Peter (P), are measured along the two axes, and the two curves represent the possible utility frontiers (maximal pairs of utilities) depending on whether technology T1 or technology T2 is employed. T1, which is the status quo (perhaps in 1920), involves a rail transport system with few goods or people carried on highways. T2 involves an extensive road system like the one currently in use in the United States. The utility status quo in 1920 is, let us suppose, point S, and the result of changing to an automobile technology will (other things being equal) result in the utilities for Rachel and Peter represented by point Z. Z is not a Pareto improvement over S since Peter is worse off, but Z is a potential Pareto improvement over S because Rachel can pay compensation to Peter and the economy can move along the T2 curve to Z′. However, one cannot conclude, as Hicks and Kaldor hoped, that T2 is more efficient than T1 –​or that it involves a greater capacity to satisfy preferences, a larger real income, a bigger “pie” –​because S here is also a potential Pareto improvement over Z: one can move from S along the T1 curve to S′, which is an actual Pareto improvement over Z.3 Furthermore, even if the choice were The obvious way that one might try to fix this problem by stipulating that there is an increase in the capacity to satisfy preferences if and only if R is a potential Pareto improvement over S and not vice versa does not work. As Gorman (1955) shows, this proposal leads to intransitive rankings of social states. 3

9.5  Objections to Cost–Benefit Analysis

163

between S and Z′ rather than between S and Z, so that what was at stake were an actual Pareto improvement rather than merely a potential Pareto improvement, one would still not be justified in concluding that T2 is more efficient than T1. In just the same way, the efficiency of relocating polluting industries in LDCs discussed in the World Bank memorandum depends in large part on the lopsided distribution of wealth. Indeed, one can interpret R and P as standing for “rich” and “poor” and take T2 to be the current technology (with polluting industries located in rich countries), Z the current level of preference satisfaction for representative members of rich and poor countries, and S′ the Pareto improvement that relocating the polluting industries will make possible. Though S′ is a Pareto improvement over Z, it cannot be compared to Z′, which could be achieved by redistribution within Z without shifting polluting industries. Only when one utility possibility curve is inside the other can one make a “pure” efficiency comparison that does not take a particular distribution for granted. A second problem with relying on cost–​benefit analysis to guide policy is that policy depends on the unreflective preferences that analysts deduce from people’s market choices. But some preferences, such as preferences for communities free of urban sprawl, are hard to signal when one buys groceries, cars, or even homes. Furthermore, people’s preferences for policies respond to arguments and may be different after public debate than they were before: substituting cost–​benefit analysis for public deliberation means that people’s preferences are never subjected to such challenges. In addition, it is questionable whether decisions about roads or schools or tanks should be equally responsive to all preferences, whether or not they are weighted by wealth and hence reflected in willingness to pay. As already argued repeatedly, only some of the preferences people express are a guide to their well-​being. Should a decision about the size of the military budget, for example, respond only to the strength of people’s preferences (or to their willingness to pay), or should it rather look to the reasons for their preferences? In circumstances of uncertainty, people’s preferences among policies depend on their beliefs about the consequences of the policies as well as on their ranking of outcomes. When people have mistaken beliefs about the constitution of the exhaust from the factory down the road, their willingness to pay to avoid breathing it will be an unreliable indicator of their true preferences, let  alone of the exhaust’s welfare consequences. People may not know the consequences of alternatives and hence may not be able to

164

Welfare Economics

identify which alternative they would prefer if they did know the consequences. Economists often finesse the problems of uncertainty by supposing that individuals possess subjective probability distributions over all the possible outcomes. But to suppose this involves extreme idealization, and there is little justification for respecting preferences based on largely fictitious probability distributions. A third problem with simple cost–​benefit analysis is that its measurement usually relies on people’s willingness to pay, and in addition to reflecting preferences among outcomes, willingness to pay depends on expectations concerning what is appropriate to purchase and for what price. One would obtain very misleading information concerning the relative value of a car and a spouse by investigating how much someone would be willing to pay for each! Willingness to pay, like the amount of money one would require in order to consent to an unwanted change, also depends on wealth. Since preferences in simple cost–​benefit analysis are weighted with dollars and since the poor have fewer of these, their preferences count for less (Baker 1975). If the benefits of a policy accrue to the poor and the costs to the rich, what the poor would be willing to pay for the new policy might not be enough to compensate the rich. The policy might nevertheless have had large net benefits if there had been a more even distribution of wealth, and someone willing to make interpersonal comparisons might be convinced that the welfare of the poor would increase more than the welfare of the rich would decrease. For this reason (among others), most proponents of cost–​benefit analysis regard the version considered here as too simple and argue that willingness to pay should have “equity weights” (Adler 2015; Fankhauser et al. 1997). Finally, simple cost–​benefit analysis –​like other methods of evaluation that employ the Pareto criteria –​ignores questions of justice; and, unlike the Pareto principle, a practice of maximizing net benefit often endorses policies that make some people worse off. The compensation considered is only hypothetical. Some people win and some lose. When there are winners and losers, questions of fairness are crucial. If each policy had different winners and losers  –​so that, in the long run, every individual won as often as she lost –​then the unfairness of individual policies taken separately might wash out. But without equity weights, the bias built into cost–​benefit analysis against the preferences of the poor suggests that the unfairness will not wash out. Exactly those people whom policy makers should be most concerned to protect are those who are most likely to be harmed.

9.6  Cost–Benefit Analysis and Social Welfare Functions

165

9.6  Cost–​Benefit Analysis and Social Welfare Functions Although cost–​benefit analysis arose as an effort to extend the reach of welfare economics in the face of skepticism about making interpersonal comparisons, the problems we have documented suggest that it cannot be justified that way. Except when one policy is Pareto superior to another, there is still no way without making interpersonal comparisons to judge whether one policy promotes welfare better than another. That realization does not mean that cost–​benefit analysis is worthless, because it is possible to defend it as a practical way of implementing what economists call a social welfare function –​that is, as operationalizing some method of appraising social states in terms of interpersonally comparable measures of individual welfare. A social welfare function can be represented abstractly as W = f(w1, w2, …, wn), where wi is the welfare of the i’th individual in the society. Social welfare functions can be sensitive to the distribution as well as levels of individual welfare, but they are restrictive in asserting that social welfare depends only on individual welfare. Utilitarianism yields one of many possible social welfare functions, and some economists have defended cost–​benefit analysis as a way to implement a utilitarian policy of maximizing total welfare. Here is how the basic ideas of cost–​benefit analysis are introduced and defended in one text: The only basic principle is that we should be willing to assign numerical values to costs and benefits, and arrive at decisions by adding them up and accepting those projects whose benefits exceed their costs. But how are such values to be arrived at? If we assume that only people matter, the analysis naturally involves two steps. First, we must find out how the decision would affect the welfare of each individual concerned. To judge this effect we must ultimately rely on the individual’s own evaluation of his mental state. So the broad principle is that “we measure a person’s change in welfare as he or she would value it”. That is, we ask what he or she would be willing to pay to acquire the benefits or to avoid the costs… The second step is to deduce the change in social welfare implied by all the changes in individual welfare. Unless there are no losers, this means somehow valuing each person’s £1… And if income is not optimally distributed [then] most economists would argue that it should be redistributed by cash transfers rather than through the choice of projects. But what if we think that cash will not be distributed, even if it should be? Then we may need to value the poor person’s extra £1 more highly than the rich person’s. (Layard and Glaister 1994, pp. 1–​2)

In this text, Layard and Glaister claim that the “only basic principle” is that the considerations bearing on policy evaluation must always be comparable and quantifiable. This is a basic principle –​and a controversial one –​but

166

Welfare Economics

it is hardly the only one, and Layard and Glaister immediately assume another: that only people matter. This second principle is controversial, too, and it has been explicitly challenged by those concerned with the consequences of human actions for animals (Singer 1975) and the nonsentient environment (Stone 1996). Layard and Glaister then slide from the premise that only individual people matter to the conclusion that only the welfare of individual people matters. They do not say why one should believe that the only thing that matters about individual people is their welfare. What about, for example, their rights and liberty? The next sentence asserts that we should accept the individual’s own evaluation of her welfare (which Layard and Glaister equate with a mental state, despite the official identification of welfare with preference satisfaction). From comments they make later, it seems that they would argue for this assertion on the basis of antipaternalistic and democratic principles. Their reliance on such principles suggests that Layard and Glaister do not, in fact, believe that only welfare matters. The final paragraph then equates the net benefits of projects with their aggregate welfare consequences. Of course, these cannot be determined without making interpersonal welfare comparisons. If one assumes that willingness to pay can serve as an interpersonally comparable measure of individual welfare, then projects whose benefits exceed their costs will increase total welfare. As Layard and Glaister themselves point out, this assumption is implausible, and they suggest that people’s willingness to pay needs to be adjusted so that each £1 is equally valuable to everyone. Without distributional weights, cost–​benefit analysis is not a plausible way to operationalize utilitarianism. Consider a change in policy that benefits rich people and harms poor people. Suppose that the rich are willing to pay a small amount more than the poor demand in compensation. So this change has a net benefit. But given diminishing marginal utility, the increase in welfare to the rich will be smaller than the decrease in welfare to the poor. A social welfare function that endorses cost–​benefit analysis in this case would place a greater weight on the welfare of the rich than on the poor (Le Grand 1991, ch. 3). Unweighted cost–​benefit analysis may also give you cycles of “improvements” returning to the starting point (Blackorby and Donaldson 1990). Only if one imposes distributional weights on the willingness to pay of different individuals is it possible to use cost–​benefit analysis to operationalize utilitarianism or any other ethically plausible social welfare function. But note that deciding on equity weights and an underlying social welfare function from which these weights should derive reintroduces many of the normative questions that economists hoped to avoid.

9.7  Putting Cost–Benefit Analysis to Use

167

Regardless of the justification for basing social policy on willingness to pay, weighted or unweighted, it is in practice costly to ask people how much they would pay or how much compensation they require, and their answers may be neither truthful nor accurate. Welfare economists have accordingly devised ingenious methods of gleaning such information from data on prices and quantities traded or sometimes from responses to surveys. Much of cost–​benefit analysis is devoted to devising, criticizing, and improving methods of imputing costs and benefits.

9.7  Putting Cost–​Benefit Analysis to Use Cost–​benefit analysis has played and will continue to play a significant role in policy evaluation. Indeed the U.S. government requires that cost–​benefit analyses be conducted when it issues significant regulations governing product safety or environment protection. That leaves us with a difficult question: what role can cost–​benefit analysis reasonably play in evaluating policies given the limitations we have identified? Given (a) the wealth bias of simple cost–​benefit analysis, (b) the absence of a strong rationale for relying on unweighted willingness to pay as an indicator of the capacity of policies to satisfy preferences, and (c) the need for some distributional principles, cost–​benefit analysis without distributional weights is of limited use in many circumstances. More refined versions of cost–​benefit analysis that employ equity weights or explicitly incorporate assumptions about the social welfare function are more promising, despite the greater difficulties in formulating and applying them. However, even they respond only to some of the considerations that should guide the evaluation of policy and, moreover, cost–​benefit analyses depend on the alternatives that are on the table. Some policy alternatives with the greatest net gain may be missing from the analysis. It thus seems wise to treat even more refined versions of cost–​benefit analysis as no more than a partial guide to policy evaluation, rather than as a way to find the “right” answer, all things considered. A further important limitation on the applicability of cost–​benefit analysis is its reliance on preferences (typically in the form of willingness to pay) as the measure of individual welfare. As discussed at some length in Chapter  8, people’s preferences among alternatives can be regarded as a reliable guide to their welfare only if they are competent evaluators and their preferences are self-​interested and do not depend on false beliefs. Whether people are in general competent evaluators is questionable, but let us be optimistic and assume that they generally are. It does not follow

168

Welfare Economics

that they are equally competent at evaluating every kind of alternative, and their choices are not always self-​interested or informed by correct beliefs. With respect to some alternatives, the assumptions that permit economists to measure welfare by preferences are reasonable approximations to the truth. With respect to other alternatives, people’s beliefs are woefully off the mark, their preferences are not self-​interested, or their evaluative competence is dubious. For example, when considering environmental policies, including policies for coping with climate change, economists should recognize that the preferences people express (and may reveal in their choices) are often not self-​interested, and their beliefs about the pace and effects of climate change tend to be woefully uninformed. On top of that, psychologists and behavioral economists have provided strong evidence that people are not very good at evaluating risks of outcomes with low probabilities. A naïve application of cost–​benefit analysis that aims to assess climate change policies by examining willingness to pay for alternative policies would not be a good way of figuring out which policy most enhances people’s welfare. In light of the objections to simple cost–​benefit analysis without distributional weights, it would be unreasonable to apply it indiscriminately even as a rule of thumb. A cost–​benefit rule will likely be less worrying in cases where the winners and losers are close to equal in ability to pay, where uncertainty about consequences is limited, where people’s preferences reflect their judgment concerning which policies best serve their interests, and where most people understand the kinds of reasons recommending different policies. A cost–​benefit analysis might thus provide a reasonable basis for deciding whether to put a baseball field or a soccer field in a public park or where best to locate a rail line when environmental damage is not an issue. It would be harder to defend using such analysis to decide whether a public park or a free clinic for destitute people were a better use for a particular piece of land. Politicians and the public at large should be skeptical of the policy assessments generated by the use of cost–​benefit analysis. There are many ways these analyses can go wrong and many ways that biases can intrude into the methods of determining net benefit. But suspicion is not contempt, and suspicions can be overcome. The results of a cost–​benefit analysis can be informative, but they can never show that a particular policy is the “scientifically” right one to adopt. On issues of sufficient import to warrant serious public deliberation, the findings of cost–​benefit analyses, however relevant and important they may be, should not dictate policy over other considerations that citizens deem relevant.

9.8  Welfare Economics

169

9.8  Welfare Economics Although economists might describe welfare economics in modest terms, we have argued that it represents an ambitious philosophical project that has implications regarding many of the most fundamental questions in moral philosophy. And while it is true that welfare economists can avoid committing themselves to any philosophical theory of welfare, doing so comes at the cost of limiting the applicability of their models to circumstances in which it is reasonable to regard people as well informed, self-​interested, and evaluatively competent agents. If one evaluates policies exclusively in terms of their consequences for individual welfare, and if one takes welfare to be the satisfaction of preferences, and if one denies that it is possible to compare how well off different people are, then it seems there is little that can be said about which policies and institutions are better. Orthodox welfare economists are in precisely this uncomfortable position. Jettisoning these self-​imposed shackles, they have outlined a formal framework –​the social welfare function –​that has the potential to remove these constraints, and they have managed to salvage a practical method of evaluation –​cost–​benefit analysis. Because the fundamental evaluative apparatus of contemporary normative economics is so narrow, cost–​benefit analysis rests on shaky and incomplete foundations, and its applicability is limited. A central problem, which is widely acknowledged in principle but often ignored in practice, is that the separation that Hicks and Kaldor envisioned between questions concerning efficiency and distribution  –​between the size of the pie and the way it is sliced –​cannot in general be accomplished. Because distributions can influence the willingness to pay by which cost–​ benefit analysis weighs preferences, there is no general way to separate questions concerning the capacity to satisfy preferences from distributional questions. One cannot reasonably evaluate policies, institutions, or states of affairs exclusively in terms of their success at satisfying the interpersonally noncomparable preferences of individuals. For, as we have argued, preference satisfaction is a highly fallible indicator of welfare, and values such as freedom, equality, and justice also matter. In the right circumstances, a cost–​benefit analysis with proper distributional weights can contribute useful information to a social decision problem. Consider the earlier example of a town council trying to decide between putting a soccer field or a baseball field in a public park. Suppose that those who favor soccer are no wealthier than those who favor baseball, and the policy choice will not affect the distribution of income. Willingness

170

Welfare Economics

to pay in this case seems to be a reasonable indication of intensity of preference, and it is plausible here to take preference satisfaction as a measure of well-​being. So the town council might reasonably rely on information provided by a cost–​benefit analysis to conclude that the soccer field should be built instead. In circumstances where there are large differences between the wealth of those who favor a policy and those who oppose it; where the consequences will be lasting, serious, and uncertain; where the policy choice raises other moral questions (such as questions of equity); or where costs and benefits are difficult to measure and some have not been captured, the application of cost–​benefit analysis will be much more questionable. But, it is in cases with precisely these features –​cases such as determining which neighborhood should be destroyed for a new highway, or whether a dam that threatens a community’s way of life should be built –​that policy makers often seek the guidance provided by cost–​benefit analysis. In making limited use of the findings of a cost–​benefit analysis in cases such as these, policy makers should insist on equity-​weighted results, and they should be careful not to focus myopically on a single alternative to the status quo. The fact that some alternative provides net benefits is not sufficient for determining whether it should be instituted. Provided that one takes account of distributional presuppositions and consequences and recognizes that cost–​benefit analysis is a highly fallible source of information, such analyses under the limited conditions we have suggested can be of use. And given the absence of alternative measures of benefits and harms that are accurate and unbiased, it will be tempting to rely on cost–​benefit analysis. Though we see no better alternative, we have pointed out a number of objections to the naïve use of cost–​benefit analysis. Policy makers need to choose among policies that benefit some people and harm others, and cost–​benefit analysis provides a framework for doing so. But this framework is imperfect. It cannot substitute for the rational deliberation needed to improve beliefs (which may also change preferences), to explore the full range of policy alternatives, and to articulate and refine the values that are at stake.

Suggestions for Further Reading There is an extensive literature concerning interpersonal comparisons. Elster and Roemer (1991) is an excellent collection. Other notable contributions include Robbins (1935, ch. 6; but for a somewhat different view,

Questions for Study and Discussion

171

see also Robbins 1981), Harsanyi (1977b, ch. 4), Davidson (1986), Gibbard (1986), MacKay (1986), and Hausman (1995). The inconsistency between rationality and the Pareto principle in circumstances of uncertainty is proven and discussed in Hammond (1983), Broome (1989), Seidenfeld et al. (1989), Levi (1990), and Mongin (1995). For critical discussion of the argument in favor of cash over in-​kind provision of benefits, see Hochman and Rogers (1969), Okun (1975), Thurow (1977), Kelman (1986), and Blackorby and Donaldson (1988). For general discussions of cost–​benefit analysis see Mishan (1971, 1981), Sugden and Williams (1978), Ng (1983), Layard and Glaister (1994), Adler (2015), and Boadway (2016). For more sophisticated treatments of compensation and net benefit, see for example Drèze and Stern (1990). Especially via the work of Richard Posner (1979), the compensation criterion has had a major impact on legal theory –​see also Coase (1960), Baker (1975), and Coleman (1984). Coase’s work has spawned a line of thinking that seeks to address externalities by means of clearer rights assignments rather than government action. See Coase (1960), Demsetz (1964, 1967), and Cowen (1988).

Questions for Study and Discussion 1. Can economists defend their exclusive focus on welfare by arguing for a division of labor, whereby they focus exclusively on what will promote welfare and leave questions about other normative considerations, such as equity or freedom, for others to consider? 2. Explain how it is possible for X to be Pareto optimal, while Y is not, yet for X not to be a Pareto improvement over Y. Is it possible for X to be a Pareto optimum and for Y to be a Pareto improvement over X? Is it possible for X and Y both not to be Pareto efficient, but for Y nevertheless to be a Pareto improvement over X? 3. What is the difference between minimal benevolence and the principle of the personal good? Which is more plausible ethically? 4. How important is Pareto optimality? Can you think of circumstances in which an outcome that is not Pareto optimal is better than one that is? 5. What is an externality, and why do externalities lead to inefficiencies? What can be done about externalities? 6. What do you think of the general argument for providing the poor with cash rather than specific benefits such as housing vouchers,

172

Welfare Economics

free health care, and food stamps? What factors determine which approach should prevail in a particular case? 7. What is good about a potential Pareto improvement? 8. List all the criticisms that this chapter makes of cost–​benefit analysis. What do you think is the strongest objection? Are the objections, in your view, sufficient to rule out the use of cost–​benefit analysis? How much does incorporating distributional weighting in cost–​benefit analysis matter in your assessment? 9. What features of cost–​benefit analysis have value in operationalizing utilitarianism, and what features limit its effectiveness for that purpose? 10. One argument for the use of cost–​benefit analysis is that there is no good alternative. Allowing policies to be decided by the give and take of political argument with all the biases introduced by special interest groups is, to say the least, unlikely to be an improvement. So, one might argue, the criticisms of cost–​benefit analysis developed in this chapter are beside the point. What do you think?

PART THREE

LIBERTY, RIGHTS, EQUALITY, AND JUSTICE

Though well-​being is obviously important, other moral considerations, such as freedom, rights, equality, and justice, are also important. Some people risk their lives pursuing these values. Even if economists had no interest in any values except welfare, they would still need to understand these other values in order to comprehend the goals of policy makers and to help devise policies to achieve them. In fact, we suspect that economists care as deeply about freedom and justice as do other people, even if these values are absent from their models. Adam Smith objected to most government interference in the market as diminishing welfare, but at the same time he maintained that to prevent people “from employing their stock and industry in the way that they judge most advantageous to themselves, is a manifest violation of the most sacred rights of mankind” (1776, book IV, c­ hapter 7, part 2). Values such as freedom, rights, and justice may be difficult to understand, but welfare is not a simple notion either. In assessing economic outcomes, processes, institutions, and policies, normative economics should encompass a much wider range of moral concerns than welfare. In Part III we shall be discussing important moral notions other than welfare, under the separate headings of liberty and rights (Chapter  10), equality (Chapter  11), and justice (Chapter  12). Although it is convenient to divide the discussion in this way, these considerations are closely related. For example, different theories of justice have implications for what liberties and rights are important and for the extent to which inequality should be viewed as a moral problem. Our way of classifying the issues is an expository convenience, and readers should not take it too seriously. Readers should also not consider our discussion of other important moral notions exhaustive. For example, the deepest moral commitments of many of those concerned to protect wilderness areas or of those concerned with 173

174

Liberty, Rights, Equality, and Justice

community are not captured in these chapters. In these chapters, we also discuss three kinds of moral theories:  libertarianism (Chapter  10), egalitarianism (Chapter  11), and contractualism (Chapter  12). It is natural to link libertarianism with liberty and rights, but liberty and rights are also important to nonlibertarians. It makes sense to discuss egalitarianism while discussing concepts of equality, but concepts of equality are important in theories of justice that are not egalitarian. It is natural to place the discussions of justice and contractualism in the same chapter, because contractualist theories focus on justice. But justice is important in all moral theories. We want particularly to emphasize that the importance of the concepts introduced in this part extends beyond the importance of the particular theories that characteristically invoke them.

TEN

Liberty, Rights, and Libertarianism

Many people have a passionate commitment to individual liberty or freedom (which we here take to be the same thing). Some object to paternalistic laws (such as those requiring the use of seat belts) as infringements on freedom, regardless of whether these laws increase human welfare. Many favor protecting the freedom of those with unusual lifestyles or unpopular religious convictions even if doing so diminishes welfare. Social deliberation often treats protecting freedom and enhancing welfare as independent goals. It is ironic that normative economics focuses exclusively on welfare, because most economists value individual freedom very highly. When leading economists criticized socialism, for example, they not only questioned whether it is efficient but also argued that economic and political power must be kept separate in order to protect individual liberty (Friedman 1962, ch. 1; Hayek 1944). Economists also value the prosaic liberties that are part of market life such as the freedom to change jobs, to start a business, or to move from place to place within a country. They worry about measures that increase the power of the state. Consider that many economists favor taxes or exchangeable emission rights over direct state regulation of pollution not only because the taxes or exchangeable emission rights are purportedly Pareto superior to the regulation but also because state regulation limits freedom directly rather than via prices and property rights. By increasing the reach of government, such regulations may also threaten freedom indirectly. Similarly, economists often favor cash over in-​kind transfers both because cash transfers are more efficient and because they leave the recipients with more choices. The efficiency case has been a part of “scientific” welfare economics because it has appeared to rely only on uncontroversial moral premises, while the argument in terms of freedom usually has been reserved for “unscientific” essays because its moral premises have appeared to be 175

176

Liberty, Rights, and Libertarianism

controversial. The focus on efficiency has also been fostered by a belief that Pareto efficiency itself promotes liberty, because it values outcomes that fully accommodate the voluntary choices of individuals. But the link between efficiency and liberty is tenuous. For example, consider a problem such as homelessness. Though those who have no place to live are almost always badly off, one cannot appreciate what is at stake for policy without considering what being homeless implies for a person’s effective liberty. Moreover, it is possible to have a Pareto optimal equilibrium in which homelessness is widespread. The concerns about liberty that underpin economists’ normative judgments should be incorporated systematically into their methods of evaluation. Doing so raises problems of definition, of moral justification, and of weighting. In other words: (1) What is freedom? (2) How can claims about the moral importance of freedom be defended? and (3) How can liberties be incorporated along with other values into a coherent scheme of ethical evaluation?

10.1 Freedom Gerald MacCallum (1967) helpfully suggests that “freedom” should be regarded as a relation among three elements: (1) an agent; (2) obstacles, costs, or risks of some kind; and (3) objectives or possible outcomes. An agent is free from some obstacles, costs, or risks to do or become something. An agent is free in some particular regard, when there are no obstacles of a specific kind preventing or discouraging the agent from doing something. For example, Karen may be legally free to purchase bread, but not beer, while not economically free to purchase either, because she has no money. If she is stranded on a deserted island, Karen is not physically free to purchase anything. In order to pin down the meaning of the very abstract term “freedom” it is helpful to think about different obstacles to different actions or outcomes for different agents. Many measures that enhance the freedom of one person simultaneously restrict the freedom of others. For example, a restriction on your freedom to smoke gives me the freedom to enjoy a smoke-​free environment. In this way, questions about freedom require further specification:  whose freedom, and for what purposes? When unspecified, the idea of “freedom” can quickly become empty rhetoric. People are free and unfree in various ways. Some freedoms are important; others are not. And people disagree about which are more important than others. It is worth distinguishing four concepts of freedom. In his influential essay “Two Concepts of

10.1 Freedom

177

Liberty,” Isaiah Berlin argued that social policy should aim to protect “negative” liberties rather than to promote “positive” liberty. One protects negative liberties by placing few social or legal obstacles in the way of people’s efforts to satisfy their preferences. In terms of MacCallum’s three-​ place relation, negative liberty is the freedom of (1)  individuals from (2) constraints that are intentionally imposed by other people (3) to do as the individuals choose. Negative liberty has nothing to do with internal impediments to action such as having poor information about the consequences of an action. Promoting positive liberty (as Berlin understands positive liberty), in contrast, involves reference to a conception of the self as rationally self-​ determining. Positive liberty in this sense is the freedom of individuals from both external constraints and the internal constraints of irrationality, poor information, or mistaken self-​understanding to achieve rational self-​determination. In a famous passage from The Social Contract, Rousseau wrote that anyone who refuses to obey the general will shall be forced to do so by the whole body; which means nothing more or less than that he will be forced to be free… Along with the civil state, man acquires moral liberty, which alone makes him truly master of himself; for the impulse of mere appetite is slavery, and obedience to self-​imposed law is liberty. (1968, p. 64)

The defender of positive liberty sees the agent’s significant objectives not as whatever the agent prefers (whether or not it is “the impulse of mere appetite”) but what is truly in the agent’s interests. When people speak of autonomy, they often have in mind positive liberty, but “autonomy” can also refer to a capacity to govern oneself, a condition of self-​governing, an ideal, or a matter of moral authority (Feinberg 1986, p. 28). The capacity to be autonomous depends on factors such as education, information, and the ability to reflect, evaluate alternatives, and act on one’s choices. The condition of autonomy involves satisfying one’s own preferences rather than having them satisfied by others. It involves actively accomplishing one’s objectives rather than merely enjoying well-​being. Autonomy also involves aspects such as self-​possession, individuality, authenticity, integrity, self-​control, initiative, and responsibility (Feinberg 1986, pp. 32, 40–​44). The term “positive liberty” is also used instead to refer to opportunity or what one might call “effective freedom.” Consider, for example, the contribution that a motorized wheelchair makes to the freedom of someone who cannot walk. It does not increase negative liberty insofar as it does

178

Liberty, Rights, and Libertarianism

not address constraints imposed by other people, but it does increase the disabled person’s effective freedom, because it lessens other obstacles that prevent the disabled person from moving freely. So there are two very different notions of positive liberty:  fully rational ideal self-​determination and, more simply, effective freedom or opportunity. A fourth (“republican”) notion of liberty (Pettit 2014) arguably falls between Berlin’s two concepts. It insists on the importance of negative liberty, but it denies that negative liberty is enough, if the absence of interference depends on the goodwill of others or an easily revocable restraint on the part of government. Slaves whose masters permit them to live freely may possess negative liberty, but they lack republican liberty. The republican notion of freedom resembles the positive liberty Berlin identifies with self-​determination, but its focus is on the social conditions for self-​determination and not on the internal state of the individual chooser. On the republican view, an agent is free to the extent that her negative liberty is secured by appropriate social conditions including rights, laws, and norms. This freedom is often identified with the status of a citizen in a republic. Libertarians (whose views we will discuss at the end of this chapter) focus on a set of negative liberties. They typically regard limitations on choices to be of moral concern only if they arise from the deliberate acts of government or other people to limit what the agent can do. Consider the difference between a person’s not being able to buy a Bible because the government bans the sale of Bibles and a person’s not being able to buy a Bible because he or she lacks the money to do so. Libertarians maintain that only the first situation limits freedom in the relevant sense, because the “obstacle” in the second situation does not arise from actions, policies, or laws aiming to limit what people can do. Similarly, those who maintain that competitive markets always promote freedom do not deny that market outcomes can place barriers in people’s way. Instead, they deny that these barriers, as the unintended consequences of voluntary interactions, are the kind of obstacles that limit negative liberty. The paradigm cases of limits on negative liberty, such as the suppression of religious observances, can also be understood as rights violations. We will later discuss some challenges for libertarians. For now, though, we will note that many people do not have such a narrow view of liberty. For many purposes what matters is people’s effective freedom to do something and not whether the obstacles in their way are the result of a deliberate policy or choice. Berlin himself argued that the barriers that others

10.1 Freedom

179

unintentionally place in one’s way may be morally significant limitations on freedom. For example, should those concerned about freedom not be concerned about the limitations on choice that arise from poverty and homelessness? The extent of an individual’s freedom (in the most general sense of “freedom”) depends on how many possibilities are open to her, how dependent those possibilities are on the choices of others, how easy they are to attain, how important in her plan of life these possibilities are, and how the society at large values these possibilities. Any assessment of the extent of liberty in a society requires in addition that one aggregate the liberties of different persons (Berlin, 1969, p. 130). Assessing people’s freedom involves an amalgam of many disparate elements. Some political philosophers have suggested ways of cutting down the complexity. Rawls, for example, produces a relatively short list of “basic liberties,” including freedom of conscience and freedom of expression. Justice does not require stringent protection for liberty as such, but for basic liberties. Rawls further distinguishes between liberty as a formal notion and the worth of liberty, which depends on having the means to exercise the relevant freedom. Billionaires and beggars are both free to attempt to influence candidates on the issues that matter to them, but the worth of their political liberties is hardly the same. Others have distinguished “formal” from “effective” liberty. Sen’s capabilities can also be made to correspond to morally significant freedoms. For example, if one is literate, then one is free from intellectual obstacles to reading. This usage of the term “free” highlights the important point that if one is willing to see beyond negative liberty, a political regime that countenances widespread illiteracy has much in common from the standpoint of liberty with one that bans books. An analysis of morally significant freedoms for individuals or for groups must evaluate how important various objectives are. It might allow the individual’s preferences to determine the importance of actions and states of affairs (as economists tend to do), but this is as substantive a moral decision as is any specification of the “true interests” of individuals. In addition to determining what matters and how much, an analysis of morally significant individual freedoms must specify what kinds of obstacles are of moral concern. Table 10.1 summarizes the four concepts of freedom we have introduced with the help of the rubric that MacCallum provides. Each of the four notions of freedom we have discussed can be extended to social groups such as trade unions or churches. Where they differ is with

180

Liberty, Rights, and Libertarianism Table 10.1.  Concepts of Freedom

Concept of freedom

Agent

Obstacle

Negative liberty

Individual person

Positive liberty

Individual person

Effective freedom

Individual person

Intentionally Whatever the agent imposed wishes restrictions Internal and external “Real” interests obstacles Any Whatever the agent wishes Will of others Whatever the agent wishes

Republican freedom Individual person

Objective

respect to the obstacles that are of concern and with respect to the objectives that the obstacles block.

10.2  What Are Rights? One might roughly characterize a right as protecting or promoting a way of acting or a way of being treated. A right always implies a liberty or permission. People do not have a legal right to vote if there is a law prohibiting them from voting. But rights are not just liberties or permissions. They also protect the action that the agent is free to do and thereby constrain the actions of others (see Martin, 1993). These permissions and constraints may be embodied in law or convention, or they may have the status of moral demands. Although the details need not concern us, legal analysts following Hohfeld (1919) have recognized that rights involve complex clusters of permissions and constraints regarding the actions of individuals. Rights typically involve both “privileges” for the right-​holder and correlative duties for others. Rights that are the flip side of the duties of others are called “claims.” For example, the legal right of workers to strike after their contract expires implies both a legal privilege or freedom to strike and a claim right –​that is, a legal duty on the part of others not to interfere with workers’ strike activity in certain ways. Rights also involve “powers” and “immunities.” For example, Donald Trump’s right to Trump Tower in New York includes (in addition to the privilege of owning it and claims against others) the power to alienate or terminate his right by selling the building or (improbably) giving it away. Trump’s right to vote, in contrast, is inalienable. He does not

10.3  The Importance of Rights

181

have the same power to alienate his right to vote that he has to alienate his right to Trump Tower. Although he can refuse to exercise his right to vote or he can forfeit it by committing a crime, he cannot sell it or give it away. Just as powers permit one to change one’s claim rights, immunities prohibit others from changing one’s claim rights. For example, occupational safety and health laws provide workers with immunities from efforts by firms to negotiate with workers certain transfers of rights. These laws thereby limit the powers of firms to shape the rights specified in labor agreements. Rights are clusters of claims, privileges, powers, and immunities. Claims are central in what are called “rights,” and privileges are central in what are called “liberties.” But rights are not just claims, and liberties are not just privileges. Legal rights are the most familiar kind of rights. For example, the legal right to quit one’s job consists roughly in there being laws permitting workers to quit and prohibiting others, including the state, from preventing them from doing so. Conventional rights are also straightforward. For example, a company policy may give workers the right to seventeen days of paid vacation. Parents have (or used to have) a conventional right to their children’s respect. Moral rights (including what are called “human rights”) are more puzzling. When they are not acknowledged or respected, in what sense do they still exist? When nineteenth-​century abolitionists maintained that everyone has a right to be free, they were obviously not claiming that slaves already had the legal or conventional right to be free. What were they claiming? One central reason for invoking moral rights is to assess legal and conventional rights. The abolitionists used their claims concerning moral rights to criticize law and custom in the United States for not codifying the right of everyone to be free. Similarly, moral rights have been invoked to criticize the former legal rights of the white minority in South Africa or to urge that in the United States rights to health care be recognized in law or convention. When people claim that an agent has a moral right to do something, they typically imply that the agent ought to have a legal or conventional right to do it (Mill 1863, ch. 5; Sumner 1987). Nonetheless, legal, conventional, and moral rights are distinct. Liberal societies give people the legal right to do many things that people arguably have no moral right to do, such as to lie to one’s friends, sell pornography, or commit adultery.

10.3  The Importance of Rights Rights safeguard people’s interests, and they help ensure people’s control over their choices. Very often these two purposes of rights coincide, since

182

Liberty, Rights, and Libertarianism

people typically try to choose what is in their interest and regard themselves as having an interest in controlling their own choices. So, for example, giving an artist a property right in a painting she has created might serve her interest insofar as it allows her to profit from the sale of the painting, and it gives her control over its disposition. However, sometimes these two goals –​of safeguarding interests and of providing control  –​will differ. For example, infants and nonhuman animals have interests but cannot make informed choices, and so the point of assigning them rights is to protect their interests. On the other hand, merely protecting interests ignores the value people attach to controlling their own choices. Suppose, for example, that someone makes use of a vacant private condominium owned by Beyoncé without her permission. When should such a trespasser be said to have violated her property rights? So-​called liability rules imply that paying Beyoncé a standard rental fee fully respects her property rights. Liability rules protect her interests, but they do not protect her choice. One may therefore prefer what are called property rules, which hold that her rights have been violated unless one has her consent. Theorists disagree about whether an interest-​based (“benefit”) or a control-​based (“choice”) conception of rights is the valid one (Calabresi and Melamed 1972). Coleman and Kraus (1986) argue that, depending on facts and circumstances, a particular right may protect interests or choices or both. Among the factors that determine which conception best applies to a particular case are the value for the owner of preserving his or her choice and the costs to the “invader” of gaining consent. For example, a desperate traveler breaking into a cabin high on a remote mountain on a frigid night might be required to compensate the owner but not judged to be guilty of trespass. Rights are important in economics for at least three reasons. First, clear definitions of rights, especially of property rights, promote economic efficiency. A familiar result in transactions-​cost economics is that the absence of clear property rights results in socially wasteful efforts to defend ambiguous claims and discourages investment, because investors cannot be sure that they will reap the rewards of their investment. Second, rights are often the starting points in economic analysis. For example, analyses of the distributional properties of economic arrangements often begin at a given allocation of property rights. Economists may find themselves assuming that this starting point is morally justified, but they may also criticize property rights and suggest ways of modifying them, especially by analyzing different ways in which the many different specific rights that in combination constitute property rights may be “unbundled.”

10.4  The Justification of Rights and Freedoms

183

For example, specific ownership rights can be separated from control rights: individuals might have the right to buy and sell land in protected wilderness areas without having the right to build a shopping mall there. A crucial controversy among rights theorists, which is of great importance to economists, concerns whether there are “welfare rights” –​that is, rights to, items such as food, housing, and health care  –​and what specifically these rights entail. Finally, rights, including especially inalienable rights, may be invoked to limit markets and to restrict the pursuit of economic goals. Most people today recognize that slavery violates human rights and should therefore be forbidden, regardless of whether it might contribute to economic growth. More controversially, one can criticize socialism on the grounds that it violates people’s rights by outlawing “capitalist acts between consenting adults” (Nozick 1974, p.  163) or one can criticize laissez-​faire capitalism on the grounds that it violates people’s rights to a job or to minimum subsistence. Even if there are no welfare rights, governmental provision of food, shelter, and health care might still be defended on the ground that people prefer that welfare benefits be provided and need state action to eliminate the temptation to free ride on the charitable efforts of others (Friedman 1962; Hochman and Rodgers 1969).

10.4  The Justification of Rights and Freedoms Are there moral rights, and what are they? One traditional answer is that human beings can somehow perceive the correct moral rules and recognize that they must be followed. Locke put it this way: The State of Nature has a Law of Nature to govern it, which obliges every one: And Reason, which is that Law, teaches all Mankind, who will but consult it, that being all equal and independent, no one ought to harm another in his Life, Health, Liberty, or Possessions. (1690, section 6)

It is not clear exactly how Locke comes to know that all humans are equal and independent or that “no one ought to harm another in his Life, Health, Liberty, or Possessions.” The authors of the American Declaration of Independence (who were influenced by Locke) regarded their claims concerning rights as self-​evident. Older views also appealed to the idea of a prepolitical natural law (Aquinas 2003). A second answer that is sometimes associated with Rousseau and Kant is that rational individuals who are understood to be free and equal will choose to adhere to certain moral rules. We will explore this contractualist answer in Chapter  12. A  third

184

Liberty, Rights, and Libertarianism

answer is that of the utilitarian and the economist: specifying and protecting certain individual rights promote welfare. There are other possibilities, but these are the main ways of justifying moral rights in modern Western moral philosophy.

10.5  Weighing Rights, Liberties, and Welfare The remaining problem concerning rights and liberty is incorporating them systematically into schemes of overall moral evaluation. This problem takes different forms given different accounts of why liberties and rights are valuable. According to consequentialist views such as utilitarianism, the problem is to determine which rights and liberties maximize good consequences and what priorities to assign to them. This is straightforward in principle. In practice devising effective rules may present difficult problems of strategic coordination. What about accounts that maintain that some rights and liberties are valuable independently of their contributions to welfare? One option is to include the protection or violation of rights and liberties among the consequences of an action, weighting them by their moral importance. For example, when a job applicant is not hired because of employer discrimination, the bad consequences include the violation of the applicant’s right to equal treatment as well as any welfare loss entailed by the worker’s unemployment. Yet this approach, which adds rights violations to welfare costs and benefits, seems to capture only part of the value of protecting rights. Robert Nozick’s view of moral rights as absolute “side-​constraints” on action (1974, pp. 28–​35) is diametrically opposed to any consequentialist view of the importance of rights. On Nozick’s view, an individual’s obligation is not to strive for a minimum of rights violations or a maximum of any social welfare function. No one may violate rights, even in order to prevent other rights violations. Many philosophers have questioned this absolute prohibition on trade-​offs (for example, Nagel 1983), and we suspect that most economists would agree. Indeed, Nozick himself suggests that in the unlikely event that the consequences of respecting rights were catastrophic (as they might be in case of a famine), it might be permissible to violate them after all (1974, p. 30).

10.6 Libertarianism Libertarians defend political liberty, private property rights, and market or economic freedoms. Some libertarians are welfarists who believe that

10.6 Libertarianism

185

defending such rights and liberties most enhances welfare, and historically, many who have supported libertarian policies were motivated by welfare concerns. Friedrich von Hayek, for example, argues that economic prosperity, social innovation, and political democracy are best advanced by limiting government. (Hayek 1967, 1976). Hayek’s argument for a minimal state is epistemic: it rests on the informational demands of coordinating a complex economy, demands he believes are far better handled by a decentralized market system than by a centralized authority. Efforts to regulate the complex economy in an efficient manner would require an incredibly invasive surveillance state. Because Hayek’s theory does not rest on a foundational appeal to absolute rights of private property, it is compatible with social redistribution. Indeed, Hayek claims that “there can be no doubt that some minimum of food, shelter and clothing, sufficient to preserve health and the capacity to work, can be assured to everybody” (Hayek, 1944, 120–​121; see also Hayek 1976, 55). Nozick’s nonconsequentialist brand of libertarianism is thus just one take on the doctrine, and one that not all libertarian economists would find congenial. Other libertarians in the consequentialist mode have drawn conclusions similar to Hayek’s about a minimal state and private property rights. For example, the literature on property rights and transaction costs that derives from Coase’s arguments in “The Problem of Social Cost” (1960) usually reaches libertarian policy conclusions, and it views rights as simply welfare enhancing devices to reduce transactions costs and to alleviate the suboptimalities caused by externalities. By contrast, “philosophical libertarians” maintain that freedom, which depends on the protection of rights, is the overriding political consideration (Lomasky 1987). As Narveson characterizes it, “The only relevant consideration in political matters is individual liberty” (1988, p. 7; see also Machan 1982, p. vii). In the writing of philosophical libertarians, liberty –​ the fundamental value –​is linked to a rights-​based view of justice: we are free and our society is just if no one violates our rights. What distinguishes philosophical libertarianism is the combination of this identification of justice and freedom with respecting rights and a distinctive view of the content of those rights and of the duties rights entail. For example, most such libertarians hold that redistributive inheritance taxes violate individual property rights and are morally impermissible. This position presupposes that property rights include the right to bequeath one’s property without any encumbrances. If property rights did not include the right to bequeath one’s property –​like the property of the British gentry,

186

Liberty, Rights, and Libertarianism

which was often “entailed” to the eldest son or otherwise closest male relative  –​then limiting inheritance in this way would involve no rights violation. Most philosophical libertarians defend strong and encompassing private property rights, and they deny that there are welfare rights to food, shelter, or health care. Except to rectify a past injustice there is, on their view, never a case for redistributive taxation. Libertarians generally hold that the duties that rights imply are almost exclusively “negative” duties not to interfere. For example, libertarians hold that the right to life is a right not to be killed, not a right to be kept alive by being given subsistence. In propounding such a view, libertarians are not necessarily exalting self-​interest or denigrating altruism, although some, such as Ayn Rand and her “objectivist” followers, do espouse egoism (Rand 1964). Most libertarians, like most other people, value beneficence and charity as virtues. They only insist that individuals must not be forced to be beneficent or charitable. The libertarian need not admire Mr. Bumble and the other officers of the parish who (in Dickens’s Oliver Twist) allow paupers to starve to death. Like other readers, most libertarians will find Mr. Bumble a despicable hypocrite. But libertarians who accept the legitimacy of existing property rights would see no injustice in the wretchedness of Bumble’s treatment of the indigent. On the contrary, insofar as the workhouse relies on taxation to provide its miserable relief, it is violating the rights of taxpayers by taking their property to sustain others. Justice requires that one not interfere with the pursuits of individuals, unless those pursuits violate rights. Most libertarians believe that government plays an essential role in protecting individual rights. Yet, at the same time, libertarians regard government, with its monopoly on force, as the greatest threat to freedom  –​in addition to being inefficient. The trick is to limit the functions of government to the protection of rights and to correction of market defects –​that is, to limit the tasks of government to national defense, police, and the judiciary. These must be paid for, and the prospects of covering the cost through voluntary contributions are bleak. Hence, most libertarians regard taxation as justified, provided that it is for these and only these legitimate government purposes. Libertarians view property rights as establishing claims that limit what government can do. But what property rights should people possess? Although politicians sometimes claim that people are entitled to their pretax incomes, without taxes people would not have those incomes. This point is most obvious in the case of government workers such as police, but it is not hard to see that incomes and wealth depend heavily on social

10.6 Libertarianism

187

and political institutions. Without a framework of law, a regulated money supply, and a variety of public goods from vaccinations to highways, there would be little wealth or income to protect from the “encroachments” of government. Even if, as libertarians urge, many of these services should be supplied privately, rather than by the government, people must pay for them. To believe that one’s gross income measures one’s own contribution and entitlement is an untenable but nearly unshakable illusion (Murphy and Nagel 2002). How can the libertarian defend the claim that protecting existing property rights serves freedom better than violating them in order to prevent starvation? The libertarian wants to say that taxing people to feed the starving involves coercion, unlike protecting property rights. But protecting property rights also seems to involve coercion  –​after all, your property rights, backed up by law and the police, prevent me from getting something I want. To distinguish the two cases requires adopting a “moralized” definition of coercion, which states that A is coerced only when unjustified force or threat is used against A. That means that the enforcement of “justified” property rights through legal authority  –​even when the result is starvation –​does not involve coercion. In response, some might question whether existing property rights can be justified if they permit starvation in an affluent world. They can also challenge the moralized definition of coercion. To claim that coercion only exists in the context of unjustified force would imply, uncomfortably, that confining a convicted criminal in prison involves no coercion. More generally, defining liberty as respecting rights makes it a tautology that rights protect liberties rather than a substantial claim. It is implausible to maintain that there are only “negative” rights to be left alone and that there are no rights to positive assistance. Does an infant’s right to life entail only duties on others not to interfere? Can there be justice without a right to a fair trial and hence a right to be provided with legal representation? Many libertarians recognize some rights to positive assistance (Lomasky 1987, pp. 164, 260; Otsuka 2003). But libertarians remain uncomfortable about imposing limitations on the negative liberty of some in order to provide assistance to others. Taking freedom to be the fundamental political value does not automatically commit one to a rights-​based view of morality, to the view that rights rarely if ever obligate others to give positive assistance, or to the view that enforcing existing property rights best serves the cause of liberty. Joseph Raz, for example, argues that valuing autonomy should commit one to favoring an extensive role for government in providing the public goods

188

Liberty, Rights, and Libertarianism

that facilitate the achievement of autonomy (1984, ch. 8). Philip Pettit (2014) presents arguments that the protection of liberty requires an extensive welfare state. Of course, Raz’s and Pettit’s arguments depend on interpretations of liberty that most libertarians would reject. In Robert Nozick’s well-​known version of libertarianism, alluded to earlier in this chapter, rights (whose justification does not depend on consequences) secure individual liberty, and justice is simply respecting rights. According to Nozick’s entitlement theory of justice, an outcome is just if it arises from just acquisition of what was unowned or by voluntary transfer of what agents justly own. Just acquisition is acquisition of unowned resources that makes no one worse off and violates no rights. Nozick regards transfers as voluntary if and only if they do not violate any rights. Only remedying or preventing injustices justifies redistribution or other interferences with voluntary action. A view such as Nozick’s places heavy demands on a theory of rights, which Nozick himself leaves undeveloped (Nagel 1983). Whether a state of affairs is just, then, according to Nozick, depends on its history, not on the pattern of holdings. This poses a huge problem for followers of Nozick who seek to assess the current distribution of income or wealth, because the past is, as a matter of fact, full of injustices. In Nozick’s view, these injustices must be rectified to bring about the state of affairs that would have obtained if there had been no past injustices. Rectification of rights violations is unproblematic if one is thinking about an injustice such as a teenager stealing a car. But there is no way to know what the world would have been like today without the millennia of horrific injustices that constitute the history of what has been called “civilization.” Even if followers of Nozick can ignore the dark shadow that past injustice casts upon the justification of current holdings, they still have to defend a detailed specification of property rights in order to determine which distributional policies, including tax policies, are justified. Does just acquisition of some piece of land give one rights to the airspace over the land or the minerals buried deep below the surface? Even if Nozick is right that justice is determined by entitlement –​that is, by respecting rights –​not by “pattern,” patterns might reenter the story at a higher level when one is deciding what the content of legal rights should be (Pogge 1989, ch. 1). The philosophical libertarian’s commitment to liberty is in principle independent of any welfare consequences, but libertarians would like it to be the case that protecting freedom also makes people better off, and libertarians need to show that extreme proposals, such as privatizing all streets and highways (Block 1982), would not lead to disastrous consequences. Narveson states the point as follows:  “Those who toil in the libertarian

10.7  Libertarian Paternalism

189

fields devote the lion’s share of their efforts to persuading us that the alleged benefits of the State are illusory” (1988, p. 183).

10.7  Libertarian Paternalism Many who value liberty are opponents of paternalism, understood as the view that it is sometimes acceptable to coerce people for their own good. Welfarists who take welfare to be the satisfaction of preferences also reject paternalism on the grounds that if what people prefer is best for them, then coercion cannot promote their interests. The case against paternalism that is grounded in the value of negative liberty seems to be much stronger, and some readers of Mill’s On Liberty worry that his classic argument against paternalism rests on a commitment to freedom that is incompatible with his professed utilitarianism. The case against state paternalism seems particularly secure, since state officials will rarely be in a good position to judge what is best for individuals. Moreover, granting to the state the power to coerce individuals for their own good threatens to open a path to tyranny, a path that worried Mill. The past decade has seen the defense of what is purported to be a new kind of paternalism, “asymmetric” or “libertarian” paternalism, which, its proponents (especially Thaler and Sunstein 2008) argue, is fully compatible with respect for freedom. The most compelling example concerns retirement plans. Behavioral economists have discovered that when new employees are automatically enrolled in company retirement plans (although they are free to opt out), far more enroll than when they are not automatically enrolled and need to opt in. Since few employees believe that their decision about whether to enroll in a retirement plan should depend on whether doing so is the default option, it appears that some of the choices individuals make do not promote their interests. If, in addition, policy makers can judge that in the majority of cases enrollment is the better choice, then by requiring that enrollment be the default option, government can promote people’s interests. At the same time, there is no limitation on freedom, since employees can easily and virtually costlessly opt out. Although not all of the policies that Thaler and Sunstein discuss in Nudge are paternalistic, they are all designed to remedy predictable failings in rational choice, without limiting freedom. What distinguishes nudging is its mechanism –​that is, its exploitation of failures of rationality to shape people’s behavior. If, by definition, paternalism involves coercion, it would appear that the “choice architecture” that Thaler and Sunstein propose is not paternalistic. But it nevertheless strikes most people as paternalistic, because the choice

190

Liberty, Rights, and Libertarianism

architect is substituting his or her judgment for the individual’s own (faulty) judgment, and nudging seems to involve a kind of manipulation. It does not make use of people’s rational capacities to improve their choices. Thaler and Sunstein counter that there is no nonmanipulative alternative: our choices start from a background default that depends on public policies. In more recent work (2015), Sunstein defines nudging so broadly that it includes merely informing or warning, on the one hand, and deceit, on the other. “Nudges are interventions that steer people in particular directions but that also allow them to go their own way” (2015, p. 7). On this very broad understanding, there is no point in asking whether nudges are paternalistic or in any way morally dubious. Some are and some are not. There is nothing sacrosanct about the traditional understanding of paternalism, and Le Grand and New (2014, p.  2) maintain that “government intervention is paternalistic with respect to an individual if it is intended (a) to address a failure of judgment by that individual and (b) to further the individual’s own good.” On this definition, coercing someone for his or her own benefit is neither a sufficient nor a necessary condition for paternalism. For example, Le Grand and New (2014) maintain that the state’s subsidizing of museums is sometimes a form of paternalism. It is too soon to judge how fruitful it will be to shift the moral questions concerning paternalism so that they focus on the justification of correcting lapses in rational judgment rather than on the justification of coercion. Yet, in an age of incessant advertising, when scruples about exploiting people’s rational foibles are futile, it seems that government may have a role to play in protecting autonomy as well as negative liberty. Even though nudging may threaten an agent’s freedom in some senses, it may also be essential to protecting it.

Suggestions for Further Reading Isaiah Berlin’s celebrated essay “Two Concepts of Liberty” has provoked much discussion. A valuable collection of essays inspired by Berlin’s work is Ryan (1979). MacCallum (1967) is a valuable commentary on Berlin. Schmidtz and Brenner (2010) offer a useful overview of different conceptions of freedom. For discussions of autonomy, see especially Joel Feinberg (1986, p. 31f.), Gerald Dworkin (1988, ch. 1), Raz (1984), Christman (1989), and, as self-​ determination, Pettit (2014). On the definition of rights: Hohfeld (1919) regards claims as “the limited and proper meaning” of “right.” Wellman (1985), Sumner (1987), and Thomson (1990) follow Hohfeld terminologically, though none of them is

Questions for Study and Discussion

191

committed to construing pure claims as the paradigm cases of rights. See also Feinberg (1973, chs. 4–​6), Waldron (1984), and Wenar (2005, 2013). For a defense of a “choice” construal of rights, see Hart (1955), while Raz (1984) defends a benefit construal. For general discussions of rights see Dworkin (1977), Sumner (1987), Thomson (1990), and Martin (1993), and for a sustained discussion of the difficulties of defending libertarian property rights see Waldron (1990). Amartya Sen (1982b) discusses the option of including rights violations among the consequences that a consequentialist should take into account. Thomas Nagel, on the other hand, has argued powerfully for preserving a nonconsequentialist component in rights evaluation (1986, ch. 9). Nozick (1974), Machan (1982), Lomasky (1987), Narveson (1988), and Tomasi (2013) provide systematic accounts of a libertarian viewpoint. Friedrich von Hayek’s treatise (1960) argues that a variety of consequentialist considerations support the case for limited government and wide protections for negative liberty; see also Hayek (1944). Schmidtz (1994) makes an eloquent case for the importance of property rights to human flourishing, which draws on the work of Coase (1960) and Demsetz (1967). See also Cowen (1988). For a critique of the “vulgar libertarianism” that holds that people are entitled to their pretax holdings, see Murphy and Nagel (2004). Von Mises (1941) is a classic libertarian discussion of economics. Grey (1980) discusses the disintegration of the concept of property. The most influential discussion of libertarian paternalism is Thaler and Sunstein’s Nudge (2008). For other important discussions, see Camerer et al. (2003), Thaler and Sunstein (2003a, 2003b), Hausman and Welch (2010), Le Grand and New (2014), and Sunstein (2014).

Questions for Study and Discussion 1. What are the differences between what defenders of positive liberty mean by self-​determination and what defenders of republican liberty mean? How extensive and important are these differences? 2. Does republican freedom require negative liberty? What about positive freedom: does it require negative liberty? How exactly are the four notions of freedom or liberty related? 3. Does it make sense to maintain that a kidnapper demanding ransom limits the freedom of the parents of the kidnapped child, while a surgeon demanding a fee to carry out a lifesaving operation on a child does not limit the freedom of that child’s parents?

192

Liberty, Rights, and Libertarianism

4. Does a billionaire such as George Soros or David Koch have more political freedom than someone who is not rich, or should one say instead that Soros and Koch have the same political freedoms that the rest of us have, but their freedoms are worth a great deal more than ours? Why does the answer to this question matter? 5. What do you take autonomy to be? How important is it to you personally, relative to your own developed values? How important is the protection of autonomy to desirable policies for a polity? Many cultures believe that submission to established customs is more important than autonomy. What do you think? What are your reasons? 6. Over the past generation, China has made incredible strides at improving the health and well-​being of its citizens, but it has at the same time limited their freedom of movement, speech, reproduction, assembly, and religion. Do you think that trading off among these valuable ends has been in fact necessary? If it was, do you think that the trade-​off has been worthwhile? If promoting welfare and protecting negative liberty conflict, how do you think that conflict should be resolved? How important is negative liberty? 7. Rights often consist not only of privileges and claims, but also of powers and immunities. Explain what privileges, claims, powers, and immunities are. Which powers of alienable rights do inalienable rights lack? Since inalienable rights lack some of the powers of alienable rights, one would think that they are worth less. Yet it is important to treat some rights as inalienable. Why? 8. Thomas Jefferson wrote in the Declaration of Independence that all men “are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.” Yet, as a slave owner, he was well aware that many men and women at the time were enslaved. Was his claim false? In that context, how can we make sense of Jefferson’s assertion that people are endowed with such rights? 9. One of the arguments made by those who opposed granting women the right to vote was that the interests of women would be protected by men. If that claim were true, would it still matter whether women had the right to vote? Why or why not? 10. Is it consistent for someone to be a libertarian but to reject Ayn Rand’s defense of self-​interest and condemnation of altruism? Be as clear as you can be about what understanding of being a libertarian you are using.

Questions for Study and Discussion

193

11. Why is it impossible for everyone to have a right to his or her pretax income? 12. Robert Nozick maintains that if those who in exchanging with one another on a market have a right to what they bring to the market and do not engage in any force or fraud in their interactions with others, then they have a right to whatever they receive from the exchange, regardless of how unequal the final distribution of income, wealth, or political power may be. Do you think he is right? Why or why not? 13. If the current distribution of wealth in a society derives in part from past injustices (as is certainly the case in every society), then a theory of distributive justice that depends on the history of a distribution rather than on its pattern must find that the current distribution is, at least in part, unjust –​unless, the injustice has been rectified. If it is possible to rectify the past injustices, then the libertarian would favor doing so, taking from some people those goods or funds to which they have no defensible historical title. But in the case of massive injustices, rectification does not seem feasible. What then should a libertarian recommend? 14. Do you think that it is morally permissible for the government to take advantage of psychological foibles, such as people’s laziness and their inclination to take the default option, to “nudge” people for their own good? What about nudging them to benefit others by discouraging littering or encouraging the use of public transportation? Is that more or less defensible? Why? 15. Richard Thaler (coauthor of Nudge) recently proposed these standards for all “nudging” [www.nytimes.com/​2015/​11/​01/​upshot/​the-​ power-​of-​nudges-​for-​good-​and-​bad.html?_​r=0]: a. All nudging should be transparent and never misleading. b. It should be as easy as possible to opt out of the nudge, preferably with as little as one mouse click. c. There should be good reason to believe that the behavior being encouraged will improve the welfare of those being nudged. Advertising is of course a form of nudging. Are advertisers morally obliged to adhere to these standards? Should they be legally obliged to do so?

ELEVEN

Equality and Egalitarianism

Human beings differ in countless ways. These differences can be described as inequalities in particular traits, skills, circumstances, or resources. Most of these inequalities are not ethically troubling: indeed, it would be undesirable if everyone had the same skills or the same aesthetic tastes. The world is richer for the differences among people. Nor is there anything ethically objectionable in the fact that professional athletes are more fit than college professors. Indeed, many authors (confusingly) use the term “inequality” only for differences that are morally problematic. There are, unfortunately, many morally problematic inequalities. Consider that the median wealth of American families of African ancestry is just a bit more than one-​twentieth (!) –​that is, 5 percent –​of that of families of European ancestry. In 2013, according to the World Health Organization, life expectancies were below age fifty in fourteen countries (all in sub-​Saharan Africa, except Afghanistan). At the other end of the spectrum, in nineteen countries, life expectancies were above age eighty. Inequalities in wealth and income have been growing in the United States and abroad. According to Bloomberg News, in 2013, corporations in the Standard and Poor’s 500 index paid their CEOs more than four hundred times the minimum wage, while in 1950 the ratio was 20:1. The bottom 40 percent of the U.S. population has only one-​fifth of 1 percent of all the wealth in the United States. International inequalities are even larger (although with rapid development in India and China the proportion of the world’s population that is desperately poor has declined). To give some idea of the magnitude of the global gap, in 2012 the income of the one hundred richest people in the world was greater than the combined income of half a billion of the world’s poorest. These inequalities in income and wealth, along with inequalities in health and education, are now flash points for public debate. The ephemeral 194

Equality and Egalitarianism

195

“Occupy Wall Street” movement was a massive protest against the income and power of the richest 1 percent. Some of the best-​selling recent books in economics such as Joseph Stiglitz’s The Price of Inequality (2013) Thomas Piketty’s Capital in the 21st Century (2014), and Anthony Atkinson’s Inequality:  What Can Be Done (2015) scrutinize the sources and consequences of income and wealth inequalities and suggest ways to reduce them. The current interest in economic inequality derives in part from the large increases in this form of inequality, especially in the United States and Britain, at the same time as there has been a stagnation or even a drop in the incomes of those in the bottom half of the distribution. These facts raise moral as well as economic questions. The most obvious question is simply “What’s wrong with inequality?” but this question is too simple, since inequalities in some things are not morally problematic while inequalities in others are. So we need to consider which inequalities are wrong and for what reasons. Consider, for example, four proposed compensation schemes in a company. (A) pays all the workers the same wages. (B) pays all workers the same piece rates, so that each worker gets the same compensation for every task the worker completes. (C) pays all the workers with the same seniority the same wage, but increases the wage with years of additional employment. (D) pays all the workers who have the same number of dependents the same wage, but pays workers with more dependents a higher wage. All of these treat the workers equally in some regard and unequally in others. Which compensation principle should an egalitarian favor? The meaning of equality of opportunity is just as complex. People object to laws that ban gay marriages on grounds of equality, yet such laws leave everyone equally free to marry anyone of the opposite sex. Are affirmative action programs an implementation or a violation of equal opportunity? To attach any meaning to claims about what policies promote equality of opportunity requires that one be clear about the respects in which opportunities are to be equalized. Morally problematic inequalities are found between groups as well as individuals. Although some argue that inequalities among groups matter independently of their consequences for individuals, that is not a common view; nor is it ours, and we will not discuss it further. We assume that group inequalities matter because of their implications for individuals. Nevertheless, attending to group-​based inequalities can help to identify the sources of inequalities among individuals (such as racial discrimination),

196

Equality and Egalitarianism

distinguish the consequences of inequalities, and point to effective policies to address objectionable inequalities. Additionally, inequalities in one regard often correlate with inequalities in other regards, and focusing on groups may make these correlations more obvious. For example, those who have higher incomes tend to have higher social status. They are also better educated, better nourished, more comfortably and safely housed, and appreciably healthier. Moreover, as recent scholarship has documented, they have a great deal more political influence than their poorer fellow citizens (Bartels 2008; Gilens 2012). Obviously, income and wealth are valuable only insofar as they enable people to live good lives, and those who are wealthier are not always better off. But economic inequalities are nevertheless indicators of other more fundamental inequalities. On some views, the importance of inequalities does not lie merely in their consequences. To many –​as Ronald Dworkin, James Griffin, Will Kymlicka, and Amartya Sen have all argued  –​some form of equality lies at the heart of morality. Sen argues that all moral theories prescribe that people be treated equally in some regard (1992a, p. 3): utilitarianism requires that everyone’s welfare count equally, and libertarian theories require equality of rights. Dworkin (1977) and Kymlicka (1991) maintain that conflicts between different political philosophies can best be understood as conflicts over the best interpretation of equal respect and what it entails. On a libertarian interpretation of equal respect, there is nothing objectionable in the fact that some people are much richer than others (unless they obtained their riches by theft or deceit), while, on a utilitarian interpretation, if these large inequalities fail to maximize happiness, the reason likely lies in a failure to give equal weight to every person’s welfare. In addition to inspiring many historical movements for social justice, appeals to equality play a crucial role in debates over economic policy. One reason why many criticize the massive tax cuts secured by the administration of G.  W. Bush is that they contributed to inequalities in wealth and income. One reason why many support the Affordable Health Care Act (“Obamacare”) is that it lessens inequalities in access to health care. Concerns about inequality have also been among the most important reasons for interfering with market outcomes. Inequalities may be objectionable because of their causes, because of their consequences, because of what they express about a society, or in themselves, regardless of their causes and consequences and what they express. How important are the various sources of concern about inequalities and what do they imply with respect to economic policy? For example, when

11.1  What Is Wrong with Contemporary Income and Wealth

197

the pursuit of equality conflicts with the constraints imposed by rights or with other objectives such as welfare or liberty, how much weight should be placed on equality? As with our discussion of liberty and rights in Chapter 10, we face problems of definition, justification, and weighting. We begin in Section 11.1 with the second of these questions, focusing on the causes and consequences of inequalities.

11.1  What Is Wrong with Contemporary Income and Wealth Inequalities? Consequences and Causes Concerns about the consequences of economic inequalities have focused mainly on how they affect well-​being, whether they undermine equality of opportunity, and whether they weaken democratic institutions. With respect to the causes of these inequalities, the central moral questions concern justice and desert. Do economic inequalities derive from injustices? Are they deserved? Begin with consequences. If one supports democracy, then inequalities in income and wealth are alarming to the extent that they tear the social fabric, or skew public priorities. For example, if the rich can opt out of public education or state-​supported health care, they will be less willing to pay taxes to support these institutions. Moreover, while disproportionate amounts of money do not guarantee election results, they can have significant effects, particularly when candidates are not already well known and advertisements can easily shape voters’ impressions. What is especially noteworthy about political spending in the United States is not mainly the disparity across parties, but its sheer magnitude: 7.2 billion dollars in 2014 according to the Federal Election Commission. Defenders of the current wealth distribution in the United States respond to the concerns about the possibility that huge contributions undermine democracy by pointing out that liberal Democrats do, after all, regularly win elections. But these Democrats typically have their own wealthy backers, especially at higher levels of office, and it is mistaken to believe, as this response assumes, that only Republicans defend the interests of the rich. In fact, there is good evidence that neither party is much concerned about the poor (Bartels 2008; Gilens 2012). A second consequence-​based criticism of wealth and income inequalities argues that they make people on the whole worse off. As we noted when discussing utilitarianism, if the marginal utilities of wealth and income –​ that is, increases or decreases in one’s welfare with increases or decreases in

198

Equality and Egalitarianism

one’s income and wealth –​diminish with larger incomes or greater wealth and if similar levels of income typically yield similar levels of utility to different people, then transferring income or wealth from those who have a great deal to those who have little should increase total welfare. In other words, all else being equal, the way to derive as much welfare as possible from a fixed “pie” is to slice it equally. Inequalities in this way lower total welfare. On the other hand, all else may not be equal: the social pie may grow or shrink depending on how it is cut. If there is no financial incentive for investment or work, then the economic pie will shrink drastically. By their effect on incentives, greater inequalities may then lead to an increase of total welfare, even for those at the bottom of the income distribution. Which of these effects on welfare is larger? Over the long run economic growth has apparently mattered a great deal more to people’s health, incomes, and wealth than has its distribution. Recall the words of Robert Lucas quoted in Chapter  1, “Of the tendencies that are most harmful to sound economics, the most seductive, and in my opinion the most poisonous, is to focus on the question of distribution… The potential for improving the lives of poor people by finding different ways of distributing current production is nothing compared to the apparently limitless potential of increasing production.” For example, over the past half-​century, average life expectancy for the more than 1 billion Chinese has gone from about age forty to age seventy-​three. During this same period there has been rapid economic growth and an increase in income inequality. On the safe assumption that a thirty-​year increase in life expectancy provides a huge increase in well-​being, growth in this case has been far more important than distribution. Similarly, according to the 2015 Economic Report of the President, if productivity growth had not slowed after 1973, the median household would have thirty thousand dollars of additional income today, while if income inequality had not increased after 1973, the median household would have nine thousand dollars of additional income today. So, one might conclude, it would be a mistake to institute policies that diminish inequalities because of the risk that they will undermine economic growth. But even if growth is more important to welfare than is distribution (which may not continue to be true, owing to the effects of economic growth on the environment), how large are the incentive effects of policies devoted to lessening inequalities? The fact that completely equalizing incomes would destroy incentives to work and invest does not imply that raising the top marginal income tax rate from 40 percent to 50 percent or lowering it to 30 percent would have a big effect on incentives. Currently, Republican leaders in the United States all maintain that large cuts in income tax rates

11.1  What Is Wrong with Contemporary Income and Wealth

199

would have a large and lasting positive effect on economic growth, but there is nothing in the history of the effects of tax changes during the past half-​ century to support this view. (There is, on the other hand, evidence that tax cuts, like deficit spending, have a short-​run effect as a form of economic stimulus, but whether to employ tax cuts as a way of stimulating aggregate demand is a separate question from whether changing the distribution of income will affect the rate of economic growth.) A third significant consequence of inequality is a highly uneven playing field. When some children’s lives are so radically different from others’ in terms of access to resources, one might doubt whether these children have a fair chance of finding places in college and satisfying jobs. According to the Wall Street Journal (February 3, 2015), “In 2013, 77% of adults from families in the top income quartile earned at least bachelor’s degrees by the time they turned 24, up from 40% in 1970 … 9% of people from the lowest income bracket did the same in 2013, up from 6% in 1970.” Recent work by Raj Chetty et al. (2016) has shown that the zip code into which children are born is significantly correlated with their educational attainment, their health, and their earnings (see /​www.equality-​of-​opportunity.org/​). And, of course, the passport you carry has a great effect on the choices that are open to you. Turning from the effects of inequalities to their causes, some portion of the economic inequalities in the United States and in other countries clearly results from past and ongoing injustices. For example, there is no plausible account of the huge racial disparities in wealth in the United States in which racial discrimination and the continuing legacies of slavery and Jim Crow do not play a very significant role. As important as past injustices are to explaining important aspects of current inequality, they cannot be the whole story. Even though there have been some extremely rich scoundrels, and even though there are continuing racial injustices, it might be argued (as the Harvard economist Gregory Mankiw does) that billionaires such as Bill Gates or J. K. Rowling deserve their riches and that competitive markets in general provide people with compensation that is “congruent with their contributions” (2013, p. 32). But there is plenty of conceptual space between the view that the rich are thieves and that they are fully deserving of their levels of compensation. To maintain that competitive markets provide compensation to people that is congruent with their contributions is empty if the only measure one has of the worth of someone’s contributions is the market compensation she currently receives. So the congruence claim has content only if we have some sense of what makes contributions significant or trivial. Recall the

200

Equality and Egalitarianism

example of eflornithine, which we discussed in Chapter  2. As a cure for sleeping sickness that could save thousands of lives, its production is presumably a very significant contribution. But its value on the competitive market as a cure for sleeping sickness is precisely zero, because those whose lives it would save cannot afford the cost of producing it. Or consider a wonderful new drug, Sovaldi, which cures hepatitis C.  Its maker, Gilead Pharmaceuticals, is pricing it at one thousand dollars a pill or eighty-​four thousand dollars for a course of treatment. In justification, Gilead points out that curing hepatitis C is worth much more than a mere eighty-​four thousand dollars. But the price reflects not the importance of curing the disease, but a calculation on Gilead’s part of how much they can charge without provoking a response from the U.S. Congress. Given patent law, Gilead can ask whatever they please, and given the policies of Medicare and Medicaid, the U.S.  government will pay whatever they ask. If there were unregulated competition without patent protection, Sovaldi would likely sell for much less, quite possibly less than ten dollars a pill. (It is currently available in Egypt for less than four dollars a pill.) The intrinsic value of what people do or make has little relation to market returns, which depend on all the many factors that influence supply and demand. If there were no copyright laws, Rawlings would still be a wonderful writer, but she would not be a billionaire. She would, instead, be in a position like that of Mozart, who (while contributing no less and, presumably, no less “worthy”) in the absence of copyright protection, had a good income but no fortune. When copyright and patent laws change, the market values of people’s contributions change but the worth of their contributions does not. Consider too the fates of nineteenth-​century portrait painters, most of whose incomes drastically declined with the invention of photography; hand-​loom weavers, who were driven to starvation by the introduction of power looms; and blacksmiths, whose incomes collapsed with the introduction of automobiles. In each case, the skills, ambition, and exertion of the individuals, which are the bases of judgments of worth, remained unchanged, while their compensation changed drastically. The function of prices in a competitive market is to direct resources to where they are most demanded, not to reward contribution. The idea that markets track specifically moral desert is even less plausible than the idea that they equate compensation to the value of a contribution. Sometimes politicians proclaim that the rich deserve their fortunes on the grounds they work harder than the poor. But the view that the typical bank president works harder than an impoverished single mother who has two young children and a full-​time job is absurd. In any event, it is hard (and

11.2  What Is Wrong with Inequality Itself?

201

indeed presumptuous) to measure people’s intrinsic moral desert, and there is no reason to think that this in any way corresponds to their contributions to the economy. Indeed, there is experimental evidence that, on average, the affluent are less morally admirable than those who are relatively poor (Piff et al. 2012). Even libertarians like Hayek (1945) have noted that there is no plausible story in which a person’s market rewards can respond to his inner moral worth. The information needed to track moral worth is simply not on offer in the marketplace. Finally, in evaluating contemporary inequalities, it is important not to lose sight of the fact that, particularly in a global context, inequalities usually involve misery for those who are worst off. What is shocking about the disparities in life expectancies across the world is arguably not that the difference is almost forty years but rather that life expectancy in much of sub-​Saharan Africa is less than fifty. Though it has no necessary connection to egalitarianism, a humanitarian concern for those whose life prospects are poor is surely one of the main moral grounds supporting egalitarian policies. In Joseph Raz’s words, What makes us care about various inequalities is … the hunger of the hungry, the need of the needy, the suffering of the ill, and so on. The fact that they are worse-​ off in the relevant respect than their neighbours is relevant. But it is relevant not as an independent evil of inequality. Its relevance is in showing that their hunger is greater, their need more pressing, their suffering more hurtful, and therefore our concern for the hungry, the needy, the suffering, and not our concern for equality makes us give them the priority. (1984, p. 240)

11.2  What Is Wrong with Inequality Itself? “Basic” Egalitarianism and Prioritarianism A commitment to moral and political equality is deeply embedded in American political traditions and in moral philosophy, but the basis for this commitment and its implications for specifically economic equality are controversial. Setting aside its causes and consequences, why should equality matter morally? David Miller poses this question as follows. Why should equality be thought desirable? Equality after all means a leveling of differences; it means a smoothing down of irregularities or idiosyncrasies. Although I may from an aesthetic motive decide to trim my rose bushes to an equal height or polish my wine glasses to an equal shine, to treat people in such a way would be at best perverse and at worst immoral. The pursuit of equality seems to be impaled on a fork: either the ultimate end of the pursuit is not equality at all but some other value or values which have become confused in the popular mind with equality, or

202

Equality and Egalitarianism

our societies are aiming at a goal that cursory inspection reveals to be quite monstrous. (1982, p. 73)

We shall comment on two responses to this puzzle. The first, which is the topic of this section, maintains that lessening inequality or prioritizing the well-​being of those who are worse off is itself good. The second, to be discussed in Section 11.3, and already implicit in the previous section, sees the value of equality as instrumental to objectives that are themselves egalitarian in character. The first and most direct response to Miller’s question has been given by philosophers such as Larry Temkin, Derek Parfit, Richard Arneson, and G. A. Cohen. In Cohen’s view, the purpose of egalitarianism “is to eliminate involuntary disadvantage” (1989, p.  916). Parfit writes:  “The Principle of Equality … claims that it is bad if, through no fault of theirs, some people are worse off than others” (1984, p. 26). In Temkin’s view, for some to be worse off than others undeservedly –​that is, through no fault or choice of their own –​is “comparatively unfair” (2003, p. 767). So, for example, Temkin maintains that it is unfair if John is crippled because a tree limb falls on him while he is out for a walk, since millions of others go for walks without tree limbs falling on them (pp.  772–​773). In Temkin’s view, the claim that undeserved inequality is prima facie morally objectionable is a fundamental moral intuition. Let us call this view “basic egalitarianism.” Although many thoughtful philosophers hold this view, there are good reasons to reject it. The only argument in its favor is that it supports the moral condemnation of so many of the objectionable inequalities in our contemporary world. But it is difficult to separate one’s attitudes toward inequalities from one’s reactions to the suffering and bleak life prospects that accompany the inequalities that are so evident in the world today. Because those who deny that inequality is intrinsically bad would make most of the same judgments concerning particular inequalities, basic egalitarians do not have a strong positive argument for their view. Here are three arguments against basic egalitarianism. First, the view that inequalities are intrinsically morally objectionable offers moral criticism of any unchosen disadvantage, regardless of its source. But hurricanes are not subject to moral censure when they wreak their havoc. In our view, in contrast to basic egalitarianism, moral judgment enters only with respect to the response to that havoc. The national and state response to Hurricane Katrina, unlike the hurricane itself, was morally objectionable. Second, it is questionable whether basic egalitarianism should be the basis of state policy. Basic egalitarianism requires that the state sort out

11.2  What Is Wrong with Inequality Itself?

203

voluntary from involuntary disadvantage. Is this feasible, and, if feasible, is it desirable? Third, basic egalitarianism also appears to lead to ethically absurd conclusions. Surely it is better if half the population can see than if all are blind despite the greater equality of the latter state of affairs, even if no one is responsible for their blindness. In response, defenders of basic egalitarianism might reply that equality is not the only important value and that it is also morally important to make people better off: even though there is one regard in which it would be better if everyone were blind, the welfare loss outweighs the gain in equality. Many critics of basic egalitarianism are not satisfied by this response. In their view, there is nothing good at all about an increase in equality among the sighted and the blind achieved by leveling down. The problems posed by this last objection to “leveling down” have led some philosophers to favor prioritarianism. Prioritarianism maintains that a greater weight should be placed on improving the well-​being of those who are worse off than on improving the well-​being of those who are better off. Although distribution is of no intrinsic importance to the prioritarian, in weighting the good of those who are worse off more heavily than the good of those who are better off, prioritarians, like egalitarians, will in practice endorse policies that reduce inequalities. Although prioritarians escape the leveling down objection, they are as subject to the first two objections as basic egalitarians. Indeed, it can be difficult to describe practical examples in which the policies endorsed by prioritarians and basic egalitarians diverge (Fleurbaey 2015). Is it always justified to give more weight to the claims of those who are worse off, no matter how well off these worse off happen to be? The problem in our view is that neither prioritarianism nor basic egalitarianism provides a satisfactory account of what is disturbing about inequalities. To explain what is wrong with inequalities by saying that equality is intrinsically valuable or that one should give special weight to the interests of the worse off invites virtually the same question that one was supposed to answer: Why should such inequalities matter? Why should improving the well-​being of the worse off matter more than improving the well-​being of the better off, even if the worse off are very well off themselves? The fact that some goods are “positional” (Hirsch 1976; Frank 1987) might be misunderstood as lending support to prioritarianism or to basic egalitarianism. Goods can be “positional” in two different ways. First, some things are positional goods because the benefits they provide depend on how much of them others have. An elite education is in part

204

Equality and Egalitarianism

a positional good in this sense. Regardless of the excellence of a Yale degree, a part of its benefits depends on the fact that only a small number of people have a Yale education or its equivalent. The benefits that positional goods in the first sense provide are also positional goods in a second, less canonical sense. For example, recent research has shown that health is a positional good in this second sense: how healthy one is depends on one’s position in the socioeconomic hierarchy. Income is a positional good in both senses. The worth of one’s income depends on how much income others have, and in turn income affects how much of other valuable goods one has. Since many important goods are positional goods, inequalities are worse for those near the bottom than they might appear to be. But this fact lends no support to prioritarianism or to basic egalitarianism. In any event, it is hard to imagine any set of institutions in which there will be no positional goods: there are only so many seats possible behind home plate at Yankee Stadium. Inequalities for which individuals are responsible are not objectionable from an egalitarian or prioritarian perspective. They may be objectionable in other regards, and there is no reason to suppose that basic egalitarians or prioritarians would recommend that irresponsible motorcyclists who do not wear helmets be allowed to die by the roadside. But it is only inequalities for which individuals are not responsible –​those that are the product of what Ronald Dworkin terms “brute bad luck” –​that are of concern to the prioritarian or the basic egalitarian. Elizabeth Anderson (1999) dubs this version of egalitarianism “luck egalitarianism.” Richard Arneson calls prioritarianism that takes responsibility into account “responsibility-​catering prioritarianism” (2000, p. 340). Luck egalitarians and responsibility-​catering prioritarians face the challenge of distinguishing which outcomes individuals are responsible for and which they are not responsible for or, when outcomes depend in part on choices and in part on the environment, how to apportion responsibility. One possibility is to hold individuals responsible for the expected value of their choices rather than for the actual outcomes (Le Grand 1991, p. 99). Most egalitarians who favor compensating people for being born blind oppose compensating people for having little ambition or for holding religious beliefs that induce them to starve themselves. As Scanlon (1986, pp.  116–​117) and Dworkin both emphasize, “It would be incoherent for me to regard some ethical conviction I have –​that the only important thing to do with my life is to create religious monuments, for e­ xample  –​as a limitation on the goodness of the life I can lead” (Dworkin 1990, p. 108). If the point of egalitarianism is to show equal respect for individuals, then

11.2  What Is Wrong with Inequality Itself?

205

egalitarians cannot regard their convictions and characters as misfortunes, even when they lead them to bear costs. How exactly to capture these distinctions in a coherent way is a large and unfinished project. For example, why should society draw a sharp line between a person’s abilities and her preferences? It will not do to say that she has no control over her abilities while her character is her own doing. As any parent knows, much of character and temperament is also beyond a person’s control. G. A. Cohen has argued that, rather than drawing a line between abilities and preferences, egalitarians should distinguish between those disadvantages whose acquisition and retention are voluntary and those that are not (1989, pp.  920–​934). In Cohen’s view, the purpose of egalitarianism “is to eliminate involuntary disadvantage” (1989, p. 916; see also Arneson 1989, p. 177). However, Cohen’s way of drawing the line has its own implausible implications. As already noted, it would be inappropriate for the state to compensate someone for holding religious beliefs that demand an expensive monument to a deity, regardless of whether those beliefs are involuntary. Cohen concedes this point and revises his account to rule out compensation not only for disadvantages resulting from choice, but also for disadvantages so intrinsically connected to their bearers that their bearers would not choose to be without them (1989, p. 937). But if involuntary disadvantages are not always of egalitarian concern, then the purpose of egalitarianism cannot be to eliminate involuntary disadvantage. One promising suggestion concerning how to distinguish those outcomes for which individuals should be held responsible from those that call for compensation is due to John Roemer (1998). Suppose that the egalitarian seeks to equalize welfare insofar as inequalities result from circumstances, but not when inequalities are due to people’s efforts. Roemer defines “types” as those who share the same circumstances. Depending on levels of individual effort, members of the same type will experience different outcomes. Ideally, Roemer argues, the distribution of well-​being as a function of effort should be the same for every type. For example, some low-​income residents of inner-​city neighborhoods who lack access to healthy foods will eat worse diets than their neighbors, while some will eat healthier diets. Exerting the same level of nutritional effort as an affluent resident of an upscale community, the inner-​city resident will eat a much less healthy diet, but if the level of effort is the same, so is the extent of responsibility. The egalitarian aims to provide enough compensation so that the net effect of diet and compensation on well-​being will be the same for the low-​and high-​income individuals who exert the same level of effort. Of course, implementing Roemer’s

206

Equality and Egalitarianism

proposal requires a prior (and controversial) decision concerning which “types” are important. It is worth noting that there is an asymmetry in people’s attitudes toward responsibility. If an individual is badly off as the result of a moral or altruistic choice (as opposed to carelessness), then the resulting inequality still calls for compensation. Those who contracted Ebola owing to their own choice to care for its victims are no less deserving of treatment, despite their heightened responsibility for their fate. Larry Temkin’s version of egalitarianism, which focuses on inequalities that are not deserved rather than inequalities for which individuals are not responsible, avoids this problem. However, Temkin (2015) has recently suggested that the basic egalitarian needs to be concerned with both desert and responsibility.

11.3  Egalitarian Objectives We have questioned whether there is a defensible moral case condemning all and only those inequalities in welfare that are undeserved or for which the individual is not responsible. We have suggested that when inequalities in life prospects, happiness, health, resources, or wealth are unjust, as they often are, their injustice lies in their causes or consequences or in the avoidable suffering of those who are worse off. This is not to deny that there is a kind of inequality that is wrong in itself. This is the kind of inequality that denies some people moral standing, the kind that says that some people count for nothing, or that their interests matter less than others’. Affirming this kind of equality –​holding that all members of the human community have equal moral standing –​does not itself yield the basic egalitarian, luck egalitarian, or prioritarian principles. More work needs to be done to show what the implications of this commitment actually are, and there are many competing views of what equal moral standing entails. In responding to his own question about why equality should be thought desirable, Miller emphasizes the connection between equality and other moral objectives, which are themselves egalitarian in character. Miller lists four different ends that equality (including economic equality) may serve and to which equality has an intimate connection. Scanlon (2003) adds a fifth. First, Miller maintains that equality is sometimes required in order to be fair and to prevent arbitrariness. If there are benefits or burdens to distribute, then, other things being equal, it is unfair to distribute them unequally. Giving more to some than to others for no relevant reason would be arbitrary.

11.3  Egalitarian Objectives

207

Scanlon expands upon the connection between equality and fairness by pointing out that large economic inequalities may threaten the integrity and impartiality of social institutions and practices. It is unjust, for example, if individuals accused of the same crime are less likely to be convicted if they are rich than if they are poor. Disparities such as these can call into question the legitimacy of courts, and the rule of law. The same applies to sentencing rules. A ten-​thousand-​dollar fine may be a modest setback for a top earner, but devastating to an unemployed person. (Traffic fines in Finland depend on incomes and have cost affluent speeders tens of thousands of dollars!) Second, Miller argues that some measure of equality is necessary for self-​respect. The great inequalities that characterize the United States today make it difficult for individuals at the bottom to maintain their self-​respect. These inequalities cause what Scanlon calls “stigmatizing differences in status.” Not only are the homeless, for example, impoverished and uprooted, they are also often regarded with suspicion, fear, indifference, and contempt by the more fortunate. Third, Miller points out that the duty to show equal respect often requires equal treatment. “While [people] differ profoundly as individuals in capacity and character, they are equally entitled as human beings to consideration and respect” (Tawney 1931, p. 34). Showing equal respect to people is a different matter from securing the circumstances in which people can maintain their self-​respect. Showing equal respect implies recognizing that people have the capacities to deliberate for themselves, to engage in relationships and activities that are intrinsically valuable, and to develop their capacities (Miller 1982, p. 81; see also Lukes 1977). The differences in the United States in the frequency of traffic stops for white and black drivers are not only unfair and harmful; they are also profoundly disrespectful. Finally, Miller argues that economic inequalities undermine social solidarity and create barriers to friendship, community, and love among members of a society. Large disparities in life prospects exclude the poor from the society’s achievements. As the social radical R. H. Tawney put it, “What is repulsive … is that some classes should be excluded from the heritage of civilization which others enjoy, and that the fact of human fellowship, which is ultimate and profound, should be obscured by economic contrasts, which are trivial and superficial” (Tawney 1931, p. 139). Such barriers are related to the stigmatizing differences in status to which Scanlon objects. To Miller’s list, Scanlon adds one important item. Large inequalities in wealth and income allow some people to have power over others. As we noted in earlier chapters, markets lessen the need to kowtow to the rich or high born, but with large inequalities, the poor may have little bargaining

208

Equality and Egalitarianism

power and thus may be locked into the terms that the more affluent decide to offer. Consideration of such power asymmetries plays an important role in objections to “sweatshop labor,” and they are also linked to objections to exploitation. While both Miller and Scanlon reject Temkin’s view that inequality is intrinsically bad, they show that concerns about inequality are deeply intertwined with moral ideas of fairness, self-​respect, equal respect, fraternity, and freedom from domination by others. Lessening inequalities in distribution is not merely a means to these values; under many circumstances, it may be a way of realizing these values.

11.4  The “Currency” of Egalitarian or Prioritarian Justice Because basic egalitarians maintain that inequalities are in themselves bad, and prioritarians maintain that benefits and burdens to those who are worse off are in themselves of greater moral importance, both groups need to address the question of which benefits and burdens should concern them. To use G. A. Cohen’s terminology, what is the “currency” of egalitarian justice? Should egalitarians compare people in terms of their (1) welfare, (2) opportunity for welfare, (3) resources, (4) capabilities, or something else entirely? In this section we will focus on interpretations of egalitarianism that differ concerning which goods should be equalized. Economists have had a great deal to contribute to that discussion. We shall postpone discussion of views of equality that do not focus on the distribution of divisible goods until Section 11.5. Economists might naturally interpret egalitarianism as aiming to equalize welfare and for that reason reject egalitarianism both because it relies on interpersonal welfare comparisons and because equalizing welfare would undermine incentives to work and invest. Equality of welfare is, in any case, not a very attractive ideal. Does the combination of Tiny Tim’s sunny disposition and modest wants cancel out his claims to a wheelchair (Cohen 1989, p. 918; Dworkin, 1981a, p. 241)? Few contemporary egalitarians defend equality of welfare. Those who think that welfare has a central place in egalitarianism focus on equalizing opportunities for welfare, which do not undermine incentives. Thus, Richard Arneson writes, “For equal opportunity for welfare to obtain among a number of persons, each must face an array of options that is equivalent to every other person’s in terms of the prospect for preference satisfaction it offers” (1989, p. 85). Equality of opportunity for welfare obtains when differences in the welfare people attain are due entirely to their free choices. By contrast, involuntary

11.4  The “Currency” of Egalitarian or Prioritarian Justice

209

expensive tastes, as well as disabilities, limit opportunities for welfare and call for compensation. To the extent that Tiny Tim faces fewer opportunities for welfare than an able-​bodied person, he deserves compensation. Unfortunately, opportunity for welfare is very hard to measure. Economists would have to distinguish the outcomes that are due to choice from those that are due to differences in opportunity sets. For example, Satz chose to give up practicing the piano a long time ago. Is the fact that she is not performing in Carnegie Hall due to her choice? Was that outcome ever a part of her opportunity set? Furthermore, as she adjusted her preferences in the light of her limited piano skills, how should one think of her opportunity set? And how should one compare the branching trees in her opportunity set with the opportunity sets of McPherson or Hausman, who took other paths? Le Grand (1991, ch. 6) suggests instead that egalitarians aim to equalize choice sets, leaving individuals responsible for which alternatives they choose among those open to them. Although more easily implementable than Arneson’s proposal, similar difficulties arise in determining what is and is not in the set of alternatives that are open to an individual. Ronald Dworkin argues that egalitarians should instead be concerned to equalize resources. Owing to people’s choices and characters, some will then do well and others badly, but Dworkin argues that that is up to them. Any further equalizing undercuts individual responsibility. Resource egalitarianism involves many complexities, since resources are diverse and it makes little sense for everyone to have exactly the same resources. Down parkas are not needed in Tahiti, and even with global warming there is not much demand for bathing suits in Greenland. Peanut butter is not a resource for those who are allergic to it. Resource egalitarians want everyone to have bundles of resources that are “equivalent,” rather than identical. When are bundles equivalent? One plausible answer is when nobody would want to trade her bundle of resources for anyone else’s. This “no envy” test provides one interpretation of what it is for resources to be equal. Dworkin suggests that we imagine how resources would be distributed if holdings were a competitive equilibrium in a hypothetical market in which all n people began with claims to 1/​n of every resource. Owing to differences in their tastes, people’s resources in equilibrium differ, but no one would want to trade for anyone else’s. It is possible to show, for certain simplified cases, that any resource distribution satisfying this no-​envy condition can be reached through a sequence of Pareto improving trades that start from a condition in which everyone has equal shares of the society’s resources. Following Kolm (1972), Hal Varian (1974, 1975) has proposed the label “fair” for envy-​free Pareto

210

Equality and Egalitarianism

optima and has defended this conception of fairness as ethically appealing (see also Foley 1967; Baumol 1986). An important stimulus to the study by economists of envy-​free allocations of resources has been their apparent ability to provide a standard of equity that does not require interpersonal comparisons of welfare. But there remain many problems, even in theory. The ethical appeal of resource egalitarianism derives from the intuition that, once people are equipped equally with resources, how much welfare they achieve is their own affair. But, starting from equal resources, economic outcomes can involve smaller or larger disparities, depending on social institutions such as insurance, taxation, and market structure. What rules should govern the transfer and accumulation of resources? Should egalitarians be indifferent to the magnitude of inequalities in outcomes, provided that resources are equal? Furthermore, outcomes depend on physical and intellectual abilities as well as on external resources and individual character and choice. Should not a person’s abilities count as resources, too? Productive abilities and consumption capacities are also resources –​internal resources –​and egalitarians should think about how they might be equalized or devise some compensation for inequalities in internal resources. One response would be to suggest that those factors that Dworkin and others regard as the individual’s responsibility such as a person’s willingness to expend effort may, ultimately, be traceable to their biological and personal histories. Why not then treat ambition as a resource? Furthermore, some people may possess enzymes that allow them to extract more nutrition from a meal or even more satisfaction from a movie. Following this route, one might argue that any difference in achieved welfare levels resulting from apparently equal resources is actually the product of hidden differences in resources (Roemer 1985, but see Roemer 2012 and Fleurbaey 2012). This line of thought leads to the conclusion that resources will be equal only if welfare is equal. To cope with differences in internal resources, Dworkin instead proposes a hypothetical insurance scheme (1981b, pp. 292–​323). Dworkin’s thought is that, if it were possible before birth to contemplate such matters, individuals would be willing to purchase insurance that would compensate them in the event they have the bad luck to be born with serious disabilities. Redistribution of income to improve the life prospects of those who are less well endowed with internal resources can be conceived of as implementing hypothetical insurance rather than as equalizing internal and external resources. Although a tempting thought, Dworkin’s proposal is unacceptable, because as Roemer (1985), Varian (1985), Fleurbaey (2002), and others

11.4  The “Currency” of Egalitarian or Prioritarian Justice

211

have pointed out, it could sometimes lead to directing resources away from those who are “unfortunate” rather than toward them. The alternative some commentators have pursued is to view total resources –​both internal and external –​as equal when people can achieve the same level of consumption –​that is, when people have equal amounts of goods and equal amounts of leisure. If everyone has the same tastes, then this conforms to Varian’s definition (1974, 1985) of a fair distribution:  it is Pareto optimal and envy free. But if people’s tastes differ, then outcomes involving equal goods and leisure will generally not be efficient, and the competitive equilibrium resulting from trades made after this equal-​division starting point will not be envy free. Starting with equal endowments of leisure and goods, a highly talented individual will be able to convert leisure time into additional goods at a much faster rate than will a less talented individual. In competitive equilibrium, the less talented will wish they could swap their goods–​leisure bundles for those of the more talented (Pazner and Schmeidler 1974; Varian 1985). This difficulty led economists in the 1970s and 1980s to put forward two alternative definitions of equal resources, whose “envy freeness” will not be disrupted by trade. The first of these is a “contribution fair” (Baumol 1986) or “wealth fair” (Varian 1974, 1975) distribution, which leaves talents or internal resources out of the package of what is to be equalized. Only external resources are equalized. So if no one worked (so that talents had no effect on consumption) and people’s tastes were the same, everyone would be equally well off. But people do work, and if only external resources are equalized, the more talented will be better off. The resulting competitive equilibrium will not be envy free in the sense we have defined. But, given how hard it would be for those with few talents to contribute as much to production as do the very talented, those with few talents will not envy the whole consumption–​production package. The second is an “income fair” distribution (Pazner and Schmeidler 1974; Varian 1974, 1975; Baumol 1986). Suppose everyone owns equal shares in the labor power (and hence the talents) of everyone. These shares and all external resources are divided equally. If everyone worked all the time and tastes were the same, everyone would be equally well off. But tastes are not all the same, and people cannot work all the time, so people will want to use some of their initial resources to buy back shares in their labor power so that they can have some leisure. If one defines people’s “full income” as the combined market value of their external resources, the shares they own in people’s labor, and their leisure, then the resulting competitive equilibrium equalizes full incomes  –​hence the term “income fair.” Because the

212

Equality and Egalitarianism

labor time of the talented is expensive, it will cost a talented person more to obtain free time than it will cost an untalented person. The talented, whose leisure is more expensive, will have fewer external resources and will generally be worse off. Dworkin calls this the “slavery” of the talented (1981b, p. 312), because, unlike those whose leisure is cheap, the talented will have no choice but to work long hours. Recent work generalizes and supersedes these accounts of income-​fair and wealth-​fair distributions (Maniquet 1993; Kolm 2004). To explain how, we need to clarify a fundamental difficulty with those luck egalitarian accounts that focus on resources as the metric of equality. Consider the following simple model due to Marc Fleurbaey (2008). Suppose people’s well-​being depends on the following: (1) resources, which are transferrable; (2) internal resources or endowments, which are nontransferrable and for which, if lacking, individuals are owed some compensation; and (3) effort –​ traits for which individuals are responsible and for which they should not be compensated. Whether to classify people’s tastes and preferences as endowments or effort is controversial, but the argument does not require that this question be settled. Suppose each individual’s welfare depends on her resources, endowments, and effort, whose levels or quantities can be indicated by single numbers. The luck egalitarian wants to distribute resources so that the welfare of individuals does not depend on their endowments; in other words, if individuals show the same effort, they should be equally well off. Fleurbaey calls this “the compensation principle.” At the same time, the luck egalitarian does not want individuals to be compensated for their level of effort; in other words, the external resources an individual receives should be independent of the individual’s effort. Fleurbaey calls this “the liberal reward principle.” The two principles are logically independent of each other. Providing no compensation at all is consistent with the liberal reward principle but not the compensation principle, while equalizing welfare without regard for effort would satisfy the compensation principle but violate the liberal reward principle. The two principles are not merely independent:  they are incompatible unless the contribution that effort makes to well-​being is separable from the contribution to well-​being of an individual’s resources and endowments. Suppose that welfare  =  resources + (endowments multiplied by effort). (Remember we are assuming that there are quantities representing each of these variables.) Consider two people with the same level of effort, but different levels of endowment, high endowment and low endowment. The compensation principle requires the luck egalitarian to compensate the individual with low endowment with a greater quantity of external

11.4  The “Currency” of Egalitarian or Prioritarian Justice

213

resources so as to equalize the welfare of the two individuals. But the quantity of external resources needed to compensate the individual with the low endowment depends on the level of effort. If the common level of effort among the two individuals were lower, the quantity of external resources needed to compensate the individual with the low endowment would be smaller. That violates the liberal reward principle, because the compensation depends on the level of effort. When the contributions that people’s endowments and their effort make to their welfare are not separable, there is no way to be, in Dworkin’s terms, both endowment insensitive and ambition sensitive.1 When it is impossible to satisfy both the compensation principle and the liberal reward principle, then economists and philosophers can either jettison one or seek some compromise. Because the impossibility arises from the interaction between endowments and effort, one can achieve a compromise between the compensation and liberal reward principles by equalizing welfare relative to some reference level of endowments or of effort. Doing so leads to the following two criteria: Conditional equality:  Distribute transferable resources so that if everyone had some reference level of effort, effort*, then everyone would be equally well off. Egalitarian equivalence: Distribute transferable resources so that everyone is as well off as he would have been if everybody had some reference level of endowments, endowment*, and the same quantity of external resources. Conditional equality conforms to the liberal reward principle since stipulating a reference level of effort implies that there is no compensation for levels of effort, or, in other words, that outcomes will be responsive to effort. Conditional equality does not conform to the compensation principle, but it does conform counterfactually. If everybody’s responsibility characteristics were at the reference level, then an individual’s level of endowments would have no effect on his or her utility, as the compensation principle requires. The larger the reference level of effort, the larger the difference in utility that would result from differences in endowments, and hence the more compensation that would be needed counterfactually to satisfy 1 Let welfare be W, resources R, endowments E, and effort Z. W = R +EZ. If two individuals have the same level of effort Z*, then the compensation principle demands their welfare should be the same, which means that R1 + E1Z* = R2 +E2Z*. So R1 –​ R2 = (E2 –​ E1)Z*. This relationship violates the liberal reward principle because compensation in the form of external resources, R, depends on the level of effort, Z*.

214

Equality and Egalitarianism

the compensation principle. Wealth-​fair and income-​fair distributions are extreme versions of conditional equality. In wealth-​fair distributions the reference level of effort is zero and welfare thus depends only on external resources, which are equalized. In income-​fair distributions, the reference level of effort is maximal and the compensation required is extreme. By setting the reference level of effort at some intermediate value, one can define distributions that are more ethically attractive. Egalitarian equivalence (see also Pazner and Schmeidler 1974, 1978) satisfies the compensation principle, because, with resources and endowments fixed, those who make the same effort will be equally well off. It satisfies the liberal reward principle counterfactually by making its hypothetical transfers depend only on the hypothetical equal endowments and not at all on effort. Egalitarian equivalence calls for distributing differing positive and negative amounts of transferable resources so as to place people at the utility levels they would have had with (a) their actual level of effort, (b) the reference level of endowments, and (c) a purely hypothetical equal transfer. Another way to think about egalitarian equivalence is as a way of measuring relevant inequality. One first supposes that everyone with her actual level of effort had the reference level of endowments and then asks what quantity of external resources would provide every individual with her current level of well-​being. That “egalitarian equivalent transfer” measures who is worse off and by how much –​and it does so without requiring any interpersonal welfare comparison. Egalitarian equivalence as a criterion of equality requires that one then distribute actual transferable resources in such a way that the postdistribution hypothetical egalitarian equivalent transfers would be equal. There is one other view of the currency of egalitarian distribution that we should mention. Equalizing welfare pays too little attention to individual responsibility and to problems of preference adaptation, as in the case of Tiny Tim. On the other hand, equalizing external resources –​however radical this may appear in comparison with actual societies –​seems to equalize too little. The same income does not make possible the same range of activities and experiences for those who have disabilities as for those who do not have disabilities. Rather than assimilating disabilities and talents to resources (like those who emphasize equality of resources) or focusing on their implications for welfare (like those who argue for equality of opportunity for welfare), one might instead attempt to equalize what Cohen calls “access to advantage.” In his view, egalitarians ought to look at what is, as it were, between resources and welfare. Resources and internal features of the individual determine the

11.5  Relational Equality and Equality of Moral Status

215

range of outcomes from which the individual chooses. Cohen argues that egalitarians should try to equalize this range of outcomes. This is similar to Sen’s proposal that egalitarians seek equality of capabilities, where capabilities consist of sets of those “functionings” (activities and experiences) that are accessible to an individual. Though attractive, these proposals of Sen and Cohen face many of the problems encountered by those who would equalize resources or opportunity for welfare. Was being a concert pianist ever in Satz’s capability set? Given the measurement difficulties, it might seem that the only way to implement Sen’s or Cohen’s proposals would be to focus on resources. In response to such a worry, Martha Nussbaum (2000) argues that, by specifying a list of central human capabilities and insisting that these be available to everyone, one can formulate a practical egalitarian program. The formal egalitarian criteria we have discussed are only a small sample of a wide range of technical construals of egalitarian objectives that have been developed over the past decades with respect to a wide range of allocation problems (see Thomson 2011; Fleurbaey and Maniquet 2005, 2011, 2012). Since these formal construals of egalitarianism do not limit themselves to utility information (Fleurbaey, Suzumura, and Tadenuma 2005), there are a great many possibilities we cannot survey, many of which involve complex formal theorems.

11.5  Relational Equality and Equality of Moral Status Elizabeth Anderson (1999) argues that the debates concerning what to equalize and whether to favor egalitarianism or prioritarianism have “lost sight of the distinctively political aims of egalitarianism. The proper negative aim of egalitarianism is not to eliminate the impact of brute luck from human affairs, but to end oppression” (1999, p. 288). She argues that “luck egalitarianism” –​whether formulated as equality of resources, equality of opportunity for welfare, or equality of capabilities  –​“fails the most fundamental test any egalitarian theory must meet: that its principles express equal respect and concern for all citizens” (1999, p. 289). Luck egalitarianism fails this test because it ignores disadvantages for which individuals are responsible, because it rests the case for redistribution on the inadequacies or bad luck of the unfortunate, and because it rests on intrusive judgments concerning the sources of inequalities. In place of luck egalitarianism, Anderson advocates what she calls “democratic equality,” whose focus is on securing the social conditions of freedom and the bases for equal respect. Rawls’s theory of justice (to be discussed in Chapter 12) arguably manifests

216

Equality and Egalitarianism

this kind of egalitarianism in its inquiry into what resources, rights, and liberties citizens need in order to relate to one another as equals. As has emerged by now in our discussions, egalitarianism is a family of positions. Even within the “relational equality” camp, some agree with Anderson that the point of equality is to secure the conditions for equal freedom; others focus on whether government treats its citizens fairly or whether certain inequalities are stigmatizing. What is common to these positions is a concern with the relations among people. Rather than seeking equality in the distribution of goods, relational egalitarians seek a society of equals. Securing such a society constrains inequalities in particular goods, but what is central to relational egalitarians is how individuals relate to one another, not how things are distributed among them. Although not couched in exactly this way, Michael Walzer’s “complex equality” is closely related (1983; see also Miller and Walzer 1995). Walzer points out that the principles governing distribution within disparate aspects of social life differ considerably. Resources within families are not distributed as wages, profits, or rents. Licenses to practice medicine are not sold to the highest bidder. Political life is governed by principles designed to protect equal rights. Political rights are free, and they are neither distributed as rewards nor (except in the case of convicted felons) revoked as penalties; they cannot be bought or sold (Okun 1975, pp. 6–​10; Radin 1987, 1996). Inequalities from within economic life do, of course, spill over into political life. But this is widely thought to be undesirable, and nations erect barriers aimed at limiting the extent to which wealth leads to political power. Walzer maintains that different distributive standards should prevail in different “spheres” of social life. Even within areas of activity that are recognizably economic, he argues that one should resist the supposition that they should all be governed by a single underlying standard. For example, Walzer maintains that the social significance of employment in modern industrial societies argues strongly for equal opportunity in employment. On the other hand, Walzer would deny that our society should underwrite a similar commitment to equality of opportunity in consumption. Walzer’s egalitarianism does not rule out economic inequalities, inequalities in authority within particular institutions, or even political inequalities that arise from differing abilities to persuade, organize, and lead. These inequalities are acceptable so long as there are equal political rights and liberties and so long as the inequalities in one domain do not spill over into others –​in other words, so long as the inequalities are not pervasive. A society of equals is bound to be a society bursting with differences, but these differences must not congeal into pervasive inequalities. On this view, what

11.6  The Measurement and Importance of Inequality

217

is wrong with the inequalities in wealth and income in the United States is that, given their magnitude, they necessarily lead to inequalities in status, in employment opportunities, in political power, even in health. Although Walzer argues that there is no reason to eliminate economic inequalities completely, he nevertheless provides good reason to reduce them and to rein in their influence. Furthermore, Walzer argues that some ostensibly economic inequalities, such as those involved in the authoritarian structure of economic corporations, are actually political in nature and are as inconsistent with equality in rights and liberties as are towns that are owned and controlled by corporations (1983, ch. 12). Walzer’s complex equality and other variants of relational egalitarianism (Scheffler 2003, O’Neill 2008) seem to follow naturally from concerns with fraternity, self-​respect, equal respect, and the avoidance of domination. The quotations from Tawney that we used to illustrate the concerns for fraternity and equal respect are in much the same spirit as Walzer’s views. Although Walzer’s complex equality, like other relational egalitarian views, demotes questions of distribution from the central place they occupy in luck egalitarianism, distributional questions are still important. Because relational egalitarians defend limitations on the permissible degree of economic inequality, work clarifying the magnitude and importance of inequalities within the economic sphere continues to be of great importance. Just as one can ask, “Why be concerned with equality per se?” and “Why weight the interests of the worse off more heavily?” so one can ask, “Why care about equality of moral status or ‘complex’ inequalities?” Answers to ethical questions give out when one reaches fundamental notions, and we think that at this point one is close to the foundations of morality. As Griffin puts it, “To see things morally is to grant everyone some sort of equal standing” (1986, p. 295). As we mentioned at the beginning of this chapter, virtually all moral theories find their origin in some notion of equal respect owed to moral agents, and concerns about equality of moral status derive from a concern not about distributional equality itself but rather about a particular kind of status. All should have an equal moral status because everyone should have the moral status of a person; in possessing that moral status, all are equal to one another.

11.6  The Measurement and Importance of Inequality If one can determine what one might want to equalize and why (which, as we have shown, is difficult to do), one must then say something about what weight to place on egalitarian concerns when they conflict with

Equality and Egalitarianism

218

100% fraction of income

Perfect equality X

0 100%

0 fraction of the population

Figure 11.1.  The Gini coefficient.

other moral considerations. Arthur Okun makes this question vivid by means of an analogy to a leaky bucket (1975, pp. 91ff.). Suppose that one redistributes goods from the rich to the poor in a leaky bucket. How much leakage should one accept before deciding that efforts at redistribution are too costly? Redistribution has administrative costs, effects on incentives, and effects on attitudes toward work and care of oneself. Of course, not all policies addressing inequalities need be redistributive. Changes in patent law, licensing, and immigration could have major effects on distribution, and they might not have any cost in terms of efficiency. But there are obviously sometimes trade-​offs between equality and efficiency or between equality and other values. (Equally clearly there will be cases in which, measured in dollar terms, redistribution from the more to the less well off will increase total income. Consider, for example, eliminating tax loopholes that benefit the well off and distort the allocation of resources.) When there are trade-​offs, how much weight should be placed on equality? This is a difficult question. Clearly the extent of trade-​offs will be sensitive to the conception of equality that is under examination, and it will vary depending on the society’s degree of inequality relative to its level of welfare or freedom. In order to pose these questions meaningfully, one needs some way to measure inequality. Otherwise, there is no way to compare a reduction in inequality and an increase in welfare. But comparing the inequality of different distributions turns out to be a surprisingly complicated matter, and there are a variety of different measures in use (Atkinson 1970; Temkin 1993). The most common measure is the Gini coefficient. Consider Figure 11.1.

11.6  The Measurement and Importance of Inequality

219

If one lines up the population from the poorest to the richest, this graph shows how much of the total income each fraction of the population has. If everybody’s income were equal to everybody else’s, then each additional 10 percent or 20 percent of the population has an additional 10 percent or 20 percent of the income and the 45-​degree line would describe the distribution of income. If income is unequal, then the distribution will be represented by something like the curved line in the figure. The Gini coefficient is the ratio of “X” –​the area between the 45-​degree line and the line representing the actual distribution –​to the area of the entire triangle. It ranges from 0, which would be perfect equality, to 1, complete inequality. The Gini coefficient can be used to measure inequalities in wealth as well as inequalities in income. According to the CIA, the Gini coefficient for income in Sweden after taxes and transfers was .25 in 2000, while the Gini coefficient for the United States in 2007 was .45 (up from .36 in the late 1960s). Before taxes and transfers, the differences are much smaller: .43 for Sweden and .49 for the United States. The Gini coefficient is not a perfect indicator: it can go up, even while poverty is decreasing. Societies with different distributions of wealth can have the same Gini coefficient in cases where their Lorenz curves cross. The measure also gives more weight to distribution at the center, rather than at the margins. Other measures of inequality have been proposed by Thiel (1979) and Atkinson (1970). Of course, any one-​dimensional measure will only capture some of what is of interest to us in studying inequality. Most egalitarians have contended that economic inequalities in modern industrial societies could be diminished without serious loss of efficiency or freedom and without violating anyone’s rights. Indeed, it is plausible that equalizing educational opportunity in the United States would increase, not diminish, both efficiency and freedom (though not in the libertarian’s sense). Economies such as Japan, Austria, and Sweden demonstrate that inequalities in wealth and income that are small relative to those in the United States are compatible with freedom and economic growth. The extent of these trade-​offs between equality and economic growth is ultimately an empirical question, but how much weight to place on equality is a matter of values. In the United States over the past forty years about 15 percent of national income has been shifted from the bottom 90 percent of the income distribution to the top 10 percent. Instead of damping this massive pretax shift in income, increases in the payroll tax, which fall heavily on those who are relatively poor, coupled with dramatic decreases in income, dividend, and estate taxes paid by the

220

Equality and Egalitarianism

wealthy, have increased inequalities. For economists to contribute most effectively to social deliberation about the wisdom and justice of these policies and about how to make trade-​offs between equality and other values, they need to grasp what inequalities egalitarians object to and what grounds their objections.

Suggestions for Further Reading There are three useful anthologies of works on egalitarianism: Pojman and Westmoreland (1997), Mason (1998), and Clayton and Williams (2000). See also Hirose (2014). Crucial texts addressing issues in contemporary egalitarianism and prioritarianism are Anderson (1999), Arneson (1989, 1990, 2000), Cohen (1989, 1993), Dworkin (1981a, b, 1988, 1990, 2000, 2002), Frankfurt (2015), Rawls (1971) Roemer (1985, 1987, 1988, 1996, 1998), Scanlon (1986), Scheffler (2003), Sen (1992a), Temkin (1993, 2003, 2015), and Williams (1962). For discussions of prioritarianism, see Parfit (1991) and Fleurbaey (2015). Adler (2012) brings together utilitarianism, prioritarianism, and economic work on social welfare functions. For an accessible discussion of Roemer’s way of understanding equality of opportunity, see the April/​May 1995 issue of the Boston Review, http://​ bostonreview.net/​BR20.2/​BR20.2.html. Philippe van Parijs (1990), developing suggestions of Bruce Ackerman (1981), has put forward a weaker egalitarian notion of “undominated diversity of resources” designed to allow for the fact that different people may rank alternative bundles of resources differently. Arguing from a commitment to undominated diversity and other moral principles, van Parijs has made a persistent case for an unconditional grant to be paid to every adult citizen (1990, 1991, 1995; see also www.usbig.net/​whatisbig.php). Ackerman and Alstott (1999) have defended the somewhat different idea of a stakeholder society in which each individual, upon adulthood, would be granted a sizable “stake” to be used for investment, education, or security. Something like their proposal was actually implemented for a while in the United Kingdom in the form of so-​called Baby Bonds (see Paxton 2002). Atkinson (2015) makes a number of detailed suggestions of policies to lessen inequalities. See Fleurbaey (1995, 2002) for a challenge to the direction in which the discussion of egalitarianism has gone. For reviews of technical work on fair allocation, see Moulin and Thomson (1997), Fleurbaey and Maniquet (2005), and for more recent formal construals of equality and fairness see Fleurbaey and Maniquet (2011 and 2012), Maniquet (2007), and Thomson

Questions for Study and Discussion

221

(2011). For an argument that equality and efficiency go hand in hand rather than competing, see Bowles et al. (1999).

Questions for Study and Discussion 1. Income and wealth disparities in the United States are large and growing with CEOs making four hundred times the average salaries in their firms and the bottom 40 percent of the U.S. population owning far less than 1  percent of the total wealth. Do you think that these disparities are too large, too small, or just right? Why? How should income and wealth be distributed? 2. The chapter gives the example of four different wage policies: (A) pays all the workers the same wages, (B) pays everyone the same piece rates, (C) pays all the workers with the same seniority the same wage with higher wages for those with more seniority, (D) pays all the workers who have the same number of dependents the same wage, with higher wages for those with more dependents. Which of these policies should an egalitarian favor? Why? Would different versions of egalitarianism disagree about which of these policies is most egalitarian? 3. What are the most important reasons to be concerned about wealth and income inequalities? What do you think of these reasons? 4. Do you think that inequalities among groups matter, or should policy makers be concerned only about inequalities across individuals? Do groups and group based inequalities have any importance? Why or why not? 5. Do you think that on average the rich deserve their incomes? Why or why not? How can we tell? What constitutes “desert”? 6. Do you think that there is anything better about a poor society in which there are few inequalities in income than a society in which everyone is richer than in the first, but there are large inequalities? How would you defend your answer? 7. What exactly is Miller’s reservation about egalitarianism, and what is his solution? Do you think that he has successfully solved his problem? 8. How important are the egalitarian objectives discussed in Section 11.3 compared to promoting people’s welfare? How would you defend your answer? 9. Do you think that leveling down is ever acceptable? Can you give an example in which leveling down might be ethically attractive?

222

Equality and Egalitarianism

10. Why would most economists reject equality of welfare as an objective of social policy? Do you agree? Why? 11. What are positional goods and why are they important? When people are competing for positional goods, are the results efficient? Why or why not? 12. Why is it hard to know whether people’s opportunity for welfare is equal? Is this a sufficient reason for policy not to aim at equalizing opportunities for welfare? 13. Why do luck egaliltarians maintain that equalizing material resources is insufficiently egalitarian? What should a resource egalitarian require? 14. What does it mean to say that a distribution is “envy-​free”? What is the connection between a distribution’s being envy free and its being egalitarian? 15. What is the problem with demanding simultaneously that people should be compensated for shortfalls in internal resources (i.e., that they are not responsible for the shortfalls) and that people should not be compensated for having low levels of effort or ambition (i.e., that they are responsible for their effort)? 16. Do you think that an egalitarian should be content with Walzer’s complex equality, or do you think that an egalitarian should demand more? 17. Do you think that diminishing inequalities in the United States would lessen the rate of growth and make people worse off? If so, do you think that the trade-​off would be worthwhile?

T WELVE

Justice and Contractualism

Among the questions of justice that are important to economists are those that concern the distribution of benefits and burdens among members of a society. What claims can persons legitimately make upon one another or upon the state? What burdens can the state place on its citizens, or individuals on one another? Because justice is concerned with the distribution of scarce resources in society, its subject matter overlaps with economics. Economists, like philosophers, cannot decide what principles of justice to rely on by consulting public opinion, because people are committed to different principles, and even when there is a social consensus, it may not be well considered, or it may be so abstract as to offer little practical guidance. As an example of poorly considered agreement, consider that for centuries people believed that women were the natural inferiors of men. Consider as an example of ambiguity and overabstractness, equality of opportunity: almost everybody accepts this value, but there is little agreement about what it entails. Economic evaluation presupposes at the very least well-​defined principles of justice. The principles of justice that guide policy should also be well considered. Many issues of justice are controversial, and economists would like to avoid them. But controversial questions are often unavoidable. Suppose, for example, economists are called upon to describe the economic consequences of alternative policies affecting families. Should they evaluate these policies by looking at their effects on the household or on each individual member? Should economists include the unpaid contributions women make to the household in GNP? Do changes in family law such as no-​fault divorce that harm women and children in terms of standard of living, while also allowing women and men to leave unhappy marriages more easily, increase or decrease social welfare? All these questions involve matters of justice, and one cannot choose among policies without answering them. 223

224

Justice and Contractualism

Libertarian and utilitarian moral theories, the subjects of Chapters  7 and 10, imply theories of justice. For most libertarians, justice primarily involves respecting property rights and negative freedom. For most utilitarians, principles of justice are utility-​maximizing general rules that facilitate cooperation and regulate conflicts of interest. But there is a different idea of justice than the libertarian or the utilitarian that appeals to the idea of agreement. Roughly, the idea is that, as the terms of social cooperation, principles of justice are those that people under certain conditions would rationally agree to, or at least would not find it rational to reject. Sometimes this idea of agreement is stated in terms of a social contract, although the contract metaphor should not be taken literally. This chapter describes the interest among economists and philosophers in theories of justice in the social contract tradition.

12.1  The Social Contract Idea Human interactions involve both conflicts of interest and mutually beneficial cooperation. Circumstances falling between complete harmony of interests, which would make justice superfluous, and desperate scarcity and universal mistrust (“the war of all against all”), which would render it hopeless, are called by Hume (1777) and by Rawls (1971) “the circumstances of justice.” It is tempting to think of the normative principles governing individual interactions as human contrivances to adjudicate conflicts of interest and to secure the benefits of cooperation. Norms might thus be regarded as arising from social agreement. The idea is already set forth in Plato’s Republic, but Plato denies that the principles to which individuals would agree in order to facilitate living together are truly principles of justice. In the seventeenth and eighteenth centuries, following the collapse of a medieval and Renaissance vision of society as naturally hierarchical, political philosophers revived the notion of a voluntary social contract as the source of political obligations and social norms, including principles of justice, for people who are each other’s equals. The basic notion of social contract theories is that one judges principles and institutions by asking whether individuals would agree to them under some specified conditions, often called a “state of nature.” In a state of nature individuals are conceived of as not already bound by socially created obligations but only as governed by general moral norms (which seventeenth-​and eighteenth-​century philosophers called the “law of nature”). For example, Locke argues that in a state of nature all are under the injunction to seek the peace and preservation of mankind. Locke goes on to argue that people in

12.1  The Social Contract Idea

225

a state of nature would find it rational to agree to a form of limited government with separation of powers; that is why he thinks that residents of a country are obligated to obey the state. Not all philosophers have found contractualist reasoning valuable. David Hume, for example, famously complained that no original contract can explain or justify current obligations to society:  why should anyone be bound by what some of her (hypothetical) ancestors did? And, Hume argues, if there were no reason to keep promises without a social contract, promising to adhere to such a contract could not give rise to one. However, advocates of contractualism find the link between rational consent and justice attractive. Showing that a certain principle would be agreed to by people acting rationally in certain well-defined circumstances increases its claim to guide actions, to the extent that agents find those circumstances a reasonable context for their deliberation. All contract theories test moral principles by asking whether rational agents would agree to them under properly defined circumstances, but they differ in the interpretation of the circumstances of choice, in their characterization of the agents’ rationality and motivation, and in their view of what is to be chosen. These differences lead to large variations among contract theories. Brian Barry (1989) has helpfully distinguished two broad categories of contract theory, especially as applied to theories of justice. The first, exemplified by Thomas Hobbes, links human rationality to self-​interest and conceives of agreement as the outcome of bargaining among rational self-​ interested agents. Barry refers to this idea of justice as mutual advantage. On this view, the social contract is a compromise that enables individuals with their differing abilities and resources to pursue their separately given private aims successfully. The second category, exemplified by Kant and Rousseau, aims at specifying a social contract that is not influenced by individuals’ differing bargaining power and that no rational individual who recognizes others as equals, governed by the same rules, could reject. Barry refers to this idea of justice as impartiality. The social contract determines what principles governing interactions would be endorsed by agents who are impartial among the different private ends that specific individuals have. Unlike utilitarians, who characterize equal respect in terms of counting everyone’s interests equally, or libertarians, who characterize equal respect in terms of equal rights that place independently given limits on what people can agree to, contractarians interpret equal respect as the equal moral status of the parties to the social contract. Thomas Scanlon maintains that a contractarianism based on impartiality requires one to “justify one’s actions to others on grounds that they could not reasonably reject –​reasonably, that

226

Justice and Contractualism

is, given the desire to find principles which others similarly motivated could not reasonably reject” (1982, p.  116, 1998, ch. 5). Scanlon’s formulation highlights a feature common to contractualist views: the interest not only in identifying defensible moral principles but also in giving an account of why people should feel bound by those principles. Since theorists of mutual advantage adhere closely to a view of rationality as self-​interest, a major challenge for their views is to explain why rational persons would agree to restrain their pursuit of their interests and especially why they would adhere to agreements once made. Those who see contractualism as determining what is impartially best instead assert that people may be moved by a desire to regulate their conduct by principles that are morally acceptable from a suitably detached perspective.

12.2  Justice as Reciprocity: Rawls’s Theory of Justice The most influential work in modern political philosophy is John Rawls’s A Theory of Justice (1971), an epic exposition of a contractualist political theory adapted to contemporary circumstances. It is worth explaining Rawls’s views at length because of their central role in contemporary political philosophy and because Rawls’s argument draws on insights from economics. His theory also has a great deal to offer to economists and those interested in public policy. Rawls seeks to remedy some of the problems with previous elaborations of social contract theory. In particular, his argument is explicitly normative and does not depend on any account of a previously existing state of nature; he does not assume that everyone shares the same ends; and he does not see his theory as “deriving” principles of justice from claims about what people would choose in a state of nature. Rawls also aims to provide a well-​ worked-​out alternative to utilitarianism, which he takes to be the dominant theory of justice used in public life and appealed to in the social sciences. Rawls’s arguments adhere to what he calls the method of “reflective equilibrium.” Those thinking about theories of justice begin with widely shared beliefs in some principles of justice, for example, that it is unjust to steal or that slavery is wrong. They may also have commitments to specific moral judgments, for example, that it is permissible for a mother to steal food for her starving children. The method of reflective equilibrium consists in adjusting in a back-​and-​forth way these various beliefs and commitments with the aim of achieving a body of moral beliefs that is consistent and in which specific moral judgments can be explained and justified by the moral principles one endorses. Rawls holds that his principles should be

12.2  Justice as Reciprocity: Rawls’s Theory of Justice

227

accepted by readers if and only if they are in reflective equilibrium with the reader’s considered moral judgments and more general views concerning the nature of justice. Thus, he attempts to use only modest and uncontroversial assumptions to reach his conclusions (although his critics question his assumptions and whether they justify his conclusions). The “contractualist” element in Rawls’s theory derives from his claim that his principles of justice would be rationally preferred to a list of alternatives (including libertarianism and utilitarianism) in what he calls “the original position.” In this position, individuals are to think of themselves as choosing behind a “veil of ignorance” that deprives them of knowledge of their social position, race, gender,1 occupation, income, and wealth and even their preferences for one kind of life rather than another. This bracketing of information forces individuals to be impartial in their judgment of competing principles of justice. They cannot choose principles that will be to their own particular advantage, because they have no idea of what that particular advantage is. The veil of ignorance makes vivid the compelling idea that a society’s principles of justice must be capable of agreement by all rational individuals, at least insofar as they are reasonable. It would, for example, be unreasonable for one agent to impose principles of justice on others solely on the basis of his or her own personal advantage. Rawls assumes, for the sake of his model, that agents are neither altruistic nor envious. They are also ignorant of the range of possible social outcomes and of the relative size of different groups; they are to be understood as having no way of estimating their chances of winding up well off or not. Thus their decision situation is one of radical uncertainty rather than calculable risk. While they do have basic information about human psychology, science, and the facts about how economies function, they have no information of a more personal kind that they could bring to bear on their ranking of alternative principles of justice. This thought experiment is meant to capture a widely shared understanding of the impartial manner in which principles of justice should be chosen. Though Rawls defends the veil of ignorance as an appropriate device for choosing the principles of justice that should govern the basic institutions of society, there are other circumstances to which he deems this kind of reasoning inappropriate. For example, if one is deciding whether Mary In A Theory of Justice, Rawls takes individuals behind the veil of ignorance to be heads of households. Okin (1989) argues that, like most theorists of justice, Rawls ignores the problems of justice within the family and the problems of justice posed by children and other dependents. 1

228

Justice and Contractualism

ought to pay George $250, then one needs to know what George’s rights and entitlements are and whether Mary made promises to George or stole his property. In deciding what individuals should do –​what careers they should pursue, for e­ xample –​the original position is not the right way to represent the basis of the decision. Nor does Rawls think that this decision procedure should be directly applied to questions of international or global justice. For that purpose, Rawls (2001b) proposes that the original position be recast as applying to the various “peoples” of the world, with each people deliberating over ground rules for global society from behind a veil of ignorance concerning its relative standing on the world stage. The crucial point is that the original position procedure is meant to be applied only to the question of which principles of justice should govern what Rawls calls the “basic structure of society” –​the economy, the political system, the family structure. These are the aspects of society whose influence is inescapable, involuntary, and profound. Looking to the original position is not a procedure that is appropriate to use in deciding what movie a group of friends should see. There is, however, some controversy about Rawls’s sharp distinction between the principles that should govern institutions and those that should govern our individual decisions (Cohen 2009; Murphy 1999). James Buchanan draws a related contrast between the justice of rules and the justice of behavior within rules (Brennan and Buchanan 1985). Rawls’s framework borrows from one developed independently by John Harsanyi and by William Vickrey, although Harsanyi and Vickrey used the framework to defend utilitarianism. They argued that, when faced with uncertainty about one’s social role, the rational choice would be to maximize the expected value of one’s utility, averaging over all the possibilities. The proper social principle is thus to maximize the average utility of everyone in society. Rawls objects that the principle of average utility is too risky to choose in this special situation. Maximizing utility for the whole society could require completely sacrificing the interests of some. It is unreasonable to expect people to agree to sacrifice their entire life prospects to benefit others to whom they have no special tie. Put more simply, Rawls rejects maximizing average utility because it is in principle compatible with terrible outcomes for some people such as slavery. Rawls argues that from behind the veil of ignorance rational agents would reject average utility as the supreme principle of justice, because it cannot guarantee a satisfactory minimum. Furthermore, the situation of uncertainty, as Rawls defines it, makes the calculation of expected utility virtually meaningless:  with no

12.2  Justice as Reciprocity: Rawls’s Theory of Justice

229

knowledge of the range of possible social groups or of the relative size of groups, one is pushed toward approaches to choice under uncertainty that do not require probability calculations. Finally, one cannot maximize utility (the satisfaction of preferences) without knowing what one’s preferences are. Rawls argues that, in the special circumstances of choosing principles of justice behind the veil of ignorance, it is rational to maximize the prospects of the social group that is worst off. It must be possible to justify social arrangements to everyone. When those among the worst off inquire why justice is compatible with their doing so badly, they can be reassured that there is no way to arrange society so that the worst-​off group would be better off than members of this group are. More specifically, Rawls argues that –​for societies that are at least moderately well off –​rational choosers would seek first to safeguard their basic political and personal liberties and would then judge basic social and economic institutions according to their tendency to promote the primary goods of those in the least-​well-​off social group. No one would find their basic liberties threatened, and social and economic institutions would be designed to advance maximally the interests of the worst-​off group. It is important to note that these interests are to be judged not by the subjective standard of preference satisfaction but by the more objective notion of amounts of primary social goods, a notion that even individuals deprived of knowledge of their own view of the good would agree on as a standard for assessing basic institutions. In A Theory of Justice, Rawls describes social primary goods as “all purpose means” whose distribution is socially determined. Primary goods are those things that contribute to everyone’s well-​being or at least do not diminish the extent to which anybody’s preferences are satisfied. Rawls includes among primary goods wealth and income, opportunities, and the social bases of self-​respect. In later work, in contrast, Rawls takes primary goods to be those things that individuals need in order to exercise what he takes to be the two central moral powers of individuals: the capacity to set one’s objectives or to define what one takes to be good and the capacity to consider principles of justice and to make one’s behavior conform to what justice allows. In their final elaboration (2001a, pp. 42–​43), Rawls’s principles of justice are as follows. [1]‌ Each person has the same indefeasible claim to a fully adequate scheme of equal basic liberties, which scheme is compatible with the same scheme of liberties for all;

230

Justice and Contractualism

[2]‌ Social and economic inequalities are to satisfy two conditions: first, they are to be attached to offices and positions open to all under conditions of fair equality of opportunity; and second, they are to be to the greatest benefit of the least-​ advantaged members of society (the difference principle).

Social and economic inequalities are measured in primary goods. Fair equality of opportunity entails that those similarly talented and motivated should have the same chances to occupy positions in society regardless of the social class into which they are born. Together these two principles define a conception of justice for a democratic society, a “society of equals.” When these principles are satisfied, all of a society’s members are being treated with equal concern and respect:  even those who are least advantaged by this arrangement have no legitimate grounds for complaint. Utilitarianism, because it allows that some can be sacrificed for the good of the many, and because it does not firmly secure the basic liberties and opportunities, fails to treat those who are disadvantaged with equal concern and respect. Rawls specifies that the first principle is, for any society with enough resources to provide the basic liberties for all, lexically prior to the second. That is to say, basic political and civil liberties are secured; they are not to be traded off for economic advantage. Moreover, the first part of the second principle is lexically prior to the second –​the interests of the least well off are not to be advanced at the sacrifice of the principle of fair equality of opportunity. The “priority of liberty,” as the first principle is sometimes labeled, is not easy to capture in economic models, and economists are uncomfortable with a lexicographic ranking, which here rejects the notion of trading off liberty for other valued ends (but see Cooter 1989). Rawls maintains that, after one attains a basic minimum of resources, further means to pursue one’s goals are valueless unless one preserves the basic liberties that permit individuals to pursue these goals effectively. To permit liberty to be exchanged for resources betrays a failure to understand the importance of one’s deepest commitments. Economists have also raised doubts about the difference principle –​the view that society should arrange inequalities to maximize the primary goods of the least advantaged. If one moves outside the Rawlsian framework and takes the difference principle as applying to welfare rather than primary goods, it implies that (subject to the first principle and fair equality of opportunity) society should arrange inequalities to maximize the well-​being of the least well off. This maximin rule is as simple as the familiar utilitarian rule to maximize average well-​being, and economists

12.2  Justice as Reciprocity: Rawls’s Theory of Justice

231

have investigated implications of these two ideals (see, e.g., Phelps 1973; Alexander 1974; Musgrave 1974). Rawls rejects this welfarist interpretation of the difference principle. In particular, a policy that aims to maximize the well-​being of the least well-​ off individual sets up an impossible informational demand for any state. How could any state agency ever know whose preferences are least well satisfied? Rawls seeks instead an objective standard of disadvantage, and he defines the least advantaged in terms of “representative” individuals or social groups, not particular individuals. In that way, Rawls’s view, unlike the view that seeks to maximize the well-​being of the worst off, recognizes that outcomes depend on individual choices. As Rawls puts it, the surfer who wants to surf all day in Malibu is not entitled to public aid:  such a person does not simply by dint of his choices count as the least advantaged. Rawls’s theory defines the worst off as that group whose members have the fewest primary goods. If Rawls is to hold surfers accountable for their preferences, he has two moves open to him. First, he can add leisure to the list of the primary goods, a modification he considered in response to an objection raised by Richard Musgrave (1974). Alternatively, he can press on the importance of reciprocity in his theory. The surfer who spends all of his time on the beach and expects others to provide for him is a canonical example of a free rider. No rational person behind the veil of ignorance would consent to be in a society where others could ride free on their labor simply because of the latter’s expensive tastes. To simplify the task of constructing a theory of justice, Rawls idealizes and assumes that no one making the agreement suffers from disabilities or other health problems. Philosophers such as Nussbaum and Sen (1993), Nussbaum (2000), and Kittay (2013) have accordingly criticized him for ignoring important aspects of the human condition such as dependency, disability, and ill health, but in abstracting from these difficulties, Rawls never denied their importance. Indeed, in his last book, Rawls (2001a, pp. 173–​174) briefly seconds Norman Daniels’s suggestion that fair equality of opportunity implies a right to health care. More recently, epidemiologists have argued that there are pervasive and significant health differences between those of higher and lower social and economic status that cannot be explained only by differing access to health care (Marmot 2004; Wilkinson and Pickett 2009). Not only do the poor have much worse health and much lower life expectancies than those who are not poor, but even those who are relatively affluent, enjoy high status, and have excellent health care have worse health than those

232

Justice and Contractualism

who are at a still higher status. Researchers find that life expectancy drops twenty years over the few miles that separate downtown Washington, D.C., from its impoverished neighboring communities. While it is possible that poverty and inequalities in education rather than inequalities in wealth are driving some of the disparities and that inequalities in health are driving the inequalities in wealth (Deaton 2013), researchers find over and over a gradient whereby those who are wealthier are also healthier. For example, in England engineers and doctors live on average three years longer than schoolteachers. Even in egalitarian Sweden, Swedes who have Ph.D.s have lower mortality than Swedes who have B.A.s (Marmot 2004). Daniels speculates (2007) that a society governed by Rawls’s two principles of justice would largely eliminate these inequalities. Given how important health is to living well and given the magnitude of these effects, we think it is questionable whether the principles of justice governing the basic structure can be worked out, as Rawls hoped, independently of inequalities in human health and vulnerability. In interpreting Rawls’s difference principle as requiring maximizing the minimum utility level of individuals, economists may thus have missed what is of the greatest potential value in Rawls’s theory of justice: its emphasis on the institutional focus of principles of justice (which it shares with the work of James Buchanan and others concerned with “constitutional political economy”). Rawlsian theory should lead economists to consider how to design institutions that will minimize the need for egalitarian redistribution, and to focus on the best ways to provide the resources that enable individuals who have very different preferences to relate to one another in a society of equals. Rawls’s institutional focus also helps to answer frequent criticisms that at the same time that Rawls can say to the worst off that society cannot be organized to make the worst off any better off, he cannot say the same to other social groups. Critics argue that those who are better off can complain that Rawls’s principles ask them to sacrifice for others, that the product of their efforts is being taken for the benefit of others. These complaints are, however, from Rawls’s perspective, confused. Production is social. It depends on social institutions and policies and on the choices and efforts of many people past and present. Prior to the determination of principles of justice, the legitimate shares of the social product that go to members of various social classes are simply undetermined. There is no nonnormative benchmark or prior specification of rights against which principles of justice benefit or harm social groups. Those who are better off are not as well off as they might be if society were organized in accordance with other

12.3  How Can Rawls’s Principles Guide Policy?

233

principles, but this is the only sense in which they are asked to “sacrifice.” They are not asked to give up anything that is rightfully theirs.

12.3  How Can Rawls’s Principles Guide Policy? As we have seen, Rawls’s contribution to political philosophy holds special interest for economists. The argument from behind the veil of ignorance makes effective use of notions of rational choice, decision making under uncertainty, and maximization subject to constraints –​all theoretical ideas with which economists are comfortable. In addition, Rawls investigates the implications of his theory for the provision of public goods, the role of redistributive taxation, justice in saving and capital accumulation, the role of competition, and macroeconomic policy toward unemployment. Rawls also considers the issue of whether capitalism or socialism is more compatible with justice, concluding that while his theory rules out both complete laissez faire and centralized command state planning, variants of either capitalism or socialism might satisfy the two principles of justice. Whether they do depends on how these systems work out in practice. As one might expect from a theory that focuses on the least well-​off group, the institutional implications of Rawls’s theory are egalitarian. He favors equalizing educational opportunities for persons of similar capacities and motivation, and he supports policies that would block the influence of wealth on political decisions. Rawls’s ideal society would, however, rely only minimally on redistributive taxes and transfers as an equalizing measure. On grounds of efficiency, liberty, and opportunity, Rawls emphasizes the importance of redressing inequalities in the starting points of members of successive generations. Such redress might require investing more in the education of the poor so that talents are cultivated more widely. A final feature of Rawls’s work that should be of interest to economists concerns what he calls the “stability” of principles of justice. Would a society that conformed to his two principles of justice lead people to behave in such a way that those principles would continue to be upheld over time? The answer will depend on how psychological, sociological, political, and economic forces work themselves out over time. Will children raised under just institutions embrace the principles of justice, or will they be free riders seeking to undermine them? Can institutions be sustained that are compatible with the basic liberties and ensure that large inequalities do not undermine the principles of justice over time? These are deep questions about political sociology and moral psychology, made all the more interesting in Rawls’s treatment because he frames them explicitly as relevant to the

234

Justice and Contractualism

assessment of theories of justice. If a set of principles of justice for a society cannot be stable over time, that is a mark against them on Rawls’s view, as on the view of economists who insist that social theories must never ignore incentives. Some major features of Rawls’s work are more closely tied to his contractualist commitments than are others. In particular, one might find the original position framework helpful without being persuaded by the conclusion that Rawls’s specific principles of justice would be chosen; conversely, his two substantive principles may be appealing to those who are not persuaded by the argument from the original position. Does it make sense to appraise a medieval society or a nomadic tribe by the standards of Rawls’s principles of justice? Is it plausible that the same principles of justice apply to representative democracies, feudal manors, and nomadic herders? Although some have read A Theory of Justice as a timeless and explicitly liberal moral theory, a reading that the high abstraction and lofty rhetoric of the book may have invited, Rawls makes clear in his later writings that he sees his work as a historically situated political theory of justice for a pluralist and democratic society. A major goal of Political Liberalism (1993) and Justice as Fairness (2001a) is to help defend his theory as a workable account of justice for a modern, pluralist democracy whose members disagree about many things. Among these are deep and enduring disagreements about religion, about the role of markets, and about the scope of individual liberty. How is it that people with very diverse points of view can agree about the two principles of justice? For such a theory to be successful, it must not be anchored in distinctively liberal views of the person and of morality, like those of Kant or of J. S. Mill. If it were, then the theory could command the adherence only of those who share its liberal vision of the person and of the good life. But, Rawls argues, it is a fact of modern life, which he calls the “fact of reasonable pluralism,” that any society that secures the basic liberties will produce people who hold a range of conflicting moral convictions and religious beliefs. Reason alone is inadequate to produce agreement; only coercion could maintain a society in which everyone’s underlying beliefs were, say, those of a Kantian liberal or of a Roman Catholic. Given the impossibility of uncoerced agreement about all of liberal morality among free citizens, the problem for political philosophy in these circumstances is to seek out agreement on terms of fair cooperation among people who hold differing comprehensive moral, metaphysical, and religious views. Rawls believes that his two principles could be the product of what he calls an “overlapping consensus.” That is, people holding a wide diversity

12.3  How Can Rawls’s Principles Guide Policy?

235

of comprehensive views (provided that such views accept the fact of pluralism) can agree on the two principles of justice. This agreement is not, however, driven simply by the need for some agreement to secure social cooperation. Although different people may have different grounds for accepting the two principles of justice, in each case there is a genuine moral commitment. For example, a conscientious Protestant might endorse a political conception of justice from which follows a commitment to liberty of conscience, because coerced declarations of faith are of no value within the framework of her religion, while a religious skeptic can endorse the same political conception –​including its implication that liberty of conscience should be upheld –​out of a judgment that no right answers are to be had in this realm. They agree on general principles of justice to govern their common lives and accept the more particular principles they entail, but for different reasons. For political purposes they can agree, but they do not agree “all the way down.” Rawls suggests more broadly that what may begin as simply a modus vivendi –​a tactical agreement that toleration is preferable to continued warfare (as in the wake of the Reformation) –​may grow into a more genuine consensus as the value of a free and fair political culture becomes apparent. This casts the argument for the two principles of justice in a different light. Although some might regard the original position in A Theory of Justice as a device to help people to converge on their most basic philosophical convictions, Rawls’s later works make it clear that it serves as a device by which people who disagree (and who expect to continue disagreeing) about what constitutes a good life can arrive at a mutually agreed upon way to live together. Rawls does not regard this agreement as a mere practical compromise but rather as constituting justice. As we have noted, the Rawlsian approach contrasts directly with views like libertarianism and utilitarianism that take the same fundamental principles to govern individual behavior and social institutions and policies. In this regard, utilitarianism, with its focus on the bearing of outcomes on individual welfare, is closer to conventional economic analysis. Although the concern of libertarians with freedom is congenial to many economists, its focus on process rather than outcome is less conventional. A theory like Rawls’s, in contrast, focuses on rules and institutions, and although it assesses them in part by their consequences, it also evaluates them by examining whether they are rationally acceptable. Furthermore, if Rawls can make the case that his two principles can command an overlapping consensus, then theorizing about justice will

236

Justice and Contractualism

not require any prior agreement concerning what constitutes well-​being or concerning what rights people have.

12.4  Other Contract Theories Theorists of justice as mutual advantage reject the features of the Rawlsian contractualist framework that limit the role of self-​interest in contractarian reasoning. The challenge for a contractualist, on this view, is to solve the problem Hobbes had earlier formulated: How can it be rational for a self-​interested actor to conform his behavior to moral rules? Many contemporary thinkers have been inspired by Hobbes: they try to show that moral choice is in fact rational choice under conditions of individual freedom. According to David Gauthier, for example, justice is the result of bargaining among expected utility maximizers with full common knowledge of everyone’s capabilities, endowments, and preferences. No actor in their settings will agree to accept less than she could obtain in the absence of agreement; nor will any expect the others to accept anything less than they would receive if there were no agreement. The minimum others will accept thus fixes a maximum each agent can receive from the bargaining. In Gauthier’s view, rational agents who are unable to make interpersonal comparisons of well-​being will agree to a distribution of benefits and burdens that requires of each the same “relative concession” –​that is, that gives each the same percentage of the maximum increase in well-​being he could get from the bargain. (Because an equal relative concession is not always efficient, Gauthier defends instead the principle that the largest relative concession should be as small as possible.) Those who are advantaged initially will emerge with advantages from the bargain. Gauthier argues that from the perspective that regards justice as promoting mutual advantage, one cannot ignore the differences that agents bring to the table, and therefore one should accept a much less egalitarian theory of justice than Rawls defends. James Buchanan and his colleagues defend a mutual advantage version of contractualism that, like Gauthier’s theory, is less egalitarian than Rawls’s. However, like Rawls’s and Harsanyi’s theories, Buchanan’s sees no role for bargaining, because he believes that the radical uncertainties implicit in deciding on a constitution undermine the differences among individuals that ground bargaining. Buchanan argues that self-​interested agents who are concerned about “constitutional” choices  –​choices involved in setting the general rules of the game –​face so many uncertainties that they might as well be behind a veil of ignorance. Since the gains from constitutional design may be very large, self-​interested agents should be concerned

12.4  Other Contract Theories

237

about the rules of the game and so have good reason to step behind “the veil of uncertainty” (as Brennan and Buchanan 1985 style it). Even though the motivational assumptions in this constitutionalist view are more self-​ interested than in Rawls’s, the element of impartiality introduced by not knowing how alternatives will affect one’s own situation provides a strong common element between the views. In Buchanan’s view, politicians and government bureaucrats should be modeled as self-​interested. Rawls is explicit in rejecting this assumption: “Ideal legislators do not vote their interests” (1971, p. 284). Brennan and Buchanan believe that, recognizing the ubiquity of self-​interest, agents behind the veil of uncertainty will want to limit the scope of government discretion, especially with respect to distribution. They will endorse less egalitarian arrangements than Rawls does because they are concerned that self-​interested political agents will misuse the authority to redistribute in order to promote their own ends. Following Hume and James Madison, Brennan and Buchanan maintain (1985, p. 59), that for the purposes of constitutional design, “every man ought to be supposed a knave.” Although Rawls and Buchanan disagree about the force of self-​interest, they are to some extent also offering different kinds of theories. Rawls defends an “ideal theory.” He is asking about what a just society would look like without corruption, error, and lawbreaking. Questions about how to deal with imperfect compliance with the laws and other perturbations of a just society are undoubtedly important questions for a theory of justice to tackle, but Rawls believes that it is important first to clarify the ultimate objective. However, as Rawls insists, even an ideal society must meet a test that it can stably reproduce a society of people who will willingly obey its rules. Buchanan’s disagreement with Rawls might then be formulated in either of the following ways. First, he might argue that Rawls’s judgments about human nature are too optimistic, and that even in ideal conditions a stable society must ask less of people’s unselfish dispositions than Rawls does. Alternatively, rather than seeing his disagreement with Rawls as resting on different views of human psychology, Buchanan could instead argue that he is stepping away from ideal theory on the grounds that ideal circumstances for fostering human beings who will meet the requirements of justice are not to be expected in any reasonable period. What is needed is nonideal theory to address the noncompliance of individuals who may be far more concerned with feathering their own nests than with promoting cooperation. In either case, Buchanan argues that individuals adopting a constitutional perspective would not want to leave redistribution to the

238

Justice and Contractualism

discretion of inescapably self-​interested politicians. If they favor measures to equalize resources, then they should prefer to embed particular distributive rules directly into the constitution  –​which might, for example, specify a flat tax coupled with a negative income tax (payments to low-​income persons or families that decline with rising income). Such constitutional distributive rules (like the institutions mandated by Rawls’s principles of justice) should not be considered redistributive, since they would figure into the underlying definition of property rights (see also Buchanan 1975). The general idea that moral principles or social norms can be usefully interpreted as the outcome of self-​interested bargains –​that is, the research program suggested by the notion of justice as mutual advantage –​has over the past half-​century received wide attention. This sort of contractualism makes heavy use of the tools of game theory, and its findings may have both explanatory and normative force. In exploring mutual advantage models and in comparing their findings to contractualist work that emphasizes reciprocity or impartiality, it is important to remember their different starting points. What people would agree to if they were motivated as Scanlon assumes is a different matter from what self-​interested people would agree to. Which approach is more fruitful depends on philosophical theses about the nature and point of reasoning about justice, which are not themselves answered from within either of these standpoints.

12.5  Challenges to Contractualism Before closing this chapter on justice, we call attention to two major challenges to contractualist theories of justice and, more specifically, to Rawls’s theory of justice. In The Idea of Justice, Sen argues that ideal theory, even one for which he has great respect, such as Rawls’s theory, ultimately is irrelevant to the central problems of justice in our time. There are two main aspects of Sen’s case. First, he insists that the task for contemporary theorists is “enhancing justice and removing injustice” (2009, p. ix). What fires the passion for justice is the experience of injustice. To develop abstract theories of ideal justice is at best an indirect way of tackling the important problems; at worst it is a form of self indulgence that offers no guidance to tackling injustice. The fact that an ideal society would adopt Rawlsian fair equality of opportunity tells us nothing, according to Sen, about what a policy maker should do now to make public school funding less unfair. Second, owing to the variety of considerations of justice, Sen is skeptical about the possibility

12.5  Challenges to Contractualism

239

of formulating a complete theory that everyone would accept. He asks us to consider what justice says about which of three children should have a toy flute if one child made it, another child can play it, and a third has no other toys (2009, p. 13). Are theorists likely ever to formulate a theory of justice that adjudicates among these separate considerations and that would be agreed to by all? Finally, Sen criticizes the exaggerated ambitions of theorists who feel that philosophy must do something much grander than merely inform specific judgments (2009 pp. 241f.). Not only are theorists unlikely to develop a unified theory of justice, but, Sen maintains, such a theory is often not needed. In particular contexts some considerations of justice will not be relevant and so no weighting of their importance compared to others is needed. From the other flank, as it were, G.  A. Cohen in Rescuing Justice and Equality (2009) criticizes Rawls’s theory on the grounds that it is not ideal enough. As we have seen, Rawls’s framework conceives of a moral division of labor on matters of distributive justice between the state and individual citizens. The just state creates a structure of just rules that leads people who are making otherwise self-​interested decisions about their jobs, pay, and levels of effort to maximize the primary goods of the worst off. Citizens are obliged willingly to obey the rules, but they are not obliged actively to make decisions in their own working life that support the interests of the less advantaged. Cohen argues that this dualism should be rejected: “Rawls calls for compliance in daily life with law that serves principle, rather than compliance in daily life with principle itself ” (p. 375). For example, if marginal tax rates are raised on high earners, it is, in Rawls’s system, permissible for the high earners to avoid the tax by reducing their effort levels, even though sustaining their effort could help the worst off by generating more tax revenue for programs that help the needy. Cohen grants that what justice demands of private citizens must be limited by their legitimate claim that they have their own lives to live, but he insists that nonetheless justice demands that citizens display in appropriate measure an “egalitarian ethos” as they conduct their private lives. The inequalities that the difference principle allows, primarily to serve as incentives for productive labor, are, Cohen argues, a compromise with ideal justice. On his view, any distribution that enhances the primary goods of the worst off through providing incentives to the more talented can always be improved on in a distribution in which the more talented participate willingly while accepting more modest rewards for their efforts. There may be practical reasons for adopting a structure of rules that does not rely on

240

Justice and Contractualism

the talented adopting this egalitarian ethos, but such a concession to the limits to compliance is not appropriate in an ideal theory of justice. The dispute between Rawls and Cohen here is methodological. Although both are concerned with ideal justice, Rawls seeks a theory that integrates and balances a variety of considerations, empirical and ethical, and that can be endorsed through a process of reflective equilibrium. Cohen instead seeks maximum clarity about what distributive justice is. In his view the unfairness inherent in the difference principle shows that, while it may be part of the most desirable or even the most just policy, it cannot be just “through and through” (p. 2).

12.6  Conclusion: Social Contract Reasoning and Economics Problems of justice can be addressed from many different perspectives, and the general moral theories we have already discussed  –​libertarianism and utilitarianism –​imply theories of justice. Libertarianism appeals to economists who share its commitment to the value of negative freedom and its hostility to government interference in economic matters. Utilitarianism appeals to economists who share its focus on individual well-​being. Contractualist approaches may have a harder time getting a hearing in economics, but we have tried to show that there are good reasons why students of economics should take them seriously. In particular, all forms of contractualist reasoning are concerned with rational agreement. (In an earlier formulation, Rawls even went so far as to say that his approach made the theory of justice a branch of the theory of rational choice.) Most contract theorists are also concerned to draw conclusions regarding matters of great interest to normative economics such as the justification of property rights, the legitimacy of redistribution, and the scope of liberty. Finally, forms of contract theory offer well-​worked-​out alternatives and complements to the welfarist theories of normative economics. The reawakening of interest in theories of the social contract has been an important stimulus to economists’ focus on normative theory, and we hope that its influence will grow. This chapter only scratches the surface of an immense literature concerned both with the systematic questions of justice that Rawls, Harsanyi, Gauthier, Buchanan, Sen, and Cohen have been concerned with and with much more particular (though not less important or less challenging) questions concerning what justice demands with respect to the distribution of health care (Brock 2003) or with respect to the apportionment of representatives to states in the U.S. Congress as their populations change (Young

Suggestions for Further Reading

241

1995). On both macro-​and microlevels, justice is a subject to which economic tools and concepts have a great deal to contribute, and it is a crucial corrective to the exclusive focus on welfare that characterizes much of normative economics.

Suggestions for Further Reading Historical texts on the social contract are especially important for appreciating its purposes and significance. Landmark texts include Hobbes (1651), Locke (1690), Rousseau (1762 [1968]), and Kant (1785). Kymlicka (2001), Swift (2013), and Wolff (2006) are useful general introductions to political philosophy with extensive discussions of justice. The literature on Rawls is exceptionally large and rich. The anthology edited by Daniels (1976) and the books by Barry (1973), Wolff (1977), Pogge (1989), and Kukathas and Pettit (1990) are good starting points. Rawls’s economic framework was heavily influenced by James Meade’s classic (1964). Scanlon (1998) develops a “reciprocity” view, including an intriguing account of moral reasons. The many important contemporary works on justice that may be of particular interest to economists include a large part of the egalitarian literature discussed in Chapter 11. General works on justice include Clayton and Williams (2000), Barry (1989, 2005), Tomasi (2013), Gauss (2007), and Brighouse (2004). On global justice, see Pogge (2002), Tan (2004), Brock and Brighouse (2005), and Risse (2012). On local justice Elster (1993) and Peyton Young (1995). On rectificatory justice, see Elster (2004). On challenges to conceptions of justice as the interaction of independent agents, see Sandel (1982), Okin (1991), Kittay (2013), and Nussbaum (2006), As noted in the text, there has been a good deal of contractualist writing in the “mutual benefit” vein. Along with Gauthier and Buchanan, Hampton (1986) and Kavka (1986) have outlined general theories of justice that adopt this Hobbesian starting point. Schotter (1981), Sugden (1986), and Taylor (1987) consider the recurrent problems of coordination and conflict that individuals face and the sorts of institutions and conventions they will come to accept to resolve them. Ken Binmore (1994, 1998, 2005) starts out with a bargaining framework like Gauthier’s, a veil of ignorance like Harsanyi’s (which, Binmore argues, has been built into the conception of fairness by evolutionary processes), and then endorses a view of justice that is very similar to Rawls’s theory, on the grounds that egalitarian views such as Rawls’s are self-​enforcing solutions to bargaining problems.

242

Justice and Contractualism

Questions for Study and Discussion 1. The chapters on utilitarianism, libertarianism, and egalitarianism have already presented several theories of justice. How do these theories of justice compare to the contractualist theories discussed in this chapter? 2. Robert Nozick famously quipped that social contracts are not worth the paper they are not written on. Why should one care about hypothetical social contracts, which would hardly be binding on people today even if they had been signed long ago by others? 3. What is the difference between regarding justice as a matter of mutual advantage and regarding justice as a matter of impartiality? Would theorists of justice as mutual advantage deny that justice should be impartial? Would theorists of justice as impartiality deny that justice should be mutually advantageous? 4. Does not the method of reflective equilibrium lead people merely to systematize their prejudices? Since it holds that none of our ethical starting points, whether they be considered moral judgments about specific cases, moral principles, or metaethical claims about the meaning of ethical claims, is beyond challenge and revision, how can one ever be confident in the truth of anything concerning ethics? 5. What do you think of the thought experiment of the veil of ignorance? How “thick” do you think the veil should be –​that is, what should and should not one be allowed to know in the context of deciding among purported principles of justice? 6. What is the “basic structure” of society, and why does Rawls believe that it raises distinctive questions of justice? Do you agree? Why or why not? 7. What is Rawls’s main criticism of utilitarianism as a theory of justice to govern the basic structure of society? How would a utilitarian reply? Do you find Rawls’s criticism convincing? 8. What is the difference principle? Is it reasonable to focus on the primary goods of the least well-​off group or representative agent? What if everybody else in the society could be made better off at only a small cost to the least-​well-​off group? 9. What is your overall view of Rawls’s two principles of justice? How, in your view, might they be improved?

Questions for Study and Discussion

243

10. What is “ideal theory” and what good is it? Should not a theory of justice, in Rousseau’s words, be concerned with “men as they are,” and thus conforming only imperfectly to any principles of justice? 11. We noted earlier that utilitarianism exists in two varieties. “Total” utilitarianism aims to maximize the amount of utility of all the people in the world. “Average” utilitarianism aims to maximize the utility of the average person. Which of these versions is in your view more compatible with a contractualist outlook? Why? 12. “Thought experiments” like Rawls’s original position play a significant role in both philosophical argument and economic modeling (as, for example, in models that assume perfect information). Can you think of examples (other than those in this chapter) of such hypotheticals that you have found really valuable or really misleading? What do you think makes for a good thought experiment?

PART FOUR

MORAL MATHEMATICS

Morality and mathematics may seem as unrelated as martyrdom and MTV. But it is possible to characterize moral notions formally and to prove theorems. Doing so does not banish controversy or replace moral argument, because abstract mathematical characterizations of moral notions require interpretation and defense. Just as there are disagreements concerning the formal definitions of rationality, so are there controversies about formal definitions of moral concepts and principles. Over the last seventy-​ five years economists, decision theorists, and game theorists have made exciting progress in characterizing features of human interactions and principles to govern them. This work is linked to moral philosophy. Formal models of rationality and game-​theoretic studies of incentives hold out the hope of transcending ancient conundrums concerning the relations among morality, rationality, and self-​ interest. Concepts of envy-​free allocations can facilitate the articulation of egalitarianism. Solution concepts in game theory may enrich the contractarian view that morality can be justified in terms of agreement. Theorems in social choice theory explore the implications of plausible moral axioms and test the consistency of traditional principles concerning how social policies should respond to individual interests. There is an embarrassment of riches in the work on these topics, and we can only comment on a few of these developments.

245

THIRTEEN

Social Choice Theory

Traditional welfare economics offers limited help in evaluating economic institutions and outcomes, because it rejects interpersonal welfare comparisons. Welfare economists often espouse the Pareto principle (that A is better than B if somebody prefers A to B and nobody prefers B to A), but such unanimity in preferences is rarely found. Welfare economists are sometimes tempted to say that A is better than B when A is a potential Pareto improvement over B, but this view is controversial. What else can economists contribute to social evaluation and decision making?

13.1  The Social Welfare Function and Arrow’s Theorem Social choice theory evolved out of an effort to construct better tools for evaluation. Following Bergson (1938) and Samuelson (1947), let us call any ranking of social states that depends on individual welfare a “social welfare function.” Normative principles can be regarded as constraints on acceptable social welfare functions. For example, the principle of the personal good requires that social welfare functions rank A over B if A is better for someone than B, and B is not better for anyone than A. Social welfare functions thus provide a framework for exploring normative principles. Although the framework could in principle be applied to a broad range of evaluative criteria –​including nonindividualistic ones –​ social choice theorists have focused on social welfare functions whose inputs are individual preferences. The Pareto principle, for example, can be characterized as requiring that social welfare functions rank A over B if somebody prefers A to B and nobody prefers B to A. Most economists would regard social welfare functions as by definition welfarist  –​that is, as restricting the inputs to social welfare functions to information about individual welfare. 247

248

Social Choice Theory

In the early days of investigating social welfare functions, it was hoped that thinking in terms of such functions could assist economists in identifying additional plausible normative principles that, like the Pareto principle, relate individual welfare and social welfare. If these principles had precise mathematical formulations, then economists would be able to deduce their implications. In this way, economists could carry out a sort of moral mathematics. However, much of the work on social welfare has not taken this form, owing to (a)  the limitations of the basic Paretian perspective and (b) the results of early formal inquiries into social choice and evaluation. One crucial problem with the welfarist framework is that it is relentlessly consequentialist:  what matters are outcomes. Procedural matters such as fairness or due process count only insofar as they affect outcomes. But people often rank processes as well as the states of affairs to which they lead. A more specific obstacle to the development of the social welfare function framework derives from a remarkable theorem Kenneth Arrow proved in 1951 (see also Arrow 1967). Rather than arguing for strong substantive ethical principles, Arrow proposed a set of apparently weak “procedural” principles that constrain acceptable relationships between individual values and social values. Arrow’s aim was to identify a set of principles that would be consistent with the usual standpoint of normative economists and that would have widespread ethical appeal in their own right. Yet what Arrow discovered was that, in identifying a relatively short list of apparently innocuous constraints to whittle down the set of acceptable social welfare functions, he whittled it down to nothing! Arrow proved that no social welfare function could meet all the conditions for acceptability that he specified. What are the conditions Arrow imposed? First (see Arrow 1967), there is a version of the Pareto principle: P (Weak Pareto principle) If everybody prefers A to B, then A is better than B. By “better than” we mean that the social welfare function ranks A above B. Though it is more controversial than minimal benevolence (because it lacks an “other things being equal” qualification), social choice theorists often endorse it. A second principle, which seems uncontroversial, is the following: D (Nondictatorship) Whether A is better than B should not depend on the preferences of a single individual, regardless of what all others prefer. Speaking of dictatorship suggests issues about how to choose among alternatives rather than about how to evaluate alternatives, and as we shall

13.1  The Social Welfare Function and Arrow’s Theorem

249

see, all of these conditions can be interpreted as constraints on decision-​ making methods rather than as determining the value of alternatives. It seems unobjectionable to stipulate that evaluations of alternatives should not depend exclusively on a single person’s preferences. The third condition is considerably more controversial: IIA (Independence of irrelevant alternatives) Whether A is better than B should depend on how individuals rank A and B and on nothing else. This principle gets its name because it implies that the comparison of two alternatives A and B should not depend on the presence or absence of other “irrelevant” alternatives. The name is misleading, because the principle demands much more than that alternatives other than A and B should not affect their comparative evaluation. IIA says that the social ranking of the pair of alternatives being compared –​the “relevant” alternatives –​ depends on how these alternatives stand in individual preference rankings, and on nothing else. That means that intensities of preferences do not count. Neither do factors independent of individuals’ preferences, such as whether A or B violates rights. Because welfare economists so often focus exclusively on welfare, which they take to be the satisfaction of ordinal preferences, they may find independence of irrelevant alternatives plausible, but it is not an innocuous moral claim. In actual political systems the choice between A and B typically depends on many factors in addition to people’s preferences between them, all of which are ruled inadmissible by IIA. Like many others, we believe that this condition should be rejected. Arrow also includes two formal principles that constrain social welfare functions: U (Universal domain) No matter what the preferences of individuals may be, the social welfare function must always be able to rank alternatives. CR (Collective rationality) The social ranking of alternatives must be complete and transitive. The rationale behind U, the universal domain condition, is that the method of evaluating social alternatives should not collapse when there are unusual arrays of individual preferences. The plausibility of the combination of these two conditions is a complicated matter. U allows the social welfare function to rank alternatives as equally good, but it insists that the social welfare function always offers some ranking. If the social welfare function is interpreted as a method of collective decision making rather than as a means of evaluating alternatives, then the universal domain condition seems obvious: a method of decision making must always have

Social Choice Theory

250

Table 13.1.  Intransitivity of majority voting Regan

Goneril

Cordelia

x y z

y z x

z x y

some output or other. In addition, whether a matter of social evaluation or collective choice, transitivity of social rankings seems to be a compelling requirement. It is, however, less obvious that social evaluation must always be in this way complete. Although it is desirable to have some method of social evaluation that is able to say of any two alternatives that one is better than the other or that the two alternatives are equally good, there is no obvious reason to rule out of bounds social welfare functions that are not complete. Arrow proved that the only complete and transitive social welfare function that satisfies collective rationality, the Pareto principle, and independence of irrelevant alternatives for all profiles of individual preferences is the dictatorial one, in which the evaluation of alternatives matches the preferences of a single individual. Or, to state it more dramatically (and much less precisely), the only rational and universally applicable way to amalgamate individual evaluations of alternatives into social evaluations is for a single individual to serve as the evaluator. However intuitively appealing these five conditions might appear to be, one cannot consistently endorse them all. At least one of them has to be abandoned. Mathematics here leads morality by the hand and gives it a rude shaking. These five conditions cannot be satisfied simultaneously, and so they cannot all be imposed. To illustrate what Arrow’s theorem implies, consider majority voting. It satisfies the weak Pareto condition, the nondictatorship condition, and independence of irrelevant alternatives. But it cannot satisfy both universal domain and collective rationality. Suppose there are three voters –​Regan, Goneril, and Cordelia –​and three alternative social policies x, y, and z. Universal domain requires that the social welfare function be defined for all profiles of individual preferences, including the preferences shown above in Table 13.1. Majority voting ranks x above y, y above z, and z above x and thus fails to satisfy the collective rationality condition. Arrow’s theorem tells us that every method of making social decisions or evaluating alternatives will violate at least one of the five conditions.

13.2  The Interpretation of Arrow’s Theorem

251

13.2  The Interpretation of Arrow’s Theorem The notion of a social welfare function is, as its name suggests, concerned with social welfare and thus with the evaluation of outcomes. But the framework of social choice theory is also useful for the examination of social decision procedures such as majority voting. Sen helpfully distinguishes two broad purposes of aggregation: deciding and evaluating. For example, one might want an aggregation mechanism (1) to decide which movie a group will see or (2) to determine which among a set of movies is best. The criteria for judging choice mechanisms might differ from the criteria for judging evaluation methods. Decision making and evaluating are different tasks. One reason why they differ is that decision rules may have procedural virtues. The best way to decide which movie to see might be by majority vote, since that shows respect for everyone’s wishes, whereas for the purposes of evaluation, it is unclear what is the point of aggregating individual judgments. One might thus want to impose different conditions on decision rules than on methods of evaluation. In the context of decision making, the weak Pareto principle seems compelling:  surely the social choice should conform to the unanimous wishes of the population. On the other hand, it is not obvious that what everybody prefers is better than the alternatives. We already suggested that completeness is more compelling as a condition on decision rules than on evaluation. On the other hand, a ranking of X and Y cannot express the judgment that X is better than Y unless it is transitive. Matters are still more complicated, because, as the movie example suggests, what is being aggregated and not simply the purposes of the aggregation differs among cases. Arrow writes (1967, p.  118):  “Classical utilitarianism specifies that alternative social actions be judged in terms of their consequences for people. In the present terminology I  take this to mean that they are to be judged in terms of the individual preference scales.” But notice that the preferences of an individual among alternative social policies need not have any connection at all to the consequences of those policies for that individual. The preferences of those who support policies that limit global warming need not reflect their judgment that these policies will be better for them than continuing to increase the world’s consumption of fossil fuels. Preferences can sometimes indicate individual welfare, and it is apparently for this reason that Arrow focuses on preferences. But preferences can also reflect an individual’s overall assessment of social policies, regardless of whether those policies make her better off. Sen distinguishes four objects of aggregation  –​interests, preferences, judgments, and views –​and this list could be extended to other items. Any

Social Choice Theory

252

Table 13.2.  Interpretations of social choice theories Purposes Inputs Preferences as indicators of welfare Preferences as judgments of what is socially best

Evaluation Decision making Type 1: Evaluation with Type 2: Objective: to respect to aggregate welfare promote aggregate welfare Type 3: Determination of Type 4: Objective: to an aggregative or average implement what the evaluative judgment population judges best

of these properties of individuals might be aggregated for purposes of collective decision making or evaluation. But the reasons to aggregate individual attitudes or whether aggregation is even a sensible thing to do varies depending on the attitude and the purposes to which the aggregation is to be put. Arrow’s theorem and the conditions discussed in the previous paragraph are typical of the social choice literature in treating the items to be aggregated as preferences, though, as a formal structure, Arrow’s theorem applies to other possible inputs into collective evaluation and decision making. Let us consider the two most common interpretations of the individual attitudes that constitute the inputs into social choice. Reference to “preferences” equivocates between these two interpretations. One possibility is that they are rankings of alternatives with respect to individual well-​being. The second is that they are assessments of policies or outcomes with respect to overall value or with respect to some specific objective, such as security, human rights, or economic growth. One can then draw the 2 × 2 table shown as Table 13.2. This table illustrates four possible interpretations of work relating individual preferences or judgments to social decisions or evaluations. To interpret this table, it is essential to spell out precisely what preferences are and what the social choice theorist takes them to indicate, and to make clear why the theorist seeks to aggregate them. It should by now be clear that a social welfare function of type 1 –​one that attempts to provide a social evaluation based on individual welfare rankings –​makes a claim in moral philosophy or, more precisely, normative political theory. It specifies a relationship between a social evaluation of outcomes, institutions, or policies and the welfare of individuals. A simple illustration of an aggregation mechanism based on a normative principle (i.e., of a type 1 theory) is a

13.2  The Interpretation of Arrow’s Theorem

253

utilitarian social welfare function, which defines the socially best outcome as that which maximizes the sum of individual utilities (where utility is an index of preference). By the way, a utilitarian social welfare function satisfies four of the five conditions that give rise to Arrow’s theorem (the Pareto principle, nondictatorship, universal domain, and collective rationality). It violates independence of irrelevant alternatives, since the total utility of an alternative A as compared to B does not depend exclusively on individual ordinal rankings of A and B. Note that classical utilitarianism does not attempt to aggregate individual normative judgments about what is socially best into a social judgment; that is, utilitarianism is not intended as a type 3 theory. But, as comparative social state evaluations, preferences often imply judgments, and in aggregating preferences, one may be in fact aggregating people’s judgments about what would be better for themselves, about what would be better for others and for the society or the world as a whole, or about both of these taken together. It is far from clear why one should think that the social evaluation of outcomes, institutions, or policies should conform to Arrow’s conditions. Preferences among social policies are often not reliable indicators of individual welfare. In addition to the fact that individuals often prefer what they think is right or what they think is better for someone else, regardless of its welfare benefit to them, they may lack the information, time, and ability to make any judgment about which of two alternatives is better, and owing to false beliefs and biases their evaluations are often off the mark. Indeed, with respect to many decisions, such as whether a certain drug should be available without prescription or whether the Federal Reserve should raise interest rates, a large majority may prefer that their preferences not govern the evaluation of alternatives. If one believes, as many economists do, that preferences are reliable indicators of individual well-​being and that the value of social policies depends exclusively on the extent to which they enhance individual well-​being, then the social value of policies should depend on individual preferences. But we have argued that these beliefs are false. Preferences are often not reliable indicators of well-​being, and the evaluation of social policies depends on many factors apart from their consequences for individual well-​being. So it is questionable how much a social welfare function relating individual preferences to social evaluations has to contribute to normative social and political philosophy. The type 2 and type 4 interpretations that see social choices as governed by individual preferences might appear to be on stronger ground. These views

254

Social Choice Theory

of the relationship between social choice and individual preference might appear to follow from the notion of democratic sovereignty, the idea that in a democracy the views of the population should govern. But democratic sovereignty demands only that the methods of social decision making conform to the preferences of the population. It is no violation of democratic sovereignty if the population chooses to delegate social decision making to a legislative body and to a bureaucracy that interprets and applies the legislation. On some matters, sensible people do not want their preferences to govern political decisions, because they recognize that they do not have the relevant expertise. The upshot of our discussion is that to be relevant to political philosophy, social welfare functions or social choice theory must either limit its domain of applicability severely, or it must revise its concept of welfare, or it should allow other nonwelfarist moral considerations to play a role.

13.3  Social Choice Theory and Moral Philosophy Arrow’s theorem has shaped a large portion of the field of social choice theory, and it has had a significant impact on contemporary moral and political philosophy. Yet social choice theory is not exhausted by extensions and reactions to Arrow’s theorem. Work we have already discussed –​such as the theorems of welfare economics or the formal work on egalitarianism –​ count as a part of social choice theory, as does a good deal of cooperative game theory (upon which we shall comment in Chapter 14). Much of this work aims constructively to clarify principles of morality and rationality and to determine their implications. However, the interpretation of theorems in social choice theory and even its very definition remain elusive and controversial. Sen suggests that social choice theory can be defined either by its approaches or by its subject matter, which he describes as follows: “Social choice problems arise in aggregating the interests, or preferences, or judgments, or views, of different persons (or groups) in a particular society” (Sen 1986, p. 214). Since aggregating interests and resolving conflicts are central to moral and political philosophy, their subject matter often overlaps with that of social choice theory. Social choice theorems typically raise moral questions. The most obvious concern the axioms: Do they express plausible normative principles? But one can also ask questions about subtler features of a formal model. For example, treating the relationship between profiles of individual preferences and a specific social choice as literally a function implies the uniqueness of social choice, which may be doubted. Furthermore, one can

13.3  Social Choice Theory and Moral Philosophy

255

question whether a particular aggregation mechanism as a whole is morally or politically justifiable. For instance, someone may object to a process for aggregating welfare rankings for purposes of decision making (a type 2 theory) as unfair, perhaps because it does not adequately protect the interests of minorities. Or she may object to a process for aggregating individual normative judgments into a collective judgment (a type 3 theory) as unwise because it fails to respond to the quality of the arguments supporting the judgments. Finally, she may ask (as we have) moral questions about the content of the preferences and value judgments upon which the collective decisions or evaluations depend. If the axioms and formal structure pass preliminary scrutiny, then theorists can argue from the results of social choice theory to moral conclusions. If one accepts the axioms of some valid proof, then one must accept the conclusions. Axiomatic arguments of this kind have been proposed for ethical conclusions as varied as utilitarianism, a utility analogue of Rawls’s difference principle (d’Aspremont and Gevers 1977), and welfare egalitarianism (Roemer 1988). For example, in addition to making a veil of ignorance argument for utilitarianism, John Harsanyi (1955) demonstrates that if (1) personal and moral preferences satisfy the axioms of expected utility theory and (2) the Pareto principle is satisfied, then the social evaluation of alternatives is a linear function of individual utilities. Harsanyi takes utility to be an index not of actual but rather of corrected and cleansed preferences, and he interprets this theorem as an argument for this version of utilitarianism. Others, such as Sen (1976) and Weymark (1991), have questioned this interpretation of the result on the grounds that Harsanyi has not shown why the expected-​utility representation of individual and ethical preferences is the right one to use for ethical comparisons. Harsanyi would counter this criticism by maintaining that only the expected-​utility representation captures attitudes toward risk and, implicitly, information about preference intensity (1977a; see also Mongin and d’Aspremont 1998, pp. 434–​435). Logically valid arguments show only that one cannot both accept all their premises and reject any of their conclusions. If one is convinced of the correctness of their premises, valid arguments can persuade one to accept their conclusions, but if one is convinced of the incorrectness of the conclusions, then it is equally true that valid arguments can persuade one to doubt the premises. In general, people’s confidence in the axioms of social choice theorems will derive both from how reasonable they seem to be in their own right and from how reasonable are the conclusions they entail. If the axioms entail a disturbing moral conclusion, then people will be motivated

256

Social Choice Theory

to scrutinize them more closely. This perspective is similar to John Rawls’s “reflective equilibrium” as a method for moral deliberation generally; it can be reasonable for a person to reject “axioms” that fail to accord with her most deeply held convictions. We suspect that the most important role of theorems in social choice theory has been to reveal that apparently plausible moral principles are more ambiguous and problematic than they appear. Such ambiguities and difficulties have been highlighted most conclusively by demonstrations that sets of plausible principles are logically inconsistent. Arrow’s theorem is the most famous of these demonstrations. But it is not easy to say exactly what moral lessons should be drawn, because the theorem can be read in many different ways depending on how the individual rankings are interpreted and on whether the point is to evaluate or to choose. As Sen has observed, Arrow’s impossibility result may not be as distressing as it appears because individual ordinal rankings provide little information concerning the relative importance of different alternatives. If either interpersonal comparisons of preference intensity or additional information (for example, concerning comparative income or resource levels) is permitted (violating IIA), then no analogous impossibility result will hold (Sen 1970a, chs. 7, 8).

13.4  Social Welfare Functions with Interpersonal Comparisons The conception of a social welfare function can be adapted to explorations of very different principles for evaluating social outcomes on the basis of richer information concerning individual welfare. To provide a better sense of the power and sophistication of this work and to provide further illustration of the abstract claims in the previous section, we shall sketch an argument Matthew Adler gives in his recent (2012) Well-Being and Fair Distribution for a social welfare function that gives most weight to the worst off. Although Adler offers a sophisticated preference-​based view of welfare, including an intricate view of interpersonal utility comparisons that develops the “extended sympathy” view we criticized in Chapter  7, he intends his account to be neutral among plausible accounts of welfare. So rather than taking individual preference rankings to be the inputs into the social welfare function, Adler starts with interpersonally comparable cardinal utilities. (In fact, matters are more complicated, since he allows for incompleteness in individual comparisons of outcomes, but we shall simplify and suppose that for each outcome and each individual there is a single interpersonally comparable number that represents each person’s welfare.) Like Arrow, Adler requires that his social welfare function

13.4  Social Welfare Functions with Interpersonal Comparisons

257

satisfy a universal domain condition, U, but the domain consists of individual utilities rather than profiles of preferences. In addition, the rankings implied by the social welfare function must satisfy a collective rationality condition. He then imposes a stronger version of the Pareto principle: P* (Strong Pareto) If somebody is better off with outcome x than with outcome y, and nobody is better off with outcome y, then x is a better outcome than y. By “x is a better outcome than y” we mean that the social welfare function ranks x above y. In addition, Adler imposes an anonymity condition: A (Anonymity) If the pattern of well-​being in one outcome is simply a permutation across individuals of the pattern of well-​being in another outcome, then the two outcomes are equally good. By “equally good,” we mean that the social welfare function ranks them equally. Adler regards the anonymity axiom as one way to capture the impartiality that morality demands. What matters are levels of well-​being, not which individuals have those levels of well-​being. Unlike Arrow or Harsanyi, Adler also imposes an egalitarian condition on his social welfare function: PD (Pigou–​Dalton) If (1) P is better off than Q in outcome x, and (2) P is no worse off than Q in outcome y, and (3) Q’s gain in well-​being in y as compared to x equals P’s loss, and (4) everybody else is equally well off in x as they are in y, then y is better than x. Stated less formally, the Pigou–​Dalton principle says that transfers that improve the well-​being of those who are worse off and that diminish the well-​being of the better off by the same amount make the social outcome better. This is clearly a principle that would be attractive to an egalitarian, and it makes no sense unless the measure of welfare is interpersonally comparable. Finally, Adler adds a separability condition, which we restate somewhat roughly as follows: S (Separability) If P’s welfare is w in both outcome x and outcome y, and x´ and y´ are identical to x and y, except that P’s welfare is w´ in x´ and in y´, then x is better than y if and only if x´ is better than y´.

258

Social Choice Theory

Separability means that in comparing outcomes one need not pay attention to the level of well-​being of those who are equally well off in the two outcomes. Only the differences in the welfare of individuals between the two outcomes are relevant to their social ranking. Adler then shows that these axioms imply that the social welfare function must be a “weighted” utilitarianism with larger weights attached to the welfare of those who are worse off. His theorem leaves open, however, how strongly these weights should diminish with the greater well-​being of an individual. Adler also imposes a continuity axiom in order to prove that the weighting is also continuous, but we will not discuss these details. Universal domain, collective rationality, separability, and anonymity all seem plausible. Those who would reject Adler’s conclusion must thus (1)  deny it is possible to have some sort of interpersonally comparable numerical representation of people’s welfare; (2) deny that individual welfare should govern social evaluation, which is what the Pareto principle says; or (3)  reject the mild egalitarianism in the Pigou–​Dalton principle. Axiomatic investigations such as Adler’s can help in this way to clarify the bases of disagreements about moral and political principles.

13.5  Rights, Freedom, and Fairness in Social Choice Theory With the exception of the nondictatorship condition, Arrow’s axioms seem to have little to do with fairness. Harsanyi’s axiom requiring that everyone’s utility be weighted equally can certainly be interpreted as a form of fairness, but it does not imply that the utilitarian have any direct concern with the way that welfare is distributed among people. The Pigou–​Dalton condition, in contrast, seems to embody a minimal egalitarian idea of fairness, and Adler argues that his weighted social welfare function captures all defensible concerns about the fairness of outcomes. Whether he is correct or not, his argument provides a further illustration of the value of efforts to formulate moral principles mathematically and investigate their implications. One feature of Adler’s argument worth noting is his response to a classic objection both to utilitarianism and to his own prioritarian social welfare function. Consider the following example. Peter Diamond (1967) has maintained, plausibly, that unequal outcomes are fairer if everyone has an equal chance of doing well or poorly than if the chances are also unequal. Since, by assumption, the outcome is the same, whether the chances that individuals will do well are the same or not, both utilitarianism and Adler’s own social welfare function would be indifferent to the distribution of chances. So it seems as if Adler’s social welfare function fails to capture one important kind of fairness, which concerns the distribution of expectations.

13.5  Rights, Freedom, and Fairness in Social Choice Theory

259

Table 13.3.  Fair expectations vs. fair outcomes Policy 1 s1 s2 Expectation

A 50 60 55

Policy 2 B 50 40 45

A 70 30 50

B 30 70 50

Should not Adler and others apply the Pigou–​Dalton principle to expectations as well as to outcomes? Has Adler successfully captured the distributional concerns of egalitarians in his formal theory? In response, Adler argues that appraising policies in terms of their expectations, however appealing it may initially be, violates his substantive axioms. For example (Adler 2012, p. 512), consider a society of just two people, A and B, deciding between two policies, which will have different results depending on whether the future state of the world is s1 or s2, which are equally likely. Suppose that the results of the policies are as shown in Table 13.3. The Pigou–​Dalton principle says that both of the possible outcomes of policy 1 (a 50/​50 or a 60/​40 division) are better than the 70/​30 or 30/​70 splits resulting from policy 2. So it favors policy 1. But the distribution of expected utility (55/​45) in policy 1 is more unequal than the distribution of expected utility in policy 2 (50/​50) and so, applying the Pigou–​Dalton principle to expected utilities implies that policy 2 is better than policy 1. The example shows that one has to choose. Either one should be concerned about inequalities in outcomes or one should be concerned with inequalities in expectations. Favoring equalities in both leads to contradiction. It is in findings such as these that formal work with social welfare functions shows its value. What about important considerations such as rights and freedom? Do they have formal representations within social choice theory? Decades ago Amartya Sen (1970b) argued that it is possible to represent some rights by imposing constraints on the social welfare function. If an individual has a right to determine whether some aspect of an outcome obtains (such as whether that individual brushes his teeth), then his own preference with respect to brushing or not brushing his teeth should determine the social evaluation of outcomes that differ in just this single regard. In other words, to have a right to do X is for one’s preferences to be socially decisive between social states that differ only with respect to whether one does X. Unfortunately, Sen showed that granting two people rights characterized in this way, each over just two alternatives, winds up being inconsistent with the weak Pareto principle or collective rationality. This is Sen’s “paradox of the

Social Choice Theory

260

Table 13.4.  Lewd’s and Prude’s preferences Lewd

Prude

(b) both read the book (p) only Prude reads it (l) only Lewd reads it (n) neither reads it

(n) neither reads the book (p) only Prude reads it (l) only Lewd read it (b) both read the book

Paretian liberal.” An example that Sen uses to illustrate the conflict involves two individuals, “Lewd” and “Prude,” and a copy of D. H. Lawrence’s novel Lady Chatterley’s Lover. Lewd most prefers that both he and Prude read Lady Chatterley’s Lover. If only one should read it, however, Lewd wishes it to be his prudish friend, since it would loosen him up. Prude would most prefer the book not be read at all, but would rather read it himself than further his friend’s corruption. Their preferences are shown in Table 13.4. Given minimal liberalism, Lewd should be decisive between either (b, p) or (l, n) and Prude should be decisive between either (n, p) or (l, b). Suppose Lewd is decisive over (b, p) and Prude is decisive over (l, b). Since Lewd is decisive between b and p and prefers b to p, it follows that b should be socially preferred to p. And since Prude is decisive between l and b and prefers l to b, l should be socially preferred to b. So if social choice respects minimal liberalism, then the outcome will be l, that only Lewd reads the book. But both prefer p –​that Prude read the book –​to l, and the Pareto principle consequently implies that p should be socially preferred to l. If one respects both minimal liberalism and the Pareto principle, then there is a cycle and collective rationality is violated. Allan Gibbard points out that on a slightly less restrictive formulation of the view of rights as socially decisive, certain combinations of preferences will rule out the possibility of any social choice, regardless of questions of Pareto optimality. In Gibbard’s example of the nonconformist and Mrs. Grundy (1974, p. 389), the nonconformist wants his bedroom walls to be a different color from Mrs. Grundy’s, while she wants her bedroom walls to be the same color as the nonconformist’s. If each is decisive over social alternatives that differ only with regard to the color of his or her bedroom walls, then every possible outcome will be vetoed by somebody. Gibbard (1974), Nozick (1974), Gärdenfors (1981), Sugden (1985), and Gaertner, Pattanaik, and Suzumura (1992) suggest that Sen’s paradox and Gibbard’s result arise from a mistaken characterization of rights. In their view, rights should instead be modeled as sets of strategies in suitable

13.5  Rights, Freedom, and Fairness in Social Choice Theory

261

game forms. What this means, roughly, is that whether an individual has a right to do something depends on whether the choice is up to that individual. To represent formally whether a choice is up to an individual requires a different framework than social choice theory, one that allows some way of representing the alternative actions that individuals may take. Game theory, which we will discuss in Chapter 14, provides such a framework.1 Given the difficulties concerning incorporating rights within social choice theory, it is unsurprising that there are problems in characterizing freedom as well. The obstacles that, on MacCallum’s view, help to define different notions of freedom are mainly procedural and hard to characterize within social choice theory. One way that the tools of social choice theory (stripped of their welfarist and consequentialist interpretation) have contributed to the understanding of freedom has been with respect to the question of the quantity or extent of what we called in Chapter 10 “effective freedom” or opportunity. Here is an example: Let X be the set of all possible alternatives among which people might choose. Let A, B, … be the subsets of X that represent the sets of alternatives among which particular individuals choose. Some normative economists have been interested in characterizing when A ≥ B, A > B, or A ≈ B, where “A ≥ B” is to be read “A provides at least as much freedom or opportunity as B,” “A > B” means “A provides more freedom than B,” and “A ≈ B” symbolizes “A and B provide the same amount of freedom.” These economists have attempted to formulate some plausible conditions comparing the freedom provided by different “opportunity sets” and to investigate their implications. For example, the following three axioms seem plausible: 1. Opportunity sets that permit no choice at all –​that contain only one member –​all provide the same (zero) amount of freedom. 2. If A contains all the options in B and some others in addition, then A > B. 3. If some option X is not in sets A or B and if A* and B* are the sets that result from adding X to A and B (respectively), then A ≥ B if and only if A* ≥ B*. 1 Suppose in Gibbard’s case that the nonconformist and Mrs. Grundy can each paint their own bedroom wall either blue or gray. One might then say that their rights are respected if both strategies are open to them. The rights that strategies imply only enable each individual to determine the colors of his or her bedroom. Neither can unilaterally determine whether the colors are the same or different.

262

Social Choice Theory

These three conditions imply that the opportunity or effective freedom of a set can be measured by the number of alternatives it contains.2 As Sen (1990b) has argued, this implication is unacceptable. It means, for example, that a worker who has a choice of jobs at the counter of any one of fifty different McDonald’s restaurants has more freedom than if she has a choice between ten entirely different jobs. At least one of the three plausible premises that imply this conclusion must be mistaken. We think the third axiom should be rejected, on the grounds that how much adding a new alternative X to some set A increases freedom depends on how similar it is to the alternatives already in set A. Sen also argues that the quality of the alternatives bears on the extent of the opportunities a set provides. “We find it absurd to dissociate the extent of our freedom from our preferences over the alternatives” (1990b, p. 470). One simple way of taking this into account (which Sen does not espouse) is to measure opportunity by the utility of the best alternative that a set makes possible. But proceeding this way leads the concern with freedom to collapse into a concern with welfare. Freedom is not a matter of how well off one is. A military conscript who is assigned a great job may be much better off than a civilian who has a choice among a variety of worse jobs, but the conscript does not have more freedom. Others concerned with the formal representation of conceptions of freedom have focused on the diversity of the alternatives available to an individual. The concern with diversity links up to a concern with the quality of alternatives insofar as diversity depends in some way on preferences. One might, for example, consider not just an individual’s actual preferences but also a range of different preference rankings that an individual might have or might reasonably have had. Adding anchovies to the alternative pizza toppings increases the range of an agent’s choices, even if the diner does not happen to like anchovies, unlike adding mouse droppings. Arrow (1995) thus links our interest in freedom to our interest in flexibility. One might argue that adding an alternative to an opportunity set increases freedom or opportunity if and only if the alternative is best in at least one admissible 2 Condition (i) tells us that {x} ≈ {y} and {z} ≈ {w} for any four distinct alternatives. If we add z to the set {x} and to the set {y}, then (iii) implies that {x, z} ≈ {y, z}. If we add x to the sets {z} and {w}, then (iii) implies that {x, z} ≈ {x, w}. So {y, z} ≈ {x, w} –​that is, all two-​member opportunity sets provide the same amount of freedom (which is, by (ii), more than the freedom in one-​member opportunity sets). In the same way we can show that, if all n-​ member opportunity sets provide the same amount of freedom, then all (n + 1)-​member opportunity sets provide the same amount of freedom, and by mathematical induction it follows that every two sets with the same number of alternatives provide the same amount of freedom. See Pattanaik and Xu (1990).

13.6  Other Developments of Social Choice Theory

263

Table 13.5.  The discursive dilemma

Individual 1 Individual 2 Individual 3 Society

P

If P then Q

Q

True False True True

True True False True

True False False False

preference ranking. However, this sort of view can easily lead to the conclusion that if every admissible preference ranking ranks some alternative X in the opportunity set A higher than every alternative in B, then A provides more freedom than B (Sugden 1998). But A may contain very few alternatives while B contains a large number of highly diverse alternatives. These examples provide only a small introduction to a significant literature. The axioms we have presented and discussed are far from adequate. Nevertheless, the efforts of social choice theorists to characterize the extent of freedom or of opportunity highlight serious lacunae in the philosophical literature. For example, without some coherent concept of greater or lesser opportunities, it is confusing to talk about equalizing opportunity.

13.6  Other Developments of Social Choice Theory We shall end this chapter by sketching two other formal investigations of moral and political principles that jettison the welfarism that characterizes the standard work in social choice theory such as Arrow’s theorem. The first of these examples, the so-​called discursive dilemma, is concerned with aggregating the judgments of individuals. Suppose that a society consists of three individuals who make the judgments shown in Table 13.5 concerning the truth or falsity of the propositions P and Q, and suppose that social judgment follows the majority choice. The judgments of each of the individuals in this case are consistent with the principles of logic, while social judgments violate them (List and Pettit 2002). Unlike the material discussed earlier, no ethical principles are explicitly invoked here; the issue concerns rationality. As the example shows, majority decision making can lead to logical inconsistency. How important is it that social judgments be consistent with the principles of logic? What can be done about the disturbing possibility of inconsistency? Should the social judgment concerning Q depend on what most people think about Q, or should it depend on what they think about P and if P then Q?

264

Social Choice Theory

The second example (Roemer 1988, pp. 160–​172) is concerned with the social evaluation of states of affairs that involve different distributions of welfare. However, unlike traditional work on social welfare functions, the axioms are not entirely welfarist. Roemer invites us to consider the distributional principles that should govern a hypothetical economic environment in which there is a single output, “corn,” which is produced and consumed by two individuals, Able and Infirm, who have the same preferences for bundles of corn and labor but different skills: Able can produce much more corn per hour of labor than can Infirm. If egalitarians want to compensate Infirm yet remain fair to Able, what overall egalitarian principle should they follow? Just as social choice theorists –​in their attempts to define a social welfare function –​seek conditions relating individual preference and social welfare, so Roemer proposes moral principles governing the allocation of corn to Able and Infirm. His conditions are as follows: 1. (Pareto optimality) The corn allocation should be Pareto optimal. 2. (Land monotonicity) If the amount of land increases, neither individual should be made worse off. 3. (Technological monotonicity) If it becomes possible to produce more corn with the same inputs of land and labor, neither individual should be made worse off. 4. (Limited self-​ownership) Able should not be worse off than Infirm. 5. (Protection of Infirm) Infirm should be at least as well off as he would be if Able had no more skills than Infirm has. Although these axioms all concern welfare, axioms 2, 3, and 5 make claims about how welfare is related to other factors such as land, technology, and individual skills. In this way Roemer expands or breaches the framework of social welfare functions. These axioms seem attractive: greater productivity should not harm either individual; Able should not do worse than Infirm; Infirm should not lose in absolute terms from Able’s greater skills; the allocation of corn should be Pareto optimal. Yet Roemer proves that the only method of allocating corn that satisfies these five rules is to equalize Able’s and Infirm’s utility. The weak egalitarianism that makes rules such as protection of Infirm or the monotonicity conditions attractive, when combined with limited self-​ownership and the Pareto principle, implies an incredibly strong egalitarian conclusion. One reaction would be to go back to Chapter 11 and rethink the arguments against equalizing welfare. Perhaps the criticisms described there were premature and equalizing welfare is, after all, an appropriate ideal. Alternatively, one can reconsider Roemer’s five conditions and their role in the proof of his theorem.

13.7 Conclusions

265

The point of discussing these examples is not to explore the logical consistency of social judgment or to reopen the discussion of egalitarianism. Instead, our object has been to show how the techniques of social choice theory can be employed with different inputs and for different purposes. The techniques of social choice theory yield surprising results concerning a wide range of issues.

13.7 Conclusions What do examples such as Arrow’s impossibility theorem, Harsanyi’s proof of utilitarianism, Adler’s derivation of a principle giving priority to the worst off, or Roemer’s demonstration that reasonable axioms demand the strong egalitarian conclusion that welfare should be equal show? What value is there in attempting to characterize rights formally, in laying out the discursive dilemma, or in exploring the implications of axioms concerning freedom? How fruitful is social choice theory in illuminating moral issues? Its typical subject matter –​especially the problem of assembling individual judgments or interests into social evaluations –​is of undeniable moral importance. But we think there are reasons to be ambivalent about the value of its axiomatic approach. To be clear, we are not ambivalent about the value of precision and rigor. There is no doubt that the formal approach can lead to important clarifications and can stimulate a new awareness of the depth of problems –​as the literature on the representation of rights or the measure of freedom shows. At the same time, the abstractness of these methods has pitfalls. To quote Sen again:  “Arrow was undoubtedly right in saying that ‘one of the great advantages of abstract postulational methods is the fact that the same system may be given several different interpretations, permitting a considerable saving of time’ (Arrow 1951, p. 87). But that probably is also one of the great disadvantages of these methods” (Sen 1982a, p.  190). Considerable confusion has been generated, particularly in discussions of Arrow’s theorem, by a failure to distinguish among alternative interpretations of the objects being aggregated and the purposes of aggregation. It is difficult to spend much time in the social choice literature without becoming aware of the disparity between the attention lavished on the derivation of theorems and the comparatively little attention typically devoted to their interpretation. Unless guided by ethical sophistication, social choice theory can degenerate into formal exercises that illuminate little of importance. It would be unfair to single out economists for blame: moral philosophers ought to be more involved with this work. If formalizations of

266

Social Choice Theory

plausible moral principles lead to inconsistencies or to unacceptable conclusions, then philosophers and economists need to understand what has gone wrong. It may be that the formalizations fail to express the informal principles they attempt to restate, or it may be that there are ambiguities within those principles themselves. Only careful interpretation of both the informal moral principles and their formal restatements enables one to appreciate the significance of a particular theorem. Social choice theorists have left lots of philosophical work to be done.

Suggestions for Further Reading One of the best general introductions to social choice theory is still Sen (1970a). For a more recent introductions and surveys see Arrow and Sen (2010); Gaertner (2009); Arrow, Sen, and Suzumura (1996, 1997, 2002); Barbera, Hammond, and Seidl (1999); and Sen (1999). Arrow’s theorem was first presented in Arrow (1951), a classic. For a valuable book-​length philosophical discussion see MacKay (1980). A  range of critical and constructive essays on social choice is presented in Elster and Hylland (1986). Sen first presented his paradox of the Paretian liberal in 1970. For some of the major discussions, see Sen (1976, 1983, 1986, 1992b); Gibbard (1974); Nozick (1974); Gärdenfors (1981); Riley (1989, 1990); and Gaertner et al. (1992). For further developments of a game-​form construal of rights, see Fleurbaey and van Hees (2000). Vickrey (1945), Fleming (1952), and Harsanyi (1955) present axiomatic arguments for utilitarianism. Reinterpreted and spruced-​up versions of something like Harsanyi’s argument can be found in Broome (1991b), Mongin and D’Aspremont (1998), and Mongin (2001). Harsanyi’s interpretation of his theorem as an argument for utilitarianism is challenged in Sen (1976) and Weymark (1991). Adler’s complex argument for prioritarianism (2012) is related to these arguments for utilitarianism, although in relying as it does on the Pigou–Dalton principle, it acquires an egalitarian twist. For axiomatic arguments for a utility analogue to Rawls’s difference principle, see Hammond (1976), Strasnick (1976), and d’Aspremont and Gevers (1977). Issues concerning the logical consistency of social judgments are explored in List and Pettit (2002, 2004). There is a large literature on formal representations of the extent of freedom and opportunity. See particularly Jones and Sugden (1982), Pattanaik and Xu (1990), Sen (1990b), Arrow (1995), Laslier et  al. (1998), Carter

Questions for Study and Discussion

267

(1999), and especially Sugden (1998). For further developments, see Nehring and Puppe (2002) and Bossert, Pattanaik, and Xu (2003). For some formal investigations of egalitarianism and its consequences, see Fleurbaey (1995, 2002, 2008), Kolm (1997), Roemer (2002), Fleurbaey and Maniquet (2005, 2011, 2012), and Fleurbaey et al. (2005).

Questions for Study and Discussion 1. Explain Arrow’s theorem. How important do you think that the result is? Which of the axioms that give rise to the impossibility result would you reject? 2. Why might economists and policy makers be interested in aggregating the preferences of members of the society? Should the views of citizens (perhaps via referenda) determine society’s policies? What answers to these questions do Arrow’s axioms imply or presuppose? What do you think the right answer is? Defend your answer. 3. As the example at the end of Section 13.1 shows, it is possible for majority voting to lead to an intransitive ranking of alternatives. In the case of such intransitivities, how should social decisions be made? 4. What do you think of Adler’s formal argument in defense of a prioritarian consequentialism? Are you prepared to accept his conclusion, or would you instead reject one of his assumptions? Why? 5. Diamond’s point (1967) can be illustrated as follows. Suppose that there is one indivisible unit of some good that can go to either Jack or Jill. Flipping a coin to determine who wins strikes most people as fairer than arbitrarily choosing to give the good to Jack or to give the good to Jill. Adler denies this and gives the argument illustrated in Table 13.3. Do you think that Adler is right? What are your reasons? 6. What is the difference between characterizing rights in Sen’s way (in terms of having a decisive say in social choices) and the alternative characterization in terms of what an individual is free to do without interference from others? 7. Why do you think it is difficult to give a formal characterization of rights within social choice theory? What limitations do these difficulties suggest concerning the usefulness of social choice theory? 8. What do you think of the principle that adding some new alternative X to the set of things among which an individual is free to choose increases the individual’s freedom if and only if there is some preference ranking that the individual might reasonably come to have according to which X would rank above all the other alternatives?

268

Social Choice Theory

9. A three-​judge panel is trying to decide whether Herbert is guilty of libeling Henry. To be guilty of libel, Herbert’s claims must be false and intended to harm Henry. One judge believes that both of these conditions are true of Herbert and thus believes he is guilty. The second judge believes that Herbert made a false claim but did not intend any harm, while the third judge believed that Herbert intended harm but that his claims were true. If the judges vote on Herbert’s guilt, the majority will find him innocent. If instead they take separate votes on whether his claims were false and whether he intended harm, a majority will find that his claims were false and that he intended harm, implying that he is guilty. Which procedure should they employ? Should Herbert be found innocent or guilty? Why? 10. If you do not believe that there is a moral obligation to equalize welfare, then you have to reject one of Roemer’s five axioms in Section 13.6, which together imply that policy should aim to equalize welfare. Which one would you reject and why? 11. Many of the examples in this chapter discuss the problem of reconciling seemingly innocuous axioms with conclusions derived from those axioms that seem surprising and counterintuitive. Physicists faced a problem like this in the early twentieth century, when it became clear that the observed behavior of the speed of light was incompatible with the axioms that constituted Newton’s laws of motion. Drawing on what you know about physics or other physical sciences, how would you compare scientists’ response to anomalous results that follow from highly plausible axioms with that of “moral mathematicians”?

FOURTEEN

Game Theory

Game theory was conceived as a tool to explain, predict, and guide the behavior of rational individuals in strategic interactions, and its main applications still lie in this domain. But game theory is not wedded to any particular domain, and it also applies to such disparate phenomena as computer programming and biological evolution. Game theory is as much a branch of mathematics as a theory of interactive decision making. However, in this chapter we focus exclusively on applications of game theory to strategic interactions among individuals. An interaction is “strategic” if the payoffs to some individuals depend both on their own choices and on the choices of others.

14.1  What Is a Game? A game is defined by its players, the strategies the players can choose, the payoffs to the players from the combinations of different strategies, and the capacities of the players, including the knowledge each player has concerning features of the game (such as the other players’ available strategies). Each of the italicized terms has a technical meaning. In the case of the games we consider here, players are individuals. To keep matters simple, we shall for the most part discuss games with only two players. A  player’s strategy is a complete specification of the actions the player can take whenever the player might be able to act. The set of strategies available to a player defines all the possible actions that the player could take. Before explaining the other italicized terms, it will help to provide an example. One way to represent strategic interactions is via an “extensive form” or a game “tree,” like the one shown in Figure 14.1. This interaction involves two players, player I (Jill) and player II (Jack). Jill moves first at 269

Game Theory

270

I a Up

II Left

($40, $40) (3,3)

Down

b

c Right

($0, $50) (1,4)

Left

($50, $0) (4,1)

II Right

($1, $1) (2,2)

Figure 14.1.  An extensive form game.

the node labeled a and plays Up or Down; these two moves are her two strategies. It is assumed the extensive form of the game is common knowledge. (A fact is common knowledge if all know it, all know that all others know it, all others know that all others know that all others know it, and so on.) So when player II, Jack, is choosing whether to play Left or Right, he knows whether he is choosing at node b or instead at node c. Although Jack can choose between just two actions at each node, he has four (not just two) possible strategies. His four strategies are (i) Left regardless (no matter what Jill does); (ii) Left if Jill plays Up, Right if she plays Down; (iii) Right if she plays Up and Left if she plays Down; and (iv) Right regardless. What we are calling “strategies” here are called by game theorists “pure strategies” as opposed to “mixed strategies.” A mixed strategy involves playing different pure strategies with different probabilities. If Jack resolved to flip a coin and to play Left or Right depending on whether the coin landed heads or tails, regardless of what Jill were to do, then he would be playing the mixed strategy of playing Left regardless of Jill’s strategy with probability 0.5 and playing Right regardless Jill’s strategy with probability 0.5. In the interaction depicted in Figure 14.1, the material results of different combinations of strategies are monetary payoffs. The first of the dollar amounts at each terminal node represents the earnings of the first player, Jill, and the second amount is how much Jack receives. So, for example, if Jill plays the strategy Up and Jack plays either strategy (ii) (Left regardless) or strategy (iii) (Left if Up and Right if Down), then both receive $40. But knowing the dollar outcomes is not sufficient to define the game. The “payoffs” that define the game are the player’s preferences over the outcomes. The

14.1  What Is a Game?

271

pairs of numbers along the bottom of Figure 14.1 are the utilities that the two players attach to the outcomes of the game. As discussed in Chapter 4 and elsewhere, these utilities are indices of preference. In this illustration, the utilities are ordinal utilities. That the outcome of Jill playing Up and Jack playing Left has a utility of 3 for both players means that it is the second-​ best outcome for both, nothing more. The utility Jill attaches to an outcome can depend on the result for Jack as well as on the result for her, and it may reflect preferences over features of the game besides the monetary payoffs at the end. In this particular example, Jill’s and Jack’s preferences over the outcomes match what they would be if they cared only about their own monetary return, but this need not always be the case. For example, even if Jack might prefer a state of affairs where an anonymous donor gives him $50 and Jill gets nothing to a state of affairs where a donor gives both $40, Jack might believe that in this interaction he has an obligation to reward Jill’s trust or kindness in playing Up. For this reason Jack might prefer the state of affairs where Jill plays Up, he plays Left, and they both receive $40 to the state of affairs where Jill plays Up, he plays Right, Jill gets nothing, and he gets $50. The crucial point is that the payoffs in games are preferences or utilities, not quantities of money or anything else. The monetary amounts shown in Figure 14.1 are redundant; the game is fully defined by the preferences, regardless of what the objects of the preferences might be. The final ingredient in the definition of a game is a specification of what the players know. In Figure 14.1, for example, Jack knows whether Jill has played Up or Down before he chooses, and Jill knows that Jack knows this and that Jack knows that Jill knows that Jack knows this, and so forth. In the terminology of game theory, the game shown in Figure 14.1 is one of perfect and complete information. Everything about the game is common knowledge. (That said, the full specification of the players’ knowledge is often left implicit.) The games with which we will be concerned involve players who are rational and intelligent, who have perfect recall of earlier moves in the game, and who possess common knowledge of the extensive form of the game and of their common rationality, intelligence, and memories. Game theorists usually take rationality to be expected utility maximizing. The players are “intelligent” in the sense that they can figure out anything about the game that the game theorist can. In advanced work, game theorists have found ways to relax these assumptions, and in other applications of game theory (such as evolutionary game theory) players may be unintelligent entities hardwired to particular strategies.

Game Theory

272

I a Down

Up

II

Left (3,3)

b

c Right (1,1)

II

Left

Right

(1,1)

(2,2)

Figure 14.2. Simultaneous play.

The outcomes or “solutions” to games often depend crucially on the knowledge players have. Consider the game shown in Figure 14.2, in which Jack and Jill have the same moves, but both the payoffs and the player’s knowledge differ. The dotted circle including nodes b and c is a conventional way of representing the fact that Jack does not know how Jill has chosen before he has to move. We suppose again that Jill can play Up or Down and that Jack can play Left or Right. The payoffs are different. Both players prefer the outcome of (Up, Left) to the outcome of (Down, Right), and both prefer the outcome of (Down, Right) to either the outcomes of (Up, Right) or (Down, Left). Because Jack does not know whether Jill has played Up (which would put him at node b) or Down (which would put him at node c), Jack’s strategies cannot specify different moves at node b and at node c. Whereas in the game of Figure 14.1 Jack had four strategies, in the game of Figure 14.2 he has only two strategies: to play Left or to play Right. In addition to extensive-​form representations of games, such as those shown in Figures 14.1 and 14.2, games can also be represented in “normal” (or “strategic”) form as in Figure  14.3. Figure  14.3a shows the player’s preferences for the strategy combinations in the game depicted in Figure 14.1, while Figure 14.3b shows the player’s preferences for the strategy combinations in the game depicted in Figure 14.2. (“Lu” abbreviates “play Left if Jill plays Up,” etc.) The strategic form contains information concerning possible strategies and their payoffs, but it is silent about the order of play. Game theorists are concerned with the “solution” to different kinds of games. That is, they seek to answer a normative question, “What are the most rational strategies for intelligent players to employ in the situation they are in?” and, on the assumption that people are intelligent and rational,

14.1  What Is a Game? (a)

273

Player II (Jack) Lu-Ld

Lu-Rd

Ru-Ld

Ru-Rd

Up

3, 3

3, 3

1, 4

1, 4

Down

4, 1

2, 2

4, 1

2, 2

Player I (Jill)

Figure 14.3a.  A normal form game with sequential play.

(b)

Player II (Jack) Left

Right

Up

3, 3

1, 1

Down

1, 1

2, 2

Player I (Jill)

Figure 14.3b.  A normal form game with simultaneous play.

they also answer a factual question concerning what the outcomes of games will actually be. Some games will not have a unique solution. In thinking about solutions, theorists distinguish cooperative from noncooperative games. In cooperative game theory, the players are able to make binding agreements. This makes possible more advantageous outcomes, because when players must pursue strategies without knowing whether others will adhere to agreements, the results can be self-​defeating. The so-​ called prisoners’ dilemma game (which we shall discuss shortly) illustrates this point. On the other hand, cooperative game theory does not explain what makes binding agreements possible, and for this reason it is generally regarded as less fundamental than noncooperative game theory. Except for a discussion of bargaining theory, we focus on noncooperative games. The most important solution concept in noncooperative game theory is that of a Nash equilibrium. A Nash equilibrium consists of a set of strategies such that each strategy is a best response to the strategies the other players have chosen. In Figure 14.3a, for example, the strategy pair (Down, [Ru, Rd]) is the unique Nash equilibrium strategy pair. If Jill plays Down, then Jack has no better move than to play Right (and no better strategy than [Ru, Rd] –​though [Lu, Rd] is equally good), and if Jack plays [Ru, Rd], then Jill has no better strategy than Down. Even though (Down, [Lu, Rd]) has the same payoffs, it is not a Nash equilibrium, because if Jack plays [Lu, Rd], then Jill’s best reply is to play Up rather than Down. The appeal of the notion of a Nash equilibrium is that given the strategy the other player has

274

Game Theory

chosen, neither player can do better (with respect to his or her own preferences) by deviating from the chosen strategy. In the game shown in Figures 14.2 and 14.3b, there are two Nash equilibria: (Up, Left) and (Down, Right). The latter might seem irrational, but given that Jill is playing Down, Jack would do worse by playing Left, and given that Jack is playing Right, Jill would do worse by playing Up. As the games illustrate, the outcomes of Nash equilibrium strategies need not be Pareto optimal. Some games have many Nash equilibria in pure strategies, while others have none. If one considers “mixed” as well as pure strategies, then, as Nash (1951) proved, every game with a finite set of players and a finite set of pure strategies has at least one Nash equilibrium set of strategies.

14.2  Moral Philosophy and Some Simple Games Human interactions are the subject matter of much of ethics: is it wrong to cheat on a test, if everyone else is doing so? Why should I cut down on my CO2 emissions if what I do as an individual makes no difference? Game theory, in illuminating the structure of social interactions, has clear relevance to ethics. Individuals require social cooperation to thrive, but, as philosophers have long recognized, such cooperation is problematic. In Hobbes’s state of nature, where there is little cooperation, “the life of man [is] solitary, poor, nasty, brutish, and short” (1651, p. 107). Under what conditions does cooperation emerge as a solution to a game? Hobbes thought that only an absolutist state could provide a secure basis for cooperation. Game theory can help us understand why sometimes cooperation succeeds and why it sometimes fails. It can also help us understand the strengths and weaknesses of different moral principles by treating those principles as strategies. For example, Hardin (1988) argues that the complications of strategic interactions throw a wrench into both applications and criticisms of utilitarianism. A sophisticated utilitarianism that recognizes these factors may be able to defend rights assignments as devices for simplifying strategic interactions and avoiding suboptimal outcomes. Kantian moral philosophy may profit from game-​theoretic analyses of the outcomes of everyone following the same rule (Schelling 1978; Elster 1989a; Roemer 2015). Problems of social cooperation (or social cooperation “games”) are often complicated, and it can be helpful to understand recurring patterns. Consider the prisoners’ dilemma game shown in Figure 14.4. Jill and Jack both have two strategies, and the payoffs (utilities) are as stated. Jill and Jack both have “dominant” strategies: no matter what Jack does, Jill does better by confessing; and no matter what Jill does, Jack does better by confessing.

14.2  Moral Philosophy and Some Simple Games

275

Player II (Jack) Don’t Confess

Confess

Don’t Confess

3, 3

1, 4

Confess

4, 1

2, 2

Player I (Jill)

Figure 14.4.  The prisoners’ dilemma.

Yet the result seems paradoxical, for if there were some way for Jill and Jack to cooperate and each not confess, then both would be better off. The game in Figure 14.4 is called “the prisoners’ dilemma” because of a story that goes with it. Jill and Jack are two prisoners who have committed a crime together and to whom the district attorney separately offers the choice between confessing and not confessing. Jack and Jill are in separate prison cells and cannot communicate with each other. The best outcome for Jill occurs if she confesses (and so informs on Jack) and Jack does not. Second best is if both refuse to confess. Third best is if both confess. Worst of all is not to confess when the other confesses. And so each, pursuing his or her own interests, winds up confessing. Rational self-​interest leads to an outcome that is suboptimal for the two criminals –​though not necessarily for the society. Prisoners’ dilemmas vividly represent problems of social cooperation, free riding, and public goods provision. They show conclusively that it is mistaken to suppose that the pursuit of rational self-​interest always has Pareto optimal results. Adam’s Smith’s “invisible hand” requires a more favorable choice situation than that of a prisoners’ dilemma. Individuals deciding whether to contribute to the production of a public good seem like players in an n-​person prisoners’ dilemma game. Free riding –​that is, enjoying the public good but not contributing to its provision –​is best of all. Second best is enjoying the public good and contributing one’s share. Third best is not contributing and doing without the public good. Worst is having others free ride on one’s own contribution. Whether public goods provision is best modeled as a prisoners’ dilemma is controversial (Hampton 1987), because doing without the public good may sometimes be worse than having others free ride on one’s contribution, in which case one has a game of “chicken” (see Figure 14.8) rather than a prisoners’ dilemma. In addition, problems involving public goods provision arise in a context of continuing social interactions, and it is often misleading to model them as one-​shot or single-​play games. If players face the choice of confessing or not confessing over and over again, they are not playing a prisoners’ dilemma, which is a

276

Game Theory

one-​shot game. Repeated interactions hold out the possibility of learning, and of evolving strategies that reward reciprocity but punish defection. Prisoners’ dilemmas can also model problems of moral cooperation (Parfit 1979). For example, each mother trying to save her own child from drowning in a beach accident may do less well than all would do if they cooperated in saving all the children. The collectively self-​defeating character of moral commitments that are subject to this problem is disturbing, and one might want to require that acceptable moral views not be collectively self-​defeating. But some familiar moral maxims seem to fail in this way. Not all problems of social interaction have the structure of a prisoners’ dilemma. There are also pure coordination problems (Lewis 1969; Schelling 1978), such as determining which side of the road to drive on. It does not matter whether everybody drives on the right or everybody drives on the left, but it matters a great deal that everybody drive on the same side! The normal form of a pure coordination problem is given in Figure 14.5. Pure coordination problems are “easier” in the sense that they have self-​enforcing solutions, which can arise spontaneously. Once a custom becomes established, everybody has an incentive to adhere to it. But pure coordination problems also have less bearing on many moral problems, in which there are conflicts of interest to resolve. Mixed coordination-​conflict games such as the one shown in Figure 14.6 are more relevant to moral philosophy. Games with this structure are called “battle of the sexes” games because of the following story that is used to illustrate them. A man and a woman want to go out together for an evening, and they have a choice between dinner and the theatre. Both would prefer to be together at either activity rather than splitting up and each going solo, but Jill would prefer to go out for dinner while Jack would prefer to go to the theatre. As in a pure coordination game, each player wants to do as the other does, but in this case which strategy they jointly decide upon is not a matter of indifference. Many conflicts of interest resemble battle-​of-​the-​sexes games. Now consider the rather different game in Figure  14.7. Jill and Jack might both want to do their part toward providing some collective good that requires both their efforts. It is best for each to cooperate if the other cooperates, but it is worst to cooperate if the other does not. For example, a road might be unusable because two bridges are out. Jack is able to repair only one bridge, and Jill is able to repair only one bridge. The road is very valuable to each if both bridges are repaired and worthless otherwise. All it takes in this case to make cooperation individually optimal is “assurance”

14.2  Moral Philosophy and Some Simple Games

277

Player II (Jack) Drive left

Drive right

Drive left

1, 1

0, 0

Drive right

0, 0

1,1

Player I (Jill)

Figure 14.5.  A pure coordination game. Player II (Jack) Dinner

Theatre

Dinner

2, 1

0, 0

Theatre

0, 0

1, 2

Player I (Jill)

Figure 14.6.  Battle of the sexes. Player II (Jack) Fix bridge #2 Do nothing Fix bridge #1

3, 3

0, 1

Do nothing

1, 0

1, 1

Player I (Jill)

Figure 14.7.  An assurance game.

that the other will cooperate, so this is known as the assurance problem (Sen 1967). In assurance problems, the obstacles in the way of social cooperation are much weaker than in a prisoners’ dilemma and the prospects for reconciling individual rationality and social optimality are brighter. One strategy for coping with prisoners’ dilemmas is to convert them into assurance problems by changing either incentives or personal motivations. The assurance game is also often called the “stag-​hunt” game after a passage in Rousseau’s Discourse on Inequality. If a deer was to be taken, every one saw that, in order to succeed, he must abide faithfully by his post: but if a hare happened to come within the reach of any one of them, it is not to be doubted that he pursued it without scruple, and, having seized his prey, cared very little, if by so doing he caused his companions to miss theirs.

Killing a deer is best for both, but each runs the risk of the other running off at the sign of a hare.

Game Theory

278

Player II (Jack) Slow down

Go fast

Slow down

3, 3

2, 4

Go fast

4, 2

1, 1

Player I (Jill )

Figure 14.8.  A game of chicken.

Suppose instead that Jack and Jill face the problem depicted by Figure 14.8. They are approaching an unregulated intersection in their cars. Both prefer to speed up if the other slows down. Second best is if both slow down. Third best is deferring to the other driver. By far the worst is if both speed up and crash into each other. As in the assurance game, Jill and Jack have no dominant strategy. (A strategy is dominant if and only if it leads to a better outcome than any alternative for a player regardless of what others do.) This game of “chicken” can arise when failing to provide a public good is worse than having others free ride on one’s contribution or when conflicts over possessions are costly. In such games, cooperation may be individually rational; or, as Sugden (1986) has argued, if the positions of the players are not fully symmetrical, then conventions can arise directing those who occupy different positions to play different strategies. Consider, for example, the convention of yielding to the car on the right. It enables drivers to avoid accidents and at the same time permitting one of the two cars to avoid slowing down. In addition, unlike a convention such as “smaller cars should yield to larger cars,” it imposes the burden of having to slow down more or less equally on all drivers. Property rights might be understood as conventions that evolved to solve problems with the structure of chicken games, although property rights have also been shaped by explicit legislation. As illustrated by games such as the prisoners’ dilemma, stag hunt, battle of the sexes, and chicken, there are many circumstances where there are conflicts of interest as well as advantages in cooperation (Hardin 1982 Hampton 1987; Taylor 1987, ch. 2).

14.3  Cooperation and Justice It is common for social interactions to feature both the possibility of benefits from cooperation and conflicts about how those benefits should be

14.3  Cooperation and Justice

279

distributed. In some cases, such as one-​shot prisoners’ dilemmas, we have seen that the gains to cooperation cannot be achieved by the separate endeavors of individuals who are rationally pursuing their own interests. But social life more often consists in continuing interactions, and for this reason, the possibilities for rational cooperation are not usually as bleak as the prisoners’ dilemma makes them appear. Even a structure with moves and rewards resembling a prisoners’ dilemma need not prevent self-​interested rational agents from cooperating if their interactions are repeated. Suppose that players in what would be a prisoners’ dilemma game (if it were played only once) will, with some probability, play again. The game has then changed and, in such iterated prisoners’ dilemmas, defection is no longer a dominant strategy because cooperative moves can elicit cooperative moves from others. It is provable that in such repeated games there are many equilibria and that some of them are Pareto optimal. Though the formal analysis of repeated games is complicated (Taylor 1987, chs. 3, 4), simple strategies such as tit-​for-​ tat (cooperate on the first move, then do whatever your opponent did on the previous move) do well against a wide range of alternatives (Axelrod 1984). A strategy that does strictly better against itself than does any other strategy is an “evolutionarily stable” strategy (Maynard Smith 1982). The biological terminology is not merely metaphor. Animal traits and behavior can be regarded as strategies, and game theory can be applied to the study of evolution. Tit-​for-​tat is not an evolutionarily stable strategy in an iterated prisoners’ dilemma, since other strategies (e.g., the strategy of always cooperating) do as well against tit-​for-​tat as tit-​for-​tat does when played against itself. But it can be proved that, when the probability of repeated play is high enough, tit-​for-​tat is “absolutely (collectively) stable” –​that is, it does no worse when played against itself than does any other strategy (Axelrod 1984, pp. 59, 207–​209; but see Press and Dyson 2012). When the probability of repeated play is high enough, a group of individuals who play tit-​for-​tat can “invade” a population of players who defect on every move and, if survival depends on success, eventually drive to extinction those who always defect. In this way cooperation can evolve among rational self-​interested agents, and one might be tempted to regard one chunk of morality as the conventional outcome of repeated games. Obviously, people sometimes compete for shares of divisible goods. Suppose that (1) the competition is between just two individuals and (2) if their claims are compatible then they both get what they claim, but if they collectively claim more than the total amount of the good, then they must fight. Many people would judge that in this case, all else being equal, the

280

Game Theory

fair thing to do would be to divide the good 50–​50. Why? Where do such judgments of fairness originate, and how are they justified? Although the question of justification is a subject for philosophical analysis, game theory also has something to contribute to both the justification and the explanation of fairness norms. To simplify, suppose (unrealistically, for the purposes of illustration) that everybody in the population is limited to one of three strategies: in the case of limited goods, (a) ask for 2/​3, (b) ask for 1/​2, (c) ask for 1/​3. People then meet randomly in circumstances where they compete for goods, and if their preferences are compatible, they get what they want. If their preferences are incompatible then they each get nothing. Let Jill and Jack be two members of the population who happen to meet. The normal form of the game they are playing might then be as shown in Figure 14.9. As before, the numbers within the (3×3) matrix are utilities:  2/​3 of the good is best and fighting over the good is worst. This game has three pure-​strategy Nash equilibria. Rather than exploring how rational and intelligent individuals would play this game, suppose that, in the population at the given moment, everybody is wedded to one or the other of the three strategies and that they interact with other members of the population repeatedly. In the course of these repeated interactions, some strategies prove to be successful and others not. Suppose, in addition, that there is some mechanism that selects for those whose strategies are successful or that permits individuals to update their strategies by comparing how they fared against the average payoff. This is a sketchy example of evolutionary game theory (borrowing from Skyrms 1996, ch. 1). Depending on the details of the dynamics, on whether the pairings are random, and on whether mutations are possible, one can show that populations are likely to evolve to the point where everyone has the strategy of asking for ½. This strategy is evolutionarily stable: any other strategy will do worse against it than the strategy does against itself. Although abstract and speculative, the game-​theoretic story of the evolution of an equal-​division norm shows how such a norm could have come into existence, why, once introduced, it is stable, and why it is efficient. If stability and efficiency are desirable properties for moral principles, then this account goes some distance toward justifying a norm of equal division, at least under certain conditions.

14.4  Paradoxes and Difficulties Game theory analysis does not always produce reasonable results. Suppose two players are playing a hundred-​move game, where the monetary payoffs

14.4  Paradoxes and Difficulties

281

Player II (Jack)

Player I (Jill)

Ask ²⁄³

Ask ½

Ask ¹⁄³

Ask ²⁄³

0, 0

0, 0

3, 1

Ask ½

0, 0

2, 2

2, 1

Ask ¹⁄³

1, 3

1, 2

1,1

Figure 14.9.  Divide the good.

for each move have the form of the prisoners’ dilemma game described by Figure 14.4. Suppose, for example, that the monetary payoff each round if both cooperate is $3, and that the monetary payoff if both defect is $2. If they both play conditionally cooperative strategies such as tit-​for-​tat, the outcome for both in the 100-​move game will be $300. If they both defect on every move, then the outcome for both will be only $200. Surely they should be able to find a way to cooperate in at least some of the hundred games. Yet there is an argument that apparently shows that, if the number of repetitions is known, then defecting (refusing to cooperate) on every move is the only rational strategy. The argument begins by noting that the hundredth repetition is just like the one-​shot game. Cooperating in order to elicit further cooperation from the other player is senseless, because there will be no further repetitions. So defection is the only rational action. Consider then the ninety-​ninth repetition. Since cooperation is rationally excluded on the hundredth repetition, cooperation on the ninety-​ninth round cannot elicit later cooperation, and defection cannot be “punished.” So defection is the only rational course in the ninety-​ninth repetition, too. Proceeding one move at a time, this “backward induction” argument yields the conclusion that the only rational strategy is to defect on every move. This conclusion is difficult to accept, and ordinary people manage to do a great deal better in their interactions than would rational game theorists. The apparent failure of the backwards induction argument may conceivably be explained by uncertainty about the number of plays in which the parties will be engaged. But experiments (see Dawes, van de Kragt, and Orbell 1990) show that, if individuals are permitted to form personal connections (which do not, however, involve any promises or contractual arrangements concerning their play), then they typically cooperate even in what seems to be a one-​shot prisoners’ dilemma game. Indeed, some evidence shows that students who study economics do worse in prisoners’ dilemma interactions than those who do not (Marwell and Ames 1981). Most people are

282

Game Theory

motivated by norms and commitments that constrain the strategies that they find most rational to pursue. It might instead be argued that if experimental subjects prefer the cooperative outcome, then they are not playing a prisoners’ dilemma at all. In the case of a genuine one-​shot prisoners’ dilemma, cooperating pairs do better than pairs who both defect because each player is benefited by the other. Individuals would do even better by defecting themselves while having a partner who cooperates. To understand how people play requires a sophisticated theory of choice that recognizes the role of norms –​including moral norms –​in decision making (Bicchieri 2002). The backward induction argument also leads to an apparent paradox. Suppose Jill fully accepts the backward induction argument and defects in the first game in the 100-​move game. But for some reason or other, Jack cooperates. What should Jill do on the next move? Jack’s first move decisively refutes Jill’s view of the game. It demonstrates that Jack is not rational (or does not understand the game) or that Jack does not believe that Jill is rational (or understands the game) or that Jack does not believe that Jill believes that Jack is rational or… But what if Jack’s move is motivated by his knowing that making a cooperative move will pose this perplexing problem for Jill and that such a move may thus induce Jill to cooperate in order to take advantage of Jack’s apparent “irrationality”! Is it then irrational after all to cooperate for some of the moves of a hundred-​repetition prisoners’ dilemma? It is hard to tell a convincing story of how a player should work out a suitable strategic response to contingencies that ought not to arise if all the players are rational and well schooled in game theory. Ethical applications of game theory are often controversial, because game theory is itself controversial. Not only are there problems in incorporating imperfect knowledge and imperfect rationality, but there is also disagreement about what solution concepts are appropriate even in cases of perfect knowledge and rationality. The traditional solution concept for noncooperative game theory is the Nash equilibrium, and most game theorists would hold that it is a reasonable necessary condition for rational play that solutions should be Nash equilibria. But it is not a sufficient condition: not all Nash equilibria are rational solutions. Consider the game shown in Figure  14.10. Player I  (Jill) plays Up or Down and Jack then plays Left or Right. Suppose that we have an ordinal representation of their preferences that reflects their preference for more money rather than less. Both the ($5, $5) and the ($9, $1) outcomes result from Nash equilibrium strategy pairs. If Jack has the strategy of playing Left if Jill plays Up and playing Right if Jill plays Down, then Jill’s best reply is to

14.4  Paradoxes and Difficulties

283

I a Up

II Left

($5, $5) (2,3)

Down

b

c Right

($0, $0) (1,1)

II

Left

($9, $1) (3,2)

Right

($0, $0) (1,1)

Figure 14.10.  Subgame perfect equilibrium.

play Up, and Jack’s strategy is a best reply to Up. If Jack has the strategy of playing Left regardless of what Jill does or playing Left if Jill plays Down and Right if Jill plays up, then Jill’s best reply is to play Down, and either of the strategies that has Jack playing Left is a best reply to Jill. So there are three Nash equilibria. But there is something odd about the equilibrium that results in both players’ getting $5. Jack’s threat to play Right if Jill plays Down, to which his strategy commits him, is not credible. Once Jill has played Down, Jack has a choice between getting $1 and getting nothing. If he is self-​interested and prefers more money to less, then he will play Left after all. Therefore, Jill should play Down. Only the two equilibria involving Jill playing Down are rational. They are the only “subgame perfect” equilibria (Selten 1975): they are not only Nash equilibria in the game as a whole, they are also equilibria in subgames (games that begin with intermediate nodes). The strategy pair (Up, Left if Up and Down if Right) is not subgame perfect because, in the subgame that begins after Jill has already played Down, Jack’s best move is not Right. Solution concepts that are stronger than the notion of a Nash equilibrium (such as the notion of a subgame perfect equilibrium) are needed, but there is disagreement on what they should be. These controversies are important to the ethical application of game theory, because they bear on questions such as whether retaliation is rational. Notice that the extensive-​form representation of games is crucial to identifying subgames. Whether or not normal-​form and extensive-​form representations are ultimately equivalent (Harper 1991), they have different virtues. Normal-​form representations are compact and make it easy to identify pure-​strategy Nash equilibria. Extensive-​form representations

284

Game Theory

make explicit the order of play, the player’s knowledge, and the structure of subgames. The game presented in Figure  14.10 is of particular interest owing to experimental results that concern so-​called ultimatum games. In these experiments, subjects are paired via computer in the laboratory (so that they never actually meet). One member of each pair is asked to offer any division of a sum of money –​of $10, for e­ xample –​to the other player. If the second player accepts the proposed division, then both players receive the amounts proposed; if the second player rejects the offer, then both players receive nothing. There are many Nash equilibria in the ultimatum game: for each $X less than $10, the strategy pair (Offer $X, Accept $X and no less than $X) is a Nash equilibrium. But the only subgame perfect equilibrium involves offering the smallest amount of money (since even a minimal payment is better than nothing). What happens in the experimental setting is quite different. Most of the offers are close to a 50–​50 split, and very unequal divisions are almost always rejected. The game shown in Figure 14.10 is a more restricted version in which only two proposals –​an equal division or a $9–​$1 split –​are possible. If Jill naively supposes that Jack will accept any offer, no matter how small, then she will offer a $9–​$1 split. However, if Jack is like most people, he will refuse the offer, and Jill will wind up with nothing. In reality, it is unlikely that Jill would be so foolish. Whether or not it is a shock to game theorists, it is no surprise to most people that those who play ultimatum games regularly reject very unequal offers. We think that two factors are responsible for these experimental results. First is the fact that people care about many things besides their own financial payoff. People care about fairness (particularly when they get the short end of the stick but not only then), and they are willing to forgo financial rewards in order to “send a message” to somebody who makes what they take to be an insulting offer. (If the unequal division is generated by a chance mechanism rather than proposed by the first player, people are more willing to accept it; see Blount 1995.) Since the second player thus actually prefers the outcome where both get nothing to the outcome where the proposer gets $9 and the second player gets $1, the game shown in Figure 14.10 misspecifies the preferences of the experimental subjects and so does not correctly model their strategic circumstances. The experimental results highlight the enormous difficulties in modeling a strategic interaction as literally a game. What game is it that people are actually playing? What game should they be playing? Until economists have succeeded in specifying what game it is that people are playing, game theory has little to contribute.

14.4  Paradoxes and Difficulties

285

Second, it may be reasonable for people to adopt “rules” or “policies” or to cultivate dispositions that in the real world of repeated encounters prevent them from being identified as or played for suckers. Although the experimental situation is specifically constructed as a one-​shot interaction, people may still behave according to these habits or rules. If it is easy to confuse one-​shot with repeated interactions, this behavior may be reasonable, even for players who care only about their own monetary return. Even if game theory were free of controversy and paradox, there would be grounds to hesitate before employing it to address ethical questions. First, there are qualms about its heavy reliance on information about the utilities of outcomes. Game theory is not strictly welfarist because it does not rely only on information concerning preferences. For example, nonpreference information determines what the possible strategies are and what their utility consequences will be. But evaluative applications of game theory are “quasi-​welfarist:” other ethically relevant factors –​rights, capabilities, needs, and reasons –​make no explicit appearance, and their role is usually heavily truncated (Roemer 1986a). For example, if the strategic problem is to distribute some set of commodities, then the solution will be the same regardless of what the commodities are as long as the feasible allocations and utility functions are the same. But people’s judgments about just distributions do not depend only on information about preferences or welfare; they depend also on what the goods to be distributed are. As we shall see, the outcomes reached in experimental tests of bargaining theory do not depend on preferences alone. Yaari and Bar-​Hillel (1984) have shown that people hold different ethical views about how to distribute avocados and grapefruits depending on what the fruits are wanted for, not just on how strongly the fruits are wanted. Many people would insist that justice requires a more egalitarian distribution if the fruits are wanted for their vitamins than if they are wanted for their taste. (This is analogous to a point we made earlier about the scope of the market: people care more about the equal distribution of some market goods than they do about others.) To apply game theory productively requires understanding the complexities of individual motivations and evaluations, including moral commitments. Another complication arises from the fact that individuals are capable of looking at situations from the perspective of the group. Rather than asking, “What is best for me?” they sometimes ask “What is best for us?” While some game theorists do incorporate sympathy, altruism, and even Sen’s notion of “commitment” in their models, they do not admit a perspective that considers what we should do and what the consequences will be for us. In ruling out this kind of perspective, game-​theoretic

286

Game Theory

analyses thus beg deep questions in social and political philosophy. Some philosophers (such as Gilbert 1990; Tuomela 2000, 2003; Sugden 2003; and Bacharach 2005) argue that a collective perspective is essential to the notion of social interaction and to explaining why real individuals do not play simple games the way that game theory prescribes and predicts (Dawes et al. 1990; see also Regan 1980, ch. 8; Hurley 1989, p. 145). What better explanation is there for the fact (see Bicchieri 2002) that experimental subjects are much more likely to cooperate when they have an opportunity to talk with one another, even on irrelevant subjects? The application of game theory to moral philosophy remains controversial, and the suitability of its assumptions must be addressed explicitly in each application.

14.5  Bargaining Theory and the Social Contract In order to clarify the difficulties involved in applying game theory to moral philosophy, we focus briefly on game theory that has been applied to ethical issues: two-​person cooperative bargaining theory. In John Nash’s classic analysis (1950), the two parties are supposed to have cardinal utilities that are not interpersonally comparable. The bargaining problem is to select a distribution of utility, u* to the first party and v* to the second party, from a closed convex set S of possible utility outcomes. (A set is closed if it includes its boundaries, and it is convex if, for any two points x, y in the set, it includes every point on the straight line segment between x and y.) The set S of possible outcomes contains a “threat point,” d = (ud, vd), consisting of the utilities the two individuals receive if no bargain is reached. Rational individuals will not settle for less than they can get if no bargain is made. So u* ≥ ud and v* ≥ vd. Nash proved that if the solution to the bargaining problem satisfies four further conditions, then the solution maximizes the product of the utility gains from bargaining. That is, (u*, v*) is the Nash solution when (u* − ud)(v* − vd) > (u − ud)(v − vd) for all points (u, v) in S other than (u*, v*) itself. Nash’s four conditions are as follows. 1. (Pareto optimality) There are no points (u, v) in S such that u > u* and v > v*. 2. (Invariance) Suppose (u*, v*) is the solution to the original bargaining problem in which the agents have utility functions U and V and that the utility functions are transformed to U´ and V´, where U´  =  aU + b and V´ = cV + d (here a, b, c, d are real numbers and a and c are positive). Then (au* + b, cv* + d) is the solution to the transformed

14.5  Bargaining Theory and the Social Contract

287

bargaining problem. This means that positive affine transformations of the utility functions are irrelevant to the solution in material terms. 3. (Symmetry) If ud = vd and if (v, u) is in S whenever (u, v) is in S, then u* = v*. If the agents are in symmetrical positions then the bargaining solution must also be symmetrical. 4. (Independence of irrelevant alternatives) Suppose that S´ is a subset of S and that both the threat point (ud, vd) and the solution (u*, v*) to the bargaining over the larger set S are also in S´. Then the solution to the bargaining game restricted to S´ should also be (u*, v*). (Notice that this is quite different from Arrow’s independence of irrelevant alternatives axiom IIA, discussed in Chapter 13.) These conditions seem plausible, and the Nash solution has many defenders. Yet it has awkward consequences, for if the feasible set of utility outcomes expands in person A’s favor, then the result may be that A actually receives less than before (Kalai and Smorodinsky 1975, p.  515). The Nash solution may also be morally objectionable in the extent to which it disadvantages both the poor (whose ud will be low) and the risk averse (whose utilities increase more slowly as a function of their physical shares). A further problem is that, in empirical studies, the bargains that individuals Jill and Jack reach diverge systematically from the Nash solution even when both individuals have perfect knowledge of their utility functions. In a “binary lottery” game, Jack and Jill bargain over the distribution of lottery tickets that give them a chance of winning separate prizes. Since the bargaining solution does not depend on interpersonal comparisons, one can stipulate that Jill’s utility is zero when she has no tickets –​that is, no probability of winning –​and that the utility of having all the tickets –​that is, of winning her prize for sure –​is 1. One can do the same for Jack. The utility for each player of having X% of the lottery tickets will then be X times 1 (the utility of getting the prize) plus (1 − X) times zero or simply X, the percentage of lottery tickets the player has. The Nash solution is a 50–​50 split of the lottery tickets. The Nash solution depends only on the probability regardless of what prizes Jack and Jill are bargaining for a chance to win. If both Jack and Jill will win $5 in their separate lotteries, then the Nash solution, a 50–​50 split of the lottery tickets, seems reasonable. But if Jack’s lottery pays off $5 while Jill’s lottery pays off $20, then the Nash solution of a 50–​50 split of the lottery tickets is less appealing. When the prizes differ, the actual experimental outcomes of the bargaining games are “bimodal,” with both

288

Game Theory

the Nash solution and the solution that equalizes monetary returns as the two modes (Roth and Malouf 1979). Roth, Malouf, and Murnighan (1981) argue persuasively that moral norms concerning just distributions affect the bargaining solutions. The person who stands to win the larger prize is seen as having a weaker moral claim on lottery tickets than the less advantaged person. At the same time, experimental work by Proulx et al. (1993) suggests that people’s moral views are fragile and depend on their experience in bargaining. Apparently, even purely positive bargaining theory must take cognizance of the moral commitments of the bargainers, and it is by no means obvious how this is to be accomplished formally (Gibbard 1990, esp. p. 262; Pettit 1990). There have been important attempts to apply bargaining theory in political philosophy. After all, apart from the rejection of interpersonal comparisons, Nash’s axioms are morally appealing. We want social distributions to be symmetrical and Pareto optimal, and although a violation of independence of irrelevant alternatives would not seem to be morally wrong, it would appear irrational. In Chapter 12, we briefly described David Gauthier’s application of bargaining theory to defend a theory of justice. More recently, Ken Binmore (1994, 1998, 2005) has argued for a much more complicated application of game theory to morality. Suppose that “nature” –​that is physical, chemical, biological, psychological, and technological facts –​places human beings in the position of playing repeated games with one another, such as the iterated prisoners’ dilemma. Binmore calls these many repeated interactions “the game of life.” Given what we know about repeated games, such as the iterated prisoners’ dilemma, we can be confident that the game of life has many equilibria. Some of these equilibria are much less efficient than others, and with the development of science and technology, new equilibria will be possible. Binmore speculates that principles of fairness serve to select among equilibria. Principles of fairness are in turn chosen in what Binmore calls “the game of morals,” the terms of which are set by socially determined weights that determine interpersonal welfare comparisons and by a built-​ in human propensity to consider what principles would be chosen from behind a veil of ignorance in which people believed it was equally likely that they could be any member of society. Although bargaining theory plays an important role in solving the game of morality, Binmore argues that it is illegitimate to assume that bargains will automatically be kept. When one realizes that bargains must be self-​enforcing, one winds up, Binmore maintains, with a view of justice very close to Rawls’s.

Suggestions for Further Reading

289

These examples suggest how game theory might be useful to moral and political philosophy. Despite the difficulties we have canvassed, game theory has provided an increasingly valuable conceptual framework in which to think about moral problems and problems in political philosophy.

Suggestions for Further Reading There are many introductions to game theory. Luce and Raiffa (1957) is a classic, and it is still very helpful. Harsanyi (1977b) is also accessible, though it defends some controversial positions. Good texts are Binmore (1992) and Osborne (2003). Most authors accept the view that solutions to noncooperative games should be Nash equilibria, but some (such as Bonnano 1991) disagree. For an overview of the contributions of game theory to the social sciences see Elster (2007). The prisoners’ dilemma has been discussed in hundreds of books and essays. Axelrod (1984) and Parfit (1984, ch. 2) are particularly helpful discussions. Frohlich et  al. (1975), Hardin (1982), Taylor and Ward (1982), and Hampton (1987) all question whether problems of collective goods provision are always best understood as prisoners’ dilemmas. For Sen’s discussion of the assurance game see Sen (1967, pp. 112–​124). There are other kinds of mixed coordination/​conflict games than those described here. For excellent discussions of how coordination can arise in various circumstances, particularly when there are asymmetries in the positions of different players, see Schotter (1981) and Sugden (1986). For a pathbreaking account of how people sometimes manage to solve apparently insolvable coordination problems, see Schelling (1960). For an attempt to provide a deeper theory of how coordination can be achieved, see Vanderschraaf (2001). For discussions of the backward induction argument and its difficulties, see Kreps et al. (1982), Kreps and Wilson (1982), Binmore (1987, 1988), Bicchieri (1988, 1990), Pettit and Sugden (1989), and Bonnano (1991). An excellent account of how game theory bears on multifarious equity questions is Young (1995), and a wonderful introduction to bargaining theory and its applications to moral philosophy is Barry (1989). Other useful discussions of bargaining can be found in Nash’s own short paper (1950), Braithwaite (1955), Luce and Raiffa (1957, pp. 124–​137), and Thomson and Lensberg (1989).

290

Game Theory

Critical discussion of Gauthier’s theory can be found in Paul et al. (1988). See also Gauthier’s collection of essays (1990). Skyrms (1996, 2003, 2014) offers excellent applications of evolutionary game theory to social norms and to various moral concepts.

Questions for Study and Discussion 1. Review the basic concepts: What is a game? What is a strategy? What is the difference between a strategy in a game and a move in a game? What is the difference between the normal form of a game and its extensive form? What is a Nash equilibrium? 2. Since Nash equilibrium strategies are best replies to each other, how is it possible for the outcome of a set of Nash equilibrium strategies not to be Pareto optimal? 3. Evaluate the following argument for cooperation in the case of a prisoners’ dilemma: “It is clearly better for both of us if we both cooperate. So the best thing for me to do is to cooperate.” 4. As we discussed in Chapter  5, when groups of four play collective goods games (which have, at least with respect to monetary payoffs, the same incentive structure as a prisoners’ dilemma) they initially show a good deal of cooperation, which decays when they play multiple rounds with changing group membership. However, if group members are able at a cost to themselves to “punish” others who do not contribute to the collective good, cooperative behavior becomes universal (Fehr and Gächter 2000; Camerer and Fehr 2006). Does this behavior refute game theory? How should a game theorist respond to these results? 5. Consider the following version of a “centipede” game (Figure 14.11). It gets its name from the way it looks: Jill plays first and can play down, in case she gets three-​quarters of the “pot”: $3 and Michael gets ¼ or $1, or she can play across and then the “dealer” doubles the pot and Michael can either play down, in which case he gets ¾ of the pot, or $6 and Jill gets $2. Or Michael can play across. On the tenth move, Michael has a choice between splitting the pot equally (which has grown to more than $2,000), or taking ¾ for himself. Backward induction implies that the unique rational solution is for Jill to play down on the first move. What would you do? How would you criticize the game-​theoretic solution?

Questions for Study and Discussion j

m

j

m

j

$3 $1

$2 $6

$12 $4

$8 $24

$48 $16

m

j

m

j

291 m

$1024, $1024

$32 $192 $128 $768 $512 $96 $64 $384 $256 $1536

Figure 14.11.  A centipede game.

6. Güth and van Damme (1998) report an experimental study of a variant of an ultimatum game in which the first person offers a division of some sum of money among three people, rather than two, and the second person accepts or rejects the offer. The third person does nothing. Typically the first person proposes to give little or nothing to the third person, and responders rarely reject proposals that are generous to themselves, even if they give nothing to the third person. How would you interpret the results? Why? What value, if any, do you see in experiments such as these? 7. When, if ever, can it be rational to address strategic problems from a perspective that considers what is best for “us”? Is such a perspective inconsistent with self-​interest, or can self-​interest be extended to encompass some group? 8. What do you think of the axioms of Nash bargaining theory? If they are satisfied, does it follow that the results of bargaining are fair? Why?

PART FIVE

CONCLUSIONS

We opened this volume with four examples that show the influence of ethical assumptions and commitments on work in economics. If this book has served its purpose as an analysis of the ethical presuppositions of economics and as an introduction to a range of ethical concepts and theories relevant to economic research, it should make the ethics implicit in these cases easier to understand. It should also suggest how economic analyses can benefit from subjecting their ethical presuppositions to reflective consideration. We have also –​especially in the last two Chapters 13 and 14 –​ described some of the tools that economists have developed that may be of use to moral philosophers. In this last part of our book, Part V, we wish to consider ways that the insights of economists and philosophers can complement one another. Accordingly, in Chapter  15 we revisit the four cases discussed in Chapters 2 and 3. In Chapter 16, we present ways in which grasping both ethics and economics can help in identifying good policies and principles.

293

FIFTEEN

Putting Economics and Ethics to Work

We have repeatedly argued that a better appreciation of ethics can contribute to economic analysis, and the tools and distinctions that economists have fashioned can contribute to moral philosophy. We hope that the examples we have already discussed have shown that these are not empty slogans  –​mere reminders that more knowledge and more sensitivity are usually better than less. In this chapter, we demonstrate the ways that ethics and economics complement one another by returning to the examples with which we began. We begin by discussing how a better understanding of ethics clarifies the peculiarities of the examples presented in Chapter 3 concerning involuntary unemployment and the use of overlapping generations models to study retirement savings. Then we look again at the argument that the World Bank should facilitate the transfer of polluting industries to poor countries and the efficiency argument for school vouchers as examples of standard welfare economics. With the help of the concepts, distinctions, and arguments of this book, we can offer a more sophisticated appraisal of these policies. Our aim in returning to these examples is to make the case that the richer conceptual apparatus and arguments we have explored in this volume allows us to appraise these and other policy issues in a deeper and more illuminating fashion. Whether in understanding the differences and commonalities between welfare economics and utilitarianism, elaborating different theories of well-​being, looking at various dimensions of freedom and the role liberty plays in alternative political philosophies, or clarifying the role of concepts such as equality, fairness, and rights in theories of justice, an essential goal for us has been to make these theories, concepts, and tools available to you, our readers, so as to assist you both in reflecting about abstract issues in ethics and economics and in interpreting and deciding about practical policy issues. 295

296

Putting Economics and Ethics to Work

15.1  Involuntary Unemployment and Moral Baselines We argued in Chapter  3 that controversies concerning the existence and analysis of involuntary unemployment result in part from conflating voluntary choice with rational choice and in part from normative disagreements among economists. The conflation leads some economists to regard as fully voluntary what are in fact generally considered paradigmatically involuntary choices, such as surrendering one’s wallet to a robber who threatens death otherwise (Colander quoted in Snowdon and Vane 1999, p.  233). As the example shows, a fully deliberate and rationally defensible choice need not be a voluntary choice. The fact that those who are unemployed could find work at less than the minimum wage, could steal cars, or could engage in prostitution is not by itself sufficient to make their unemployment voluntary. We argued that whether a choice is voluntary –​or, since voluntariness has degrees, the extent to which it is voluntary or involuntary –​depends on two conditions: the process that determines the alternatives among which one chooses, and whether any of the alternatives is minimally acceptable. Here we can put MacCallum’s three-​place analysis of freedom in Chapter 10 to work. Although some libertarians may regard a choice as involuntary if and only if there has been a violation of negative liberty, in our view, for an agent to choose an outcome voluntarily is for (a)  the agent to be free from (b) obstacles whose imposition is morally impermissible (c) to choose among a set of acceptable alternatives. As MacCallum’s framework shows, our construal of voluntariness thus does not coincide with any of the four concepts of freedom we discussed. The choice the robber gives the victim is involuntary both because it is morally impermissible to impose the obstacles he does and because the alternatives are not acceptable. Even if there has been no violation of negative liberty, if none of the alternatives open to an agent is “acceptable,” then we would classify the choice as involuntary. On this account, an individual cannot decide whether choices are “voluntary” without specifying the morally appropriate baseline: she needs to know what choices are minimally “acceptable” and what obstacles are “legitimate.” Since the obstacles to finding a job are most often (but not always) imposed in morally acceptable ways, whether someone without a job is unemployed involuntarily depends on whether the set of alternatives is acceptable. One way of determining which alternatives are acceptable is with reference to the social norms governing what sorts of jobs people can typically be expected to accept. When economists commit themselves to

15.1  Involuntary Unemployment and Moral Baselines

297

describing unemployment as voluntary or involuntary, they may implicitly endorse such norms or they may regard all unemployment as voluntary, because they assume that all rational choice is voluntary (insofar as there is no infringement on negative liberty). So a better appreciation of moral philosophy sheds light on what appears to be a controversy in positive economics concerning the nature and existence of involuntary unemployment. Moral philosophy clarifies the relationship among free, rational, and voluntary choices and helps to make explicit the conflicting moral norms that contribute to the controversy. Although day-​ to-​ day consumer choices are influenced by advertising, constrained by social norms, and limited by people’s incomes and laws against theft, they are typically voluntary. There are usually multiple acceptable alternatives. In addition, although the buying, selling, producing and consuming by other market participants impose limitations on one’s choices, these limitations do not violate rights or infringe upon negative liberty. In a robust job market, the choices of whether to take this job or that job or to prepare for this career or that career have many of the aspects of ordinary consumer choices. The stakes are higher, and the risks may be greater, but the obstacles are usually legitimate and the alternatives socially acceptable. Such choices are voluntary. Not all market choices are like this. Even outside the context of job loss, the constraints on choices may make them less than wholly voluntary. Consider, for example, the choice between owning and not owning a car. The Manhattan executive who prefers cabs and limos voluntarily chooses not to have a car. What about an inner-​city resident who spends hours on buses to get to the only available job? By moving in with relatives or working even longer hours, this person could own a car. There is a choice, but it is hard to see it as fully voluntary. It is questionable both whether the alternatives count as acceptable and whether the obstacles were justly imposed. If public policies favor auto travel over public transportation without addressing the consequences for those unable to drive, or if they favor public transportation linking affluent locations over inner-​city trips, or if redlining and other housing policies placed the poor in locations far from their workplaces, then it becomes harder to view inner-​city residents as choosing voluntarily whether to own a car or to be able to live on their own. Economists who invoke the notion of involuntary unemployment resist assimilating the choice to be unemployed to a normal market choice among alternative jobs. In times of high unemployment, most of the unemployed do not quit to look for a better job: they are fired and cannot find a new job that is roughly comparable to the one they have lost. It strains language,

298

Putting Economics and Ethics to Work

such economists might argue, to say that the choice not to work at subsistence wages makes one’s unemployment voluntary. Clearly, individual episodes of unemployment differ. People may lose their jobs because of poor performance, or they may refuse a job because they do not want to work under an African American or female supervisor. But unemployment rates do not double because of an epidemic of poor job performance or of racism and sexism in job seekers. People change slowly, while the unemployment rate changes quickly. “Voluntariness” is unmistakably a moral as well as a descriptive notion, with connections to other moral notions such as blame, coercion, and responsibility. Those who regard some spell of unemployment as involuntary may want to emphasize that this worker is out of a job “through no fault of his or her own” or to suggest that the alternatives facing the worker are unacceptable and that perhaps different institutions or policies could provide those who are unemployed with better choices. A claim that there is widespread involuntary unemployment at a particular time is thus a claim that there has been a “market failure” and that some remedy is needed. Critics of this view regard spells of unemployment as a normal part of a worker’s life in a market economy, or they regard all rational choices as voluntary (unless there is an infringement on negative liberty). These economists may not find spells of unemployment disturbing. Unemployment can result from the sensible choices workers make about how long to keep searching for a high-​quality job as opposed to taking a worse job sooner. In a market economy, this is to be expected. It is a feature, not a flaw; and there is no fix for this: no pattern of government intervention can overcome these fluctuations without imposing unacceptable costs of other kinds. The ups and downs of job holding and job seeking over a career are a normal part of the game. Which is the right way to think about unemployment? What attitude should one have toward these market outcomes? This is partly a question for positive economic theory, but it also rests on moral judgments concerning what constitutes voluntary choice, what societies owe to their members, and so what kinds of alternatives are acceptable and what obstacles are legitimate. Although economists can remain neutral and confine themselves to stating whether members of a society regard unemployment as voluntary or involuntary in the light of their own norms, it is hard to divorce descriptions of market processes and outcomes from attitudes of approval or disapproval or evaluative comparisons with alternatives. These linkages should not lie in the shadows, since much confusion and needless controversy

15.2  Overlapping Generations

299

could be prevented if economists were self-​conscious and explicit about the values that so many of them share. There is a further point we want to make about how moral norms influence economists’ views of involuntary unemployment. If one believes that there is a social commitment to providing all people with work that is commensurate with their skills and experience –​by government intervention, if need be –​then a good deal of unemployment is involuntary, and unemployment reflects a social failure. If one believes that there is no such social commitment, then those who are unemployed must learn not to expect anything better than they can get on the market. Like wheat farmers, who cannot charge more than the market price for the grade of wheat they grow no matter how lovingly they tend their fields, workers need to anchor their employment and wage expectations on the actual opportunities and the going wages. Even so, those who think that the unemployed are not justified in expecting better work than what is available need not believe that unemployment is on a par with a glut of wheat in an anti-​gluten era. Unemployment may indicate and contribute to diminished economic output and growth, and it may impose real suffering on both those who are out of work and those who accept much worse jobs than the ones they lost. Nevertheless, those who endorse market outcomes will not see unemployment as a market failure. On the contrary, they will regard markets as working exactly as they ought to, or anyhow as well as can be expected, given the unrealistic expectations of those who refuse the unattractive jobs that are available. Confusions aside, this is what is meant by claiming that unemployment is always voluntary and what is denied by those who insist that a great deal of unemployment is involuntary.

15.2  Overlapping Generations Attitudes toward markets are also central to the controversies concerning Paul Samuelson’s influential 1958 article on interest rates in an economy where saving is driven by the need to provide for retirement. Samuelson’s article makes some theoretical points that might by themselves be regarded as without any ethical content. He shows that in a barter economy of the kind that underpins much economic modeling, market competition in the context of overlapping generations would lead to an inefficient profile of overconsumption in the first period of life followed by penury in the second and third periods. In order to overcome this market imperfection, the economy requires an institutional structure that relies on trust  –​the

300

Putting Economics and Ethics to Work

necessary institution being either a social contract in which younger generations acknowledge an obligation to support the elderly, or fiat money whose value is taken on faith by the members of the community. Abba Lerner (1959a, b) is, however, critical of the pattern of savings and consumption that would result from this trust, because in a growing economy, it leads to lower average utility in every period than a pattern that equalizes consumption across the periods of a life. These claims by Samuelson and Lerner pertain to a highly stylized and abstract depiction of an economy. Their bearing on practical problems is questionable. We find this abstract theorizing instructive in three regards. First, although the preceding points can all be regarded as positive economics, their interest and the way they are presented and debated depend on their relevance to normative issues. Samuelson is intrigued by his model because it presents a case in which perfectly competitive markets with self-​interested agents are inefficient. Since most mainstream economists regard perfect competition as an ideal (other things being equal), this was a striking result, which drew an immediate response from William Meckling (1960 a, b). Meckling is skeptical about allowing trust to play any role and argues that, because Samuelson’s consumption pattern is infeasible, given Meckling’s view that no practice that depends on people reliably acting against their immediate self-​interest can survive, the competitive equilibrium is optimal after all. However, Meckling’s critique is mistaken. Although it would benefit the nonretirees to stop paying the retirees, if they believe that they will not themselves be supported when retired, once a social contract is in effect, or once fiat money is generally accepted, it is in the individual interest of each person who is not a retiree to accept the cash the retirees offer for the output. Without that cash, when the individual retires, he or she will starve. In this case, trust requires a convention, but it does not presuppose or require any self-​sacrifice. Moreover (and more fundamentally), the assumption implicit in Meckling’s analysis that economists should not invoke nonselfish motivations in explaining economic behavior is problematic. Even if non-​self-​interested action were required, there is no reason to rule out its possibility. As our discussion of the limits of markets showed, nonselfish motivations can and do influence economic results, and economies cannot function well without trust. Economists cannot understand behavior in the labor market without taking into account the views of fairness of employers and workers. Supply and demand do not fully determine the market wage, as there are points of friction in the hiring and search process. As the economist Peter Diamond has pointed out (1982, p. 219), “there is a wage

15.2  Overlapping Generations

301

that makes the worker indifferent between taking this job and waiting for his next job opportunity. There is a wage that makes the firm indifferent between hiring this worker and waiting for the next available worker. The bargaining problem is to agree on a wage between these two limits.” It is clear that nonselfish social norms can play a role here. Moral norms have consequences for market efficiency. Second, the model illuminates an important conceptual point about intergenerational transfers that bears on how one may think about Social Security. The conceptual point is that in real (physical) terms retired people mostly consume goods and services –​food, clothing, newscasts –​that are produced by younger, working people. At fifty, one cannot reasonably buy and store the food you will eat when you are seventy-​five! The “social contrivance” of money can obscure this point, but people at the end (and indeed at the beginning) of their lives live off the productive efforts of the people in the middle. With money in the picture, families can save money in their middle years and buy the goods they need from younger people at the end. From this perspective, Social Security is framed as a way of socializing this individual saving strategy, which leads to the thought that each generation should save enough to provide for its own old age. Alternatively, we can think of the working generation as collectively taxing itself to pay for its parents’ generation’s consumption in old age, in the expectation that they will receive the same treatment in their turn. Empirically, the U.S. Social Security system is much closer to the second model than the first, basically because it started out paying benefits to people who had not had the opportunity to save within the system. To switch from the current system to one in which each generation saved for itself would require one generation to do the heroic task of paying for its parents and saving for itself. No generation has so far volunteered for that role.1 To draw clearly the distinction between a savings and a transfer program and to distinguish among different optimality 1 The “pay it forward” plans for college finance proposed by governors of a number of states behave as the inverse of this. The proposal is to require that current students pay the cost of their education by imposing taxes that they will pay after college when they are employed. The education of their parents and grandparents, in contrast, was paid for to a large extent by their respective parents at a time when tuitions were kept low by subsides provided through the state tax system. Pay it forward proposes to break this implicit social contract. In schematic terms the current parents would in effect be saying to their kids, “My parents paid for my education and the parents of my parents paid for my parents’ education, but we don’t want to pay for yours!” This is oversimplified because the fraction of young people going to college has grown substantially, but the intergenerational logic works that way. See Bowen and McPherson (2016).

302

Putting Economics and Ethics to Work

criteria, as well as to design efficient paths of adjustment over time, require the specialized technical competence that economists possess. Third, Samuelson is oblique in making the normative points highlighted here, and his critics did not initially formulate their objections directly in ethical terms but instead in the guise of analytical disputes. We suggest that progress on both normative and positive issues would be more sustained if economists more frequently acknowledged the normative dimensions of their work. It matters whether one looks at intergenerational transfers as a utilitarian, such as Lerner, or from a perspective that is closer to a libertarian, such as Meckling’s, whereby ideally individuals save for themselves. Is fiat money a limitation on freedom (negative liberty), if one is required to accept it, and is such a limitation a reasonable price to pay for the greater lifetime well-​being it permits? Thus, for example, some economists tacitly accept the normative view that each family should receive a “fair” market return on its savings, thus looking at Social Security as a savings system. Others analyze Social Security principally from the standpoint of the adequacy of the living standards of recipients –​viewing Social Security as a transfer program. One main hope for this book is that it will make available concepts and tools that will facilitate greater clarity about normative contexts in both policy and theoretical work.

15.3  Do Pollution Transfers and Vouchers Promote Welfare? Both the efficiency argument for school vouchers and the argument that the World Bank should facilitate the transfer of polluting industries to poor countries invoke standard welfare economics. If parents are permitted to choose among competing schools, then their preferences will be satisfied at least as well as in a state-​sponsored system, and competition among schools will induce cost savings and improvements in quality. Concerns about the distributional consequences can be addressed by issuing vouchers, which can be as generous or stingy toward those who are poor as the citizenry wishes. Similarly, the massive inequalities in incomes, wages, and environmental quality between rich and poor countries and the fact that pollution damage is not a marketable good create opportunities for Pareto-​improving transfers (or industry relocations) between rich and poor countries, which the World Bank might be able to facilitate. Even uncompensated pollution transfers appear to be potential Pareto improvements, and cost–​benefit analysis would rank them above the status quo.

15.3  Do Pollution Transfers and Vouchers Promote Welfare?

303

As noted previously, there are objections to each of these proposals. In addition to distributional concerns, voucher plans must be responsive to the paternalistic and political reasons why governments support education. Economic arguments in favor of vouchers ignore crucial values such as social solidarity and a cultivated citizenry, which are not matters of either equity or efficiency in the satisfaction of preferences. Suppose, for example, that a voucher plan led to increased racial segregation. Focusing exclusively on preferences and efficiency, one might have no objection. With respect to the World Bank memorandum, we drew attention to five objections: 1. Relocating polluting industries to poor countries might do more harm than good by increasing the total amount of pollution in the world. 2. Considering only one alternative to the status quo risks overlooking still better possibilities. 3. The disparity in the initial bargaining position of rich and poor countries makes the exchange exploitative, whether or not it is mutually beneficial. 4. Welfare cannot be measured accurately by preference satisfaction. 5. Willingness to pay is not an accurate measure of preference satisfaction. The first and second criticisms do not challenge the memorandum’s normative framework, while the other three points question the moral basis of welfare economics. It is highly questionable (and indeed we have questioned) whether welfare should be measured by preference satisfaction, whether preference satisfaction can be measured by willingness to pay, and whether fairness considerations should be set to the side in this example. To be sure, it is important to recall that welfare economists are not committed to the view that efficiency in satisfying preferences is the only relevant consideration. So mainstream welfare economists could grant the third objection, though not the fourth and fifth. But in focusing exclusively on efficiency, welfare economists may exaggerate its importance as compared to alternative moral considerations, and the case for school vouchers and the World Bank memorandum illustrate this risk. The objections to both vouchers and pollution transfers point toward alternative frameworks for moral evaluation. For example, concerns about fairness might profitably be considered from the standpoint of social contract theories, such as Rawls’s. Or economists could assess policies

304

Putting Economics and Ethics to Work

involving school vouchers or exporting pollution from the perspective of freedom, rights, or equality. Let us return to the voucher proposal. Once one recognizes that efficiency in economics is usually interpreted as efficiency in promoting welfare and that economists measure welfare by preference satisfaction, one can see that the efficiency argument in defense of vouchers is not as unproblematic as it may initially appear. The choice among schools in a voucher system lies mainly with parents, but the welfare of many others depends on how children are educated. One cannot conclude that what parents choose will best satisfy all the relevant preferences. Indeed, the whole preference satisfaction framework seems inappropriate. If, as the measurement of welfare by the satisfaction of preferences supposes, people’s preferences are directed toward their own well-​b eing, then, unless the well-​b eing of parents and children coincides, satisfying the preferences of parents is not necessarily best for their children. Whether or not the preferences of parents are good indicators of their welfare, it is questionable whether what best satisfies their preferences effectively promotes the welfare of children and the society as a whole. Schools serve society and parents via serving children, and their effects on students are central to their assessment. However, the point of schooling is not to satisfy either the preferences of parents or the existing preferences of students. If the preferences students will later have as adults were given and were independent of schooling, then economists could intelligibly consider the extent to which schooling increases the capacity of students over their lifetime to satisfy their mature preferences. But those preferences are not given. Schooling aims to shape the preferences of children –​to instill values and to lead young people toward developing the capacities they need to participate fully in society. The World Bank memorandum similarly reduces the question of whether LDCs are “underpolluted” to the question of whether shifting dirty industries to the LDCs better satisfies people’s preferences. Cost–​benefit analysis measures preference satisfaction in terms of willingness to pay or to accept compensation. Applying cost–​benefit analysis here would be problematic, since the extent to which people prefer to avoid pollutants depends on their wealth and on their (possibly mistaken) beliefs concerning the consequences of the pollutants. Moreover, it is questionable whether distributional weights can address the difficulties, because such large changes in wealth would be accompanied by large changes in preferences. The memorandum partly avoids these problems, since it focuses on the actual effects

15.4  The Bearing of Capabilities and Primary Goods on Pollution

305

of additional pollution and since the revenues polluting industries would provide are supposed to compensate those who may be harmed by the pollution. Regardless of what people in LDCs believe about the effects of pollution and regardless of how their preferences would differ if they were not so poor, the welfare costs of the pollution, given the actual poverty and lack of development, is, the memorandum argues, in fact lower in LDCs. The World Bank memorandum takes for granted the status quo distribution of wealth and, unlike a great deal of welfare economics, measures all costs and benefits –​including injuries and deaths –​in terms of willingness to pay. A permanent injury to a child lowers the welfare of the child’s family much more if the family happens to live in a rich country, because the cost of the child’s care, the extent to which the child’s prospective earnings are diminished, and the amount that the family would be able, and thus willing, to pay to prevent the injury are all much greater in rich countries than in poor. Measuring preference satisfaction this way seems morally unacceptable, and clearly distributional weights are called for here. It is just as morally important to avoid crippling a poor child as to avoid crippling a rich one. Welfare economists thus commonly impute uniform wealth-​independent values to injury and death and avoid measuring these by individuals’ willingness to pay. Doing so creates conceptual strains. If dying or losing a limb is just as bad whether one is rich or poor, then why are not the values of other factors that influence health or longevity also independent of people’s wealth? Why allow willingness to pay to determine the values of some things, but not others? It appears that moral theory would be helpful here as well.

15.4  The Bearing of Capabilities and Primary Goods on Pollution Trades and Voucher Proposals Given the difficulties of measuring welfare by preference satisfaction in the cases of vouchers and pollution transfers, there may be advantages in appealing instead to more objective measures such as Sen’s and Nussbaum’s capabilities or Rawls’s primary goods. Schooling is relevant to many central human capabilities, and, although Rawls does not specifically include education as a primary good, it is relevant to opportunity and to having the capacity to exercise one’s freedoms. The capabilities affected by pollution would include the capability to breathe freely (or to continue breathing at all!), to drink safe water, and so on, but we shall focus on the consequences of pollution on health. (Rawls does not include health among the primary

306

Putting Economics and Ethics to Work

goods, but health is also relevant to opportunity, and for that reason some have argued that health should be included among the primary goods. In the case of some pollutants, such as those that increase the risk of prostate cancer (a disease that affects mainly older men), a given exposure will have a relatively minor effect on capabilities or primary goods in those LDCs (whose numbers are fortunately dwindling), in which people are likely to die young from other causes. A given dose of a pollutant may also have fewer negative health consequences if there is already very little pollution than if there is a great deal. On the other hand, the interaction among the effects of different pollutants, the generally worse health status of people in poor countries, and deficiencies in local systems for detecting and treating environmentally induced illnesses might cause pollution to limit capabilities or primary goods more severely in LDCs than in wealthy countries. Similarly, the numbers of people affected by pollution might be higher in poor countries. There is no good reason to believe that uncompensated transfers of pollution to poor countries would result in a net increase in primary goods or capabilities. Analysis of the consequences for capabilities or primary goods of compensated transfers is more complicated. The Pareto efficiency argument in favor of compensated transfers does not require any comparison of the pollution’s welfare effects in rich and poor countries, and in this sense it avoids the direct wealth bias noted earlier in the argument for uncompensated transfers. The argument relies instead on the fact that people in poor countries are willing to give up environmental quality in exchange for items such as food and for basic health care on terms that make the transfer of polluting industries from rich countries attractive to members of both poor and rich countries. These trade-​offs appear beneficial in terms of capabilities and primary goods, not merely in terms of preference satisfaction. However, as we pointed out in Chapter 9 on welfare economics, the fact that one allocation (such as a compensated transfer of pollution to a poorer country) is a Pareto improvement over the status quo may depend on the distribution of income. Thinking about distributional consequences reveals another problem. The idea of compensating a country is a cheat: transferring pollution is truly a Pareto improvement only if the preferences of every individual are at least as well satisfied after the transfer as before. But the compensation in the form of jobs in the new industry and general economic development may fail to reach the individuals who are harmed by the pollution. Since capabilities and primary goods permit interpersonal comparisons, one can ask a different question:  does a poor country that accepts a polluting industry in exchange for economic development gain in

15.5  A Utilitarian Perspective on Pollution Transfers and Educational 307

total welfare? Indeed, further questions of fairness would remain if different people benefited from the public health measures than were harmed by the pollution. The frameworks of capabilities and primary goods seem particularly useful in thinking about elementary education. Basic literacy and numeracy are central capabilities in modern societies and are among the social bases for self-​respect (a crucial Rawlsian primary good). One of the most significant failures of American schooling has been its failure to provide these for everyone. Whether a voucher system would reduce such failures compared to a public school system (other things being equal) is hard to judge, but it is perhaps a better question to ask than whether vouchers would better satisfy parental preferences than would public schooling. Would parental choice among schools lead schools to compete to nurture students’ capabilities and break the “tyranny of low expectations” that permits students from impoverished homes to pass through school without learning much? Or would a voucher system produce larger gaps than now exist in the development of basic capabilities between those whose parents are attentive and those whose parents are not? Would the lessened influence of teachers make primary education less conceptually, emotionally, or cognitively demanding? These are hard questions, but a study of reading scores and social integration might provide better evidence concerning the efficacy of vouchers than a study of the preferences of parents.

15.5  A Utilitarian Perspective on Pollution Transfers and Educational Vouchers For simplicity of comparison, assume that the utilitarian agrees with the welfare economist that welfare should be measured by the satisfaction of preferences. The difference between them is that the utilitarian, but not the typical welfare economist, permits interpersonal welfare comparisons. Since (as a first approximation) a compensated transfer of polluting industries enables citizens of both rich and poor countries to satisfy their preferences better, utilitarians would agree that a compensated transfer of polluting industries to poor countries would be an improvement over the status quo. However, a preference utilitarian will support the pollution transfer policy only if there are no available alternatives that better satisfy preferences. In addition to the status quo and to compensated transfers of pollution, consider the alternative of simply redistributing wealth. Because the

308

Putting Economics and Ethics to Work

marginal utility of income is much lower in rich than in poor countries, transferring income or resources from the rich to the poor will be welfare enhancing. (Utilitarians recognize, of course, that one needs to consider long-​run consequences and that technical and educational aid may well be better than cash, but these complications are not germane here.) How do the welfare consequences of one-​way transfers of resources from rich to poor countries compare to compensated pollution transfers like the one the World Bank memorandum contemplates? Imagine breaking the compensated transfer of pollution into two parts. Part 1 is the transfer of compensation (cash, jobs, medical facilities, etc.) from rich to poor countries; Part 2 is the transfer of pollution to poor countries. Part 1 is welfare enhancing for the reasons just discussed. Whether Part  2 is also welfare enhancing is more questionable. In rich countries, clean air is worth more in dollar terms, but is it worth more in interpersonally comparable utility terms? The best utilitarian outcome must be Pareto optimal, because any alternative that increased someone’s welfare (as measured by her preferences) without diminishing anyone else’s would have a larger total. As similar reasoning shows, Pareto improvements must increase total welfare. But some enhance welfare less than others, and some create barriers to future welfare-​increasing actions. Although compensated pollution transfers from rich to poor countries are Pareto improvements, most would not be Pareto improvements were it not for the disparities in resources between rich and poor countries. If inequalities in standards of living were greatly reduced, the principal source of the Pareto improvement –​namely, the very low incomes and wages in poor countries  –​would no longer be operating. In addition, the long-​run consequences of the less effective management of pollutants in poorer countries, which may not register in anyone’s willingness to pay, could easily tip the utilitarian scale. In the case of school vouchers in the United States, the most urgent problems and greatest failures of education are found among poor children, a significant proportion of whom are members of underrepresented minorities. Indeed, a common argument in favor of vouchers is that they provide these children with a way out of failed public schools. Vouchers might improve the education of poor children if they lessen the inequalities in resources or if they elicit better teaching. The effects of different voucher systems on the distribution of educational resources vary. But if there were a considerable reduction in material inequality, there might be few failing public schools and little interest in vouchers. On the other hand, the concern that vouchers

15.6  Other Ways of Evaluating Vouchers and Pollution Transfers

309

might disproportionately benefit affluent families would be eased by greater equality in background conditions.

15.6  Other Ways of Evaluating Vouchers and Pollution Transfers One objection to the proposal to transfer pollution is that it apparently permits rich countries unfairly to take advantage of the vulnerabilities of poor countries, and one objection to vouchers is that they might unfairly benefit the affluent. Do pollution transfers violate rights, limit liberties, or aggravate inequalities? One might think that there would be no freedom or rights-​based objection to compensated pollution transfers. After all, the transfers are purportedly mutually beneficial and would be agreed to voluntarily by both parties if there were no barriers to such exchanges. But this way of thinking is mistaken. There is a big difference between, on the one hand, George trading his watch to Mary in exchange for her briefcase and, on the other hand, some third party taking George’s watch and Mary’s briefcase and then giving the watch to Mary and the briefcase to George even if Mary prefers the watch to the briefcase, and George prefers the briefcase to the watch. What is at stake in the World Bank memorandum are not voluntary agreements among individuals in different countries to accept pollutants in exchange for compensation, but rather efforts by the World Bank to promote transfers that might approximate the results of individual voluntary exchanges. Is this move justified? Not all would agree. Most libertarians, for example, are concerned with processes rather than outcomes. On their view, one does not respect freedom by forcing people to do what they would have chosen to do, and one does not respect rights by forcing on people risks that they would voluntarily have chosen to impose upon themselves. Furthermore, because those who incur the costs of the additional pollution will not necessarily receive the compensation, there is no reason to believe that everyone would consent to the policy, and without consent from those affected by the pollution, its importation may violate their rights. Libertarians like Nozick are prepared to reject the entire welfare-​improving perspective of standard normative economics and to restrict government policies to the protection of individual rights. And even those who reject the libertarian account of rights and freedom may still have problems with the World Bank proposal according to their own accounts of how these values are best interpreted. On the other hand, it might seem easy to make a libertarian case for school vouchers, since libertarians are eager to limit the size and power

310

Putting Economics and Ethics to Work

of government. But because schooling has such important effects on children, the argument is not so simple. A vital question is what rights parents hold over their children and what rights children have (Brighouse 2000; Brighouse and Swift 2014). Do the rights of parents give them the sort of control over their children’s education that a voucher system permits? If instead one views the rights of parents over their children’s education as mainly instrumental to promoting children’s welfare interests rather than having some independent ground, then the case for or against vouchers will turn on how well vouchers promote the welfare of children. It is plausible to think that children have autonomy interests as well as welfare interests. According to some interpretations of liberalism, an essential aspect of a good life is that people be able to reflect on their values and commitments and to endorse them for themselves. This essential feature of a good life requires a social framework in which information and active criticism of alternative points of view are encouraged. This kind of liberal view may severely constrain the choices of parents on the grounds that certain types of education, such as fundamentalist religious education, discourage rational criticism of favored beliefs and judgments. Liberalism of this variety might impose substantial limitations on the educational programs for which vouchers could be used without necessarily ruling out the use of vouchers altogether. Shifting from concerns about freedom to concerns about equality raises a different set of questions. The World Bank memorandum takes for granted a background of extreme inequality. Uncompensated transfer of pollution to poor countries, which might be recommended on cost–​benefit grounds, could worsen these inequalities. Whether compensated transfers would increase or diminish inequalities depends on the compensation and the harm the pollution causes. The crucial point is that efficiency-​oriented policy analysis takes as given the unequal status quo and pays no attention to distribution. Those committed to values such as equal respect, solidarity, and fairness will reject the memorandum’s proposal and criticize the arguments that support it. Showing equal respect precludes assigning greater values to the lives, health, and comfort of those who are richer. Solidarity points toward protecting those who are least well off and toward sharing the burdens of technological advances rather than taking advantage of the poverty of the LDCs. And we have already questioned whether the proposal is fair. From an egalitarian perspective, the details of voucher programs and their consequences are also crucial. Egalitarians such as John Rawls take as their ideal a state of affairs in which the life prospects of individuals are

15.6  Other Ways of Evaluating Vouchers and Pollution Transfers

311

independent of the social and economic status of their parents. Egalitarians might well support well-​designed voucher plans that expand the opportunities of poor children, while acknowledging that schooling may not by itself overcome the enormous social and economic disparities among families. The crucial question would be whether voucher systems expand the opportunities of poor children more effectively than improvements in public education. The details are crucial. Questions about the overall justice of pollution transfers or school vouchers can be raised from a number of perspectives. To narrow the discussion, we shall focus on contractualist theories of justice within a single country. Consider the question, as it arises within the United States, of what constitutes “adequate” funding of schooling, whether through vouchers or through direct provision. Many state constitutions in the United States guarantee every child a right to an “adequate” education, and in some states the courts have become heavily involved in enforcing this right. Interpreting and enforcing this right have been complicated by the fact that in most states education is locally controlled and locally funded mainly by property taxes. Local funding means that affluent communities can afford to provide more expensive education than poor communities. In this context, courts and legislatures have to decide how much funding is required for an adequate education. A libertarian might argue that the state’s responsibility is limited to enforcing the duties families have to educate their children and assisting destitute families to enable them to provide minimal education. Call this the minimum principle. A  more ambitious requirement might require that the state prepare students for satisfactory performance in their roles as citizens and economic actors. Call this the adequacy principle. A much more egalitarian view would require that the education of children in poor communities be at least as well funded as the education of children in rich communities. This equal funding principle would likely require moving away altogether from using local property taxes to finance education. Similarly, one can list principles that might govern pollution transfers. A principle that would justify the memorandum’s proposal is to locate polluting industries so as to minimize economic costs. Let us call this the cost-​ minimizing principle. Its justice is questionable. A second might be to locate polluting industries where they will produce the least human suffering. Let us call this the pain-​minimizing principle. A third might be to locate polluting industries so that those who benefit the most from polluting industries endure most of the pollution costs. Let us call this the pay-​your-​way principle.

312

Putting Economics and Ethics to Work

These crude principles are intended only to suggest a range of alternatives whose justice can be considered. In deciding what principles should govern schooling or pollution transfers, one takes as given rights and duties that are built into the basic structure. So the moral choice here is unlike the choice among principles to govern the basic structure itself. The justice of voucher programs depends on the rights and duties of parents and children, and the justice of pollution transfers depends on property rights and rights to clean air and water. Specific problems of justice do not exist in a moral vacuum. Utilitarians will approach these questions by considering what policies fit best within a set of overall rules that maximize happiness or the satisfaction of preferences. Rawlsians will take for granted Rawls’s two principles of justice.

15.7 Conclusions We have not argued for specific policies, and our discussions of alternative construals of welfare and the relevance of nonwelfarist evaluative considerations do not easily determine whether, all things considered, it is right or wrong to encourage the migration of a polluting industry to some less developed country or whether a particular voucher system would be an improvement over current systems of government provision of schooling. Instead, we have called attention to the ethical values that are at stake in each of these examples. Rather than resolving these issues here, we have tried to show that a better understanding of the concept of well-​being and of the difficulties of measuring it, coupled with a better appreciation of alternative ways to evaluate economic outcomes, institutions, processes, and policies, permit a more sophisticated and informative appraisal of policy proposals. Not only is preference satisfaction a dubious hook on which to hang one’s policy conclusions, but many values besides welfare are relevant to the assessment of most economic policies. Mainstream normative economics has been too narrowly focused on preference satisfaction. That fact by no means makes it useless, but it can be extremely misleading if economists –​and those who listen to their advice –​do not understand its foundations and its omissions.

Questions for Study and Discussion 1. What does MacCallum’s schema for understanding freedom contribute to the analysis of voluntariness?

Questions for Study and Discussion

313

2. How much (if any) unemployment do you think is voluntary? How much responsibility do you think that government should assume for assisting those who are unemployed? What are your reasons? 3. How do the accounts of egalitarianism in Chapter 11 contribute to the debates concerning overlapping-​generations models? 4. What do the discussions of freedom, equality, and justice contribute to your understanding and assessment of the World Bank memorandum? 5. Given a better understanding of standard normative economics, has your attitude toward the World Bank memorandum or toward school vouchers changed? If so, how? 6. Why is it hard to apply the normative framework of standard welfare economics to the assessment of proposals for school vouchers? 7. Average life expectancy is rising in the United States, and affluent people live significantly longer on average than worse-​off people. How should these two facts influence thinking about how to adjust the Social Security system retirement age? Should it be the same for all? 8. How might school vouchers promote or impede equality of opportunity? 9. How would a luck egalitarian or a relational egalitarian go about assessing the World Bank memorandum? 10. How would a utilitarian go about assessing the World Bank memorandum? 11. Would utilitarians favor a proposal to replace public schools with private schools supplemented with school vouchers? What information would they need to have? 12. High-​income families spend much more on educational enrichment activities outside schools than low-​income people do, a gap that has grown over time. How does this fact influence your thinking about the application of criteria of adequacy and equality in the funding of schools?

SIXTEEN

Economics and Ethics, Hand in Hand

We have argued that a better appreciation of ethics can contribute to economic analysis, and at the same time held that the tools that economists have fashioned can be of use to moral philosophy. We hope that the many examples in this book and the detailed comments in Chapter 15 concerning applications of mainstream welfare economics have shown that these are not empty slogans or a mere reminder that more knowledge is almost always better than less. In this last chapter, we would like to do more to show how ethics and economics can complement one another; we wish to show that such a partnership is critically important. Our demonstration involves discussion of three central problems that societies today face. The first concerns the public obligation to provide health care in the face of improvements in technology and changing costs. The second concerns the rapid development of exchanges among individuals of goods and services brokered through Internet intermediaries such as Uber. These arenas of innovation have the potential to transform the character and conditions of modern life in ways that call out for economic analysis and ethical reflection. The last problem we discuss is as important as any that humans have ever faced. Solving it is crucial to the livelihoods, health, and even the lives of the whole of humanity. Economic growth and unrestrained consumption in wealthy nations combined with growing population and economic growth in the world as a whole have placed increasing strains on the environment and threaten massive climate changes with unknown but potentially devastating effects. Each of these problems demands of policy makers and citizens four things: 1. (Recognition) Policy makers and citizens must recognize the problems. Otherwise, people will not search for solutions or be willing to 314

Economics and Ethics, Hand in Hand

315

tolerate the costs that solutions may require. One of the greatest barriers to progress on the problems of climate change is the denial that they exist by politicians currying the favor of those whose incomes depend on fossil fuel industries. 2. (Will) Policy makers and citizens must have the will to address the problems. The peoples of the world must be willing to devote resources to studying and addressing them. Doing so might require sacrifice of some of their immediate interests in order to lessen the risks these problems create. 3. (Knowledge) Policy makers and citizens need knowledge to understand and address the problems. We do not know nearly enough, and some technical knowledge makes the problems worse by increasing our ability to destroy ourselves and others. But without understanding the nuances of these problems, it will never be possible to solve them. 4. (Cooperation) A  recognition of the problems coupled with the will to address them and the knowledge of how to address them will prove futile unless citizens within nations and across national boundaries are able to cooperate with one another. Some of these problems –​such as climate change and the threat of war –​cannot be solved by any single nation. Peoples and nations must be able to work together, and to do this they need goodwill, good leadership, and knowledge and organizations to help determine and carry out the best policies. Some of the prerequisites for solutions are already in place. These three problems are widely –​though certainly not universally –​recognized. The problem of marshaling political will in the face of large challenges that involve deeply conflicting interests and beliefs goes deep in both political theory and practical politics. Nevertheless, we conjecture that enough people who are in a position to contribute to the solution to these problems are willing to do so with the right leadership. Solving these problems demands technical knowledge, and advances in technology have been enormously helpful. But we do not suppose that technology is now or ever will be able by itself to solve these problems. The opportunities technology provides may carry risks and costs that are sometimes hidden and sometimes all too visible. Managing advances in technology is one of the crucial problems concerning the provision of health care. But advances in technology are, all the same, impressive. They have, for example, greatly diminished the risks of epidemics and famines, and they offer the prospect of greatly lessening

316

Economics and Ethics, Hand in Hand

our dependence on fossil fuels and greatly diminishing the environmental degradation caused by their extraction and use. But technology will not solve our problems. We need better knowledge of social processes, to which economists have a great deal to contribute, and we must also appeal to moral norms and our best understandings of human values. We must bring ethics into economics and both into our policies.

16.1 Health Care Although there is a settled consensus in every affluent country other than the United States that government has the responsibility of ensuring that all residents have access to reasonably comprehensive health care, it is worth discussing the ethical and economic arguments that have led to this consensus. This is an issue where concepts from economics are crucial to the moral issues, and moral considerations of rights, freedom, and equality are crucial supplements or corrections to an approach that focuses exclusively on welfare. A number of ethical considerations are relevant to the question of how health care should be provided:  benevolence or compassion does not let us stand by and let people suffer and die when we can easily help them; efficiency is of huge importance because of the enormous scale of health care. Freedom is important both in the sense of negative liberty, which a legal requirement to purchase health insurance appears to violate, and in the sense of effective freedom or opportunity, which ill health can severely limit. Control of infectious diseases, like national defense, has third-​party benefits and is essential to personal security. Equality with respect to life and death strikes many people as much more compelling than equality with respect to income and wealth. (Thus, large health disparities –​such as the twelve-​year difference in life expectancy between the African American and white residents of Washington, D.C. –​is viewed by many as blatantly unjust.) These considerations count in favor of ensuring that everyone has effective access to reasonably comprehensive health care. Even so, there remains the question of what policy design is best for achieving this goal. For example, why not provide those who are poor with vouchers that, like food stamps, can be used to purchase health care? Why not address the need for health care by providing additional income to the poor through measures such as redistributive taxation? At this point, some simple economic reasoning is crucial. Because health-​care needs are often unpredictable, urgent, and expensive, people need insurance. Paying for health care

16.1 Health Care

317

via savings alone would be infeasible for those who are poor and unlucky with their health, and it would be very inefficient for everyone else, who would have to impoverish themselves to have funds available to pay for health care when needed. Borrowing when one faces a health-​care emergency is hardly a feasible alternative. Insurance not only smooths out the health-​care expenditures of individuals over times, it also pools risks and so smooths out expenditures across individuals. There is also a paternalistic case for providing insurance to pay for office visits and tests that most people can easily afford, but that they might otherwise myopically avoid. So one might conclude that people who cannot pay for their own health insurance or health care should receive vouchers to purchase health insurance rather than vouchers to purchase health care. One might think that, supplemented by such vouchers, the rest could be left to the market. However, insurance itself creates two huge problems, which economists have usefully analyzed. The first, which is unavoidable, is “moral hazard.” If insurance pays for some good or service, it makes that good or service less expensive. And if a good or service is less expensive, people will consume more of it. If health insurance pays for doctor visits, people who are insured will visit the doctor more frequently. If one drug is more expensive than another but also slightly safer or more effective, people who are insured will choose the more expensive option. If people without insurance would be likely to make choices that were not in their best interest, then moral hazard may be a good thing. But if one supposes, as economists do, that people’s preferences are a good guide to their well-​being, weakening or eliminating the incentive to economize on the use of health-​care resources will lead to inefficient outcomes. Whether efficient or not, moral hazard will make health care more expensive. The prescription most economists would offer for mitigating the moral hazard implicit in health insurance would be to reintroduce financial incentives to economize on the use of health care by means of imposing copays and deductibles. Only the United States has attempted to rely heavily on this tactic, and with little success. High deductibles prevent the poor from accessing health care except when the needs are catastrophic. Low deductibles have little effect. Copays and deductibles also have large costs, many of which are invisible, because they manifest themselves in the anxiety and time of individuals who have to interpret and respond to sheaves of often-​ inscrutable health-​care bills. Moral hazard can also be limited with the help of nonfinancial incentives to avoid overusing health care. Long waits at the doctor’s office, surly staff, anxieties about diagnoses, and the petty humiliation of undressing before strangers and permitting them to probe

318

Economics and Ethics, Hand in Hand

one’s body all serve admirably in this role. (These rationing devices are of course also costly.) Finally, and most importantly, health insurance policies (whether public or private) can refuse to pay for certain services. In these ways, moral hazard can be mitigated, even if not eliminated. The second problem with insurance is adverse selection. If insurance is not mandatory and is “community rated” with “guaranteed issue” –​that is, available to all at the same premium regardless of their health history –​then it will be more attractive to those who anticipate becoming ill than to those who are more confident that they will remain healthy. So the premium must be higher than the average cost of providing health care, because the average health care costs of those who purchase insurance will be higher than the average costs in the population as a whole. Depending on how risk adverse individuals are, community rated premiums may be so high that only a small portion of the population is insured. For example, if women have the option of purchasing insurance that covers the costs of pregnancy (which in the United States is well above ten thousand dollars), those who seek to become pregnant will prefer policies that cover the costs of pregnancy. To stay in business, insurance companies will have to charge so much for those policies that few women who are not planning to become pregnant will purchase them. The result will be insurance premiums so high that the policies will be unattractive even to women who are planning on becoming pregnant. There are two cures for adverse selection. The first is to shift from community-​ rated health-​ care premiums to “risk-​ adjusted” premiums, whereby insurance companies charge different premiums to different individuals, depending on their health histories and prospects. Because individuals know more about their health than insurance companies can know, this is only a partial solution: the insured will still be somewhat less healthy than the population as a whole. In addition, administering a system of voluntary private health insurance with risk-​adjusted premiums creates significant administrative costs. Finally, in such a system those with “preexisting conditions,” whose expected care costs are very high, are unable to purchase insurance. To the extent to which health insurance companies become able to estimate future health care costs, health insurance with risk-​adjusted premiums undermines risk sharing, which is one of the main advantages of insurance. The second response is to recognize that adverse selection undermines the feasibility of leaving health insurance to the private choices of individuals as assisted by vouchers. The resulting solution to the problem of adverse selection is to make health insurance universal, guaranteeing that the

16.1 Health Care

319

average health costs of those who are insured are the same as the average health costs of the population as a whole. When large companies purchase health insurance for all their employees, the effect is similar to universal provision in terms of sharing risks, but many people are of course left out of such pools. Economics contributes to our understanding of how to deal with moral hazard and adverse selection. A  further contribution of economics is to study the effects of asymmetries of information. Doctors, nurses, and physical therapists often know much more about the patient’s health and treatment options than the patient herself does. This asymmetry creates a need for licensing and regulation by professional organizations, consumer groups, and often government enforcement. Patients need assurance that the drugs they purchase are unadulterated, that the tests they undergo are necessary, and that the treatments health professionals recommend are well advised. If governments were to play no role in protecting consumers of health care, other organizations would need to step in. And once in place, any agency that licenses or recommends some courses of action rather than others creates an opportunity for corruption or “rent seeking,” as practitioners and providers seek to use the licensing and regulation to raise their incomes. Unfortunately, in diagnosing these dangers and clarifying the peculiarities of health-​care markets, economists have no easy solutions to offer. With the help of the clarification of the problems of providing health care that economists have provided one can address a number of moral considerations. First is a simple concern: most people are unwilling to let people suffer or die, at least when their health problems can be relatively easily addressed; they are not willing to let uninsured accident victims bleed to death by the roadside. So, realistically, societies are bound to provide some health insurance for everyone, whether they choose it or not; and there is some case to be made for charging them for the privilege of living in a decent concerned society. Indeed, even from a libertarian perspective, there is a case for public health measures such as the control of infectious diseases, where (as with national defense) the benefits of practices like widespread vaccination or spraying for infectious pests cannot be limited to those who pay. Although Rawls abstracts from issues concerning health care, his principles of justice appear to have an immediate application to them. Writing from a Rawlsian perspective, Norman Daniels (1985, 2007) emphasizes the importance of health and health care to fair equality of opportunity. Two people cannot fairly compete for jobs if they have very different health

320

Economics and Ethics, Hand in Hand

status; nor do they have the same opportunities to engage in activities generally. The problems of adverse selection also provide the basis for an efficiency argument in defense of universal health care. It is easy to see the general outlines of such an argument: universal health care produces huge savings in administrative costs and shares risks, thus increasing well-​being. Even if it were to yield no improvement in health and no savings in resources, it avoids the anxieties and economic costs that result from the uninsured costs of health care, and the inefficient choices people make of jobs and even spouses as they maneuver to secure health insurance. The same considerations coupled with the simple recognition of how important health is to a person’s welfare ground a plausible utilitarian case for universal health care. The problems of health care provide a vivid illustration of how important it is to integrate the insights of economics and moral philosophy. In short: without the economics, one cannot understand how health care markets and health insurance work, and without moral philosophy, one cannot understand why the private health insurance market leads to morally objectionable outcomes.

16.2  Uber and the “Gig” Economy Uber is a ride-​booking service that uses the Internet to coordinate trip planning by individuals seeking to hire a ride and individuals seeking to provide a ride. Uber recruited its first “driver–​partner” in 2012, and now has more than 160,000 of them. Uber is, perhaps along with AirBNB (an analogous service for renting rooms or houses), the leading example of what is sometimes called “the gig economy” (a term we prefer to “sharing economy” since the modus operandi of these businesses is actually to replace “sharing” –​the friendly exchange of favors among acquaintances –​with commercial transactions). Some observers predict that internet-​driven services like these will mushroom in importance for two main reasons. First, the Internet can sharply reduce the costs of search for deals between individual buyers and sellers in “real time,” making the match process more efficient. Second, the very high level of economic and educational inequality in the United States and other countries tends to produce a significant group of people who value their time highly and are willing to pay to outsource personal services and a much larger population who are willing to provide those services at a relatively low price. (For another example: the startup InstaKart connects a busy buyer to someone who will go to the store and do his or her shopping.)

16.2  Uber and the “Gig” Economy

321

We have no way of predicting how widespread the gig economy will become, but the rapid emergence of these businesses is a striking phenomenon. This development raises in concrete terms a number of ethical issues that have figured in earlier discussions in this volume. Some of the most interesting questions arise in the context of companies that facilitate the purchase and sale of human labor, in contrast to companies (like AirBNB) that involve the rental of inanimate objects such as a room, a car, or a piece of equipment. We will focus mainly on Uber, as the most studied of the labor-​using firms, a category that includes operations like Amazon’s Mechanical Turk, which enables people to hire the services of individuals to perform Internet services, generate written copy, or even participate in social experiments. Many see the emerging gig economy as part of a larger social and economic trend that Nicholas Lemann has characterized as a movement from “the organization society” to “the transaction society” –​a movement from long-​term relationships coordinated by stable hierarchical organizations to the fleeting transactions between individuals coordinated by markets (Lemann 2012). This broader movement can be seen in phenomena like the decline of traditional defined-​benefit pensions and their replacement by defined contribution 401-​Ks, which shift the responsibility (and the risk) of managing and protecting a firm’s employees’ retirement assets from the company to the individual. Defined contribution 401-​K’s present workers with a Faustian bargain, offering them flexibility in exchange for security. The main business of companies in the gig economy is matchmaking, connecting people who have a service to sell with interested buyers, and doing that quickly and efficiently via the Internet. These companies facilitate “capitalist acts between consenting adults” (in Robert Nozick’s memorable phrase) and surely appeal strongly to libertarian sentiments. Such sentiments appear to be widespread in the tech industry. For example, Peter Thiel, a founder of PayPal and a libertarian, writes, “As an entrepreneur and investor, I have focused my efforts on the Internet. In the late 1990s, the founding vision of PayPal centered on the creation of a new world currency, free from all government control and dilution –​the end of monetary sovereignty, as it were… By starting a new Internet business, an entrepreneur may create a new world. The hope of the Internet is that these new worlds will impact and force change on the existing social and political order.”1 Early on, in keeping with such a libertarian outlook, companies including Uber and AirBNB adopted a strongly defiant antiregulatory stance, 1 www.cato-​unbound.org/​2009/​04/​13/​peter-​thiel/​education-​libertarian.

322

Economics and Ethics, Hand in Hand

claiming that they were not providing transportation or lodging, but were simply neutral brokers connecting buyers and sellers. Although leading gig economy firms have recently begun to back off from this stance under pressure from regulators, it is worth considering the “no regulation” argument in its most aggressive form. We can spell out this position as follows: Uber is simply a communications device. Indeed, despite its stock market valuation in the tens of billions, Uber likes to characterize itself as an “app.” It facilitates voluntary commercial transactions between individuals. On this view, Uber has no more responsibility for what these individuals work out between themselves than the phone company would have if you called someone you knew and offered to pay her ten dollars to drive you to the airport (or offered a larger sum to commit a fraud). Uber resists regulation on these grounds and in France even chose to operate for a time in direct defiance of a law forbidding its operation. This no regulation argument can be divided into two parts: (1) that Uber is not a party to the transaction but simply a tool or app that these cooperating parties use and (2) that there is no governmental interest in regulating private transactions regarding personal transportation, or, to extend the argument to other gig economy businesses, room rental, rental of one’s private car, and so on. Perhaps Uber’s most striking invocation of the first claim occurred in 2014, when Uber defended itself against a complaint that some of its drivers avoided accepting blind passengers with service dogs. (In the case of Uber’s brand UberX, the cars in question belong to the drivers, who may not want dog hair in their car; one driver put a passenger’s service dog in his trunk.) When sued by the National Federation of the Blind, Uber denied responsibility, claiming it had no control over the drivers, who are independent contractors. It made this assertion despite the fact that its guidance to users included advising blind passengers to inform drivers in advance of their use of a service dog. (The case has not been resolved at this writing.) In fact, Uber screens drivers and does a good deal to structure the behavior of its “driver–​ partners.” Uber specifies how revenue is shared between the company and the driver, and it dictates other practices for drivers. Recently, the California Labor Commission concluded that “defendants hold themselves out as nothing more than a neutral technological platform, designed simply to enable drivers and passengers to transact the business of transportation. … The reality, however, is that Defendants are involved in every aspect of the operation.”2 www.slate.com/ ​ blogs/​ m oneybox/​ 2 015/​ 0 6/​ 1 7/​ u ber_​ d rivers_​ r uled_​ e mployees_​ by_ california_​labor_​commission.html. 2

16.2  Uber and the “Gig” Economy

323

As to claim (2) that there is no public interest in these transactions, scale has to be recognized as an important factor. Obviously the cost of regulating people’s casual agreements to drive someone to the airport or to let a relative use a bedroom in exchange for a bottle of wine far exceeds any plausible benefit. Moreover, attempts at regulation would intrude on people’s privacy in unacceptable ways. But Uber is a large commercial enterprise operating at significant scale to arrange rides; its policies and actions may have consequences that affect persons who are not parties to the transactions. More generally, Uber’s policies and practices affect drivers, users of Uber services, and third parties in ways that may impose moral obligations on Uber and that in some cases justify legal regulations to enforce those obligations. Consider the effects on third parties first. The claim that transactions between Uber drivers and passengers (or AirBnB renters and lodgers) are purely private ignores possible externalities that may affect other people. For example, the risk of unsafe Uber cars and drivers threatens the safety and well-​being of other drivers and passengers in other cars. An additional risk is that Uber drivers may not have insurance that applies when their vehicle is used for commercial purposes. Because Uber can monitor the driving records of drivers, the safety of the cars they drive, and the insurance they carry (or which Uber carries on their behalf), the company has a moral responsibility, and arguably should have a legal responsibility, to attend to these concerns. Some observers have further asserted that Uber and similar ride services (such as Lyft) add to pollution levels and congestion in busy areas –​both classic examples of externalities. Whether these external effects are significant is a difficult empirical question. If Uber and its kin substitute for taxis or discourage people from owning or driving their own cars, their presence may have no net impact or could even reduce pollution and crowding. But if Uber instead substitutes for mass transit, its negative effects could be nontrivial. These effects could justify enforceable regulations to limit the number of taxis and ride-​sharing vehicles on the road at any one time (as in New York City’s taxi medallion system). Uber users have interests and moral claims that overlap with those of third parties. While users, unlike third parties, have chosen to rely on the safety of Uber drivers and cars, they have little ability as individuals to judge the reliability of the stranger who is about to pick them up. Uber is much better positioned to develop such information, and is also better positioned than are individual drivers to accept moral, financial, and legal responsibility when things go wrong. Of course, Uber has a commercial interest in developing a good record for safety and reliability, but these ethical considerations impose stringent obligations, regardless of the effects on Uber’s bottom line.

324

Economics and Ethics, Hand in Hand

Uber drivers (and more generally suppliers of services in the gig economy) may have moral claims on the companies that connect them to buyers. So far in the United States, Uber drivers, people who sell their services through Amazon’s Mechanical Turk, and other workers in the gig economy are treated as “independent contractors,” a legal status that deprives them of many of the rights and protections afforded to employees under federal and state labor law (including rights to bargain collectively, to be paid a minimum wage, to have unemployment insurance, to receive employer contributions to Social Security, and to have access to health insurance through their employer). There are currently active disputes about whether Uber drivers should be classified as employees. Uber exercises considerable control over its drivers, as employers do, but at the same time Uber drivers, like independent contractors, provide their own tools (cars) and have substantial discretion over their hours and places of work. Behind the legal arguments in these disputes lie moral concerns. What are the ethical bases for assuring that workers are paid a minimum wage or have the right to form a union with their fellows? If the ethical rationales for these rules apply to drivers, then Uber has a moral even if not a legal responsibility to safeguard the interests those rules protect. And governments also have a responsibility to consider whether these rules should be legally enforced against Uber. Since the status of Uber drivers is intermediate between typical workers and independent contractors, it may be that a good solution would involve defining a new category of workers –​Harris and Kreuger (2015) suggest it might be called “independent worker” –​that would protect these workers’ interests. (In the past the U.S. Department of Labor has defined other special subcategories of employees, such as “tipped” workers and those who do “piecework” in their homes, for similar reasons.) Consider, for example, the right workers have to form unions and bargain collectively. A core ethical rationale for this provision is that it reduces the inequality of power between a large employer and each of its individual employees, by permitting the workers to coordinate their bargaining positions.3 (Independent contractors do not have the right to bargain An alternative to collective bargaining as a way of neutralizing the power balance between employers and employees is active competition for workers among alternative employers. It is an interesting economic question whether businesses like Uber’s will evolve toward monopoly or toward a competitive market with multiple suppliers, at least in large cities. The premium placed on being able to make matches quickly and the likely difficulty of differentiating services across different suppliers suggest that ride brokering, like some other Internet-​provided services including search (Google) and social networking (Facebook), may be a natural monopoly, at least in the case of ride hailing at the local level. 3

16.2  Uber and the “Gig” Economy

325

collectively.) At present Uber presents to drivers a set of work rules and a rule for sharing the revenue from rides between the driver and the company on a take-​it-​or-​leave-​it basis. In the face of such one-​sided power, there is an ethical rationale for permitting collective bargaining and requiring the company to negotiate with a legitimately constituted union (should drivers choose to create one). Other worker protections may be more difficult to adapt to the gig economy. Consider regulations that are imposed on firms to protect workers such as laws governing occupational health and safety or limiting normal working hours to forty hours per week. Providing the same protections in the gig economy requires explicitly paternalistic regulations that are both intrusive and difficult or impossible to enforce. There is consequently a risk that the gig economy can undermine the worker protections for which generations of workers have struggled. For example, consider the minimum wage. Minimum wages are justified in large measure by a consequentialist argument that on balance the higher wages received by those who keep their jobs when a minimum wage is imposed will obtain benefits that outweigh the costs of some people losing their jobs. Obviously the validity of this argument depends both on empirical claims about the consequences of setting the minimum at a given level, and on the weights attached to the positive and negative effects in a particular consequentialist calculus (Wilkinson 2004). A second line of argument for the minimum wage derives from a concern with what we earlier called “effective liberty.” As we argued in Chapter 15, it is questionable whether workers who are compelled to accept jobs that pay them less than a decent social minimum are making a genuinely voluntary choice of employment or whether they are to some degree being compelled to accept employment. Hall and Krueger (2015) maintain that few Uber drivers were unemployed when they took up the work, and that in interviews the drivers generally compare the working conditions –​flexibility and autonomy –​in driving for Uber favorably even to better-​paid employment of a more conventional kind. Although not made explicit, the point seems to be that Uber drivers are making a genuinely voluntary choice to sign up  –​not being compelled by a lack of alternatives  –​and that they see themselves as being at least as well off as similar people in other jobs.4 4 Even if Uber drivers are not in any way coerced, there is no guarantee that this conclusion holds throughout the gig economy, nor that it will continue to hold if these kinds of work arrangements become much more widespread in the economy.

326

Economics and Ethics, Hand in Hand

Whatever the merits of setting a particular level of the minimum wage, the underlying purpose of a policy putting a floor under wages suggests that the floor be as nearly universal as possible. If a particular kind of work or form of contract is exempted from the minimum wage, businesses involved in that work will gain a competitive advantage and will tend to expand, thus undermining the principle that there is a standard minimum amount people should be paid for their labor. Moreover, firms will have an incentive to organize their businesses in a way that allows them to avoid meeting a minimum wage requirement. These considerations make it important to address the question of whether and how some notion of minimum wage can apply in the gig economy. It is difficult to know whether Uber drivers generally earn more than the minimum wage, currently set by federal law at $7.25 per hour. There is an obvious conflict of interest between Uber and its drivers over how many cars should be on the street offering their services at any one time. Because a major selling point for Uber is that they can supply a ride fast, and because Uber bears no costs for having idle drivers cruising around putting themselves in the best spots to get a passenger, Uber wants to maximize the number of available drivers (which drives down the average “wage”). On the other hand, each Uber driver would prefer to have fewer drivers available (other than himself or herself) to minimize downtime. One challenge in estimating an hourly wage is to determine when a driver is working. Is she working only when she has accepted a ride or is she working when she has the app on and is available for service? Moreover, under current circumstances a driver can wait for rides from more than one service at the same time (for example, Uber and Lyft). For whom is she working at that time? Drivers also have discretion over whether to “bid” for a particular ride, giving drivers some discretion over the amount they will earn per hour. Finally, because drivers in some of Uber’s brands own and operate through their own capital (a car), operating and maintenance costs need to be measured and accounted for. Little work has been done, to our knowledge, in developing such estimates. There have been reports that workers in the Amazon Mechanical Turk system typically earn between one dollar and five dollars per hour, well below the minimum wage, but the evidence for this claim is unclear. This situation represents an instructive problem at the intersection of ethics and public policy. There are, on one hand, good moral and practical reasons to prevent Uber from undercutting an established social policy of putting a floor under the amount workers should earn for their labor. Moreover, there is good reason to worry that some gig economy businesses

16.3  Environmental Protection and Global Warming

327

will wind up doing just that (whether or not that is true of ride brokering in particular). But, on the other hand, it is difficult to determine how to measure the hourly wage equivalent of the earnings of an Uber driver (or other gig economy workers). Difficult as it may be to find a satisfactory regulatory solution to this challenge, if it is not addressed, it provides a loophole in the minimum wage through which you could someday drive a truck (or an Uber car). The gig economy shows promise for connecting buyers and sellers of services in new and mutually productive ways. At the same time, innovative organizations tend to disrupt regulatory regimes and worker protection laws that did not anticipate these new developments. To maintain, as some libertarians might, that merely because productive relationships are being organized “voluntarily,” public policy can dispense with the need to regulate firms and guard the interests of workers would be a mistake, in both economic and ethical terms. Rather, policy makers need to look to the underlying moral purposes and economic rationales that justified these existing structures of law and regulation (to the extent that were justifiable) and ensure that these purposes are well served under novel schemes of economic organization.

16.3  Environmental Protection and Global Warming The last set of questions concerns environmental protection and global warming. Problems of economic measurement loom large here, because the consequences of human activities for climate, landscape, plants and animals, and air and water all have to be valued in some way. In considering how environmental economists attempt to quantify these predicted consequences, we return to the problem of articulating the proper role of normative economics  –​of finding some way to make use of its analytical techniques and the opportunities for quantifying costs and benefits that it offers without allowing them to distort or overwhelm other ethical considerations. One might distinguish three different kinds of environmental problems: pollution of land, water, and air; depletion of natural resources, particularly farm land, fuels, fish, and forests; and preservation of species of plants and animals, places of natural beauty, and cultural landmarks. These three problems are not distinct from one another and are causally interrelated. One can also distinguish environmental problems by whether they are important mainly because of their consequences for people’s material interests or for other reasons. Preventing or lessening the effects of global

328

Economics and Ethics, Hand in Hand

warming is important mainly because of the grave harm global warming is likely to do to human beings; preventing the extinction of the Siberian tiger, in contrast, will not save human lives or increase GDP. Within the framework of mainstream normative economics, the treatment of those problems of pollution, depletion, and preservation that bear on people’s interests is apparently straightforward: one compares the extent to which various alternative policies satisfy preferences, using market data and willingness to pay (possibly corrected with equity weights) as measures of the extent to which preferences are satisfied. In doing so, however, one immediately confronts a serious ambiguity that arises because people disagree about the consequences of activities and policies for pollution, depletion, and preservation and about the consequences of these for health, income, and all the other things people value. Should one measure the extent to which global warming will affect welfare by imputing its costs from the amount people are willing to pay for increased fuel efficiency in their cars (even if this is based on faulty information), or should one look to science for a prediction of the consequences of global warming and elicit information on people’s willingness to pay to avoid those consequences? Although there are arguments on both sides, in our view the reasons in favor of satisfying preferences do not justify adhering to mistaken popular beliefs or to preferences that reflect mistaken beliefs (Hausman and McPherson 1994). Welfare economics should instead rely on the best supported estimates of the consequences of activities and policies and should attempt to screen out the effects of popular misapprehensions by eliciting preferences only for the consequences there is good reason to expect. Welfare economics provides indispensable tools for assessing different environmental policies, including those involving the mitigation of climate change. But the application of a welfarist framework for examining the significance of pollution, depletion, and preservation and for assessing possible policy responses has limits. As we have pointed out in earlier chapters, people may care about other things beyond maximizing aggregate welfare: for example, they may care about respecting human rights and obtaining restitution for past harms; they may want to ensure that the benefits and burdens of climate change are fairly shared. To address distributional concerns, it is important to make interpersonal comparisons and to provide distributional weights. The most sophisticated economic analyses of climate change thus attempt to incorporate some of these ethical considerations into their models and explicitly build in different weights. For example, some economic models incorporate prioritarian considerations such that the marginal value of social welfare diminishes with increasing

16.3  Environmental Protection and Global Warming

329

well-​being. Other models attempt to capture nonmarket effects of climate change (such as the possibility of changing cultural values due to a changing natural environment). We think that it is important that where relevant and feasible, ethical values are explicitly incorporated into economic models dealing with climate change. And, in cases in which such integration of ethics into economics is not possible, it may be helpful to supplement economic analysis with other types of information and reasoning. There are three central ethical concerns relevant to assessing policies about climate change:  justice, the proper treatment of catastrophic risks, and the value of the environment and of other living things. Dealing with climate change raises difficult questions of justice concerning how to distribute the costs within any given generation and how to distribute the costs and benefits across generations. Consider first problems of justice that arise within a single generation. In addition to questions about why people today should pay costs that benefit future generations, climate change involves differential costs and benefits for members of the same generation. Consider that not only is Bangladesh a poor country with limited capacity to mitigate the effects of climate change, but its low-​lying deltas and mudflats will be among the hardest-​hit places. Coping with climate change is expensive. For example, the large-​scale study led by Nicholas Stern (2007) concluded that GDP would need to fall 1  percent immediately  –​and for the long term  –​to address climate change. If that amount were shared equally, it would be equivalent to a reduction of $104 in consumption each year by every person on the planet. But if the amount were borne by the developed countries alone, each member of those countries would need to pay $667. When these numbers are compounded over a person’s lifetime, they call for very large sacrifices. Who should make these sacrifices? How should the benefits and burdens of climate change be distributed among people living today? Not only are the peoples of the world differently situated with respect to the costs of climate change and not only do individuals have different amounts of resources, but countries have different degrees of responsibility for the problem. How should we compare the responsibility of the developed nations, with their high, but slowly growing, levels of energy use, to that of the developing countries whose energy levels are quickly growing, although from a lower starting point? A variety of principles of justice have been proposed for dealing with questions concerning how to distribute the current costs and benefits:  a backward looking principle of “polluter pays,” a principle of equal per capita emissions rights, and an ability-​to-​pay principle are among the most

330

Economics and Ethics, Hand in Hand

debated (Singer 2002). All of these principles are controversial. And even when policy makers decide which principle or array of principles is best, they still need to decide how to weight a very disparate array of goods. While it is plausible to weight the benefits to the poor more heavily than those to the already well off, and to trade off various market goods against each other, how does one weigh the fact that some island nations will disappear unless fairly drastic steps are taken? If most people do not care about this (other than the small number of islanders themselves), does it therefore receive almost no weight? To answer this, one needs to decide whether people have any rights to live in the places they currently inhabit. This question cannot be answered by economics. Or consider: does it matter that climate change can lead to a loss of the “natural” environment? To answer this, the peoples of the world need to decide how much importance to attach to the fact that future people’s preferences will likely adapt to their changing circumstances. Not having experienced some of the deep pleasures that nature brings to people today, their descendants may not miss what they have never known. There is no way to avoid ethical issues when dealing with the distribution of the costs and benefits of mitigating and adapting to climate change. Deciding how to value future contingencies, cultural goods, and different rights that people have is complicated. In addition, the ethical problems posed by climate change do not only concern distribution within a generation: distribution between generations is also involved. How should we distribute benefits and burdens between generations? Some have argued that the fact that future generations will be richer than current generations means that the former will be in a better position to make sacrifices. The burdens should therefore fall more heavily on them, and the current generation is not obligated to take costly actions now such as those recommended by the Stern Report. There are two ethical issues at stake here in addition to technical questions about the relative costs of preventing and reversing climate change. The first ethical issue concerns the rate at which the costs and benefits of future consumption should be discounted. The second concerns the discount rate we should apply to the well-​being of people living in the future. These are not the same. Consider the discounting of goods and services. Economists regularly discount the value of consumption in the future –​money can be invested (so that something that will cost one thousand dollars in ten years requires putting away a good deal less than that today) and it is reasonable in many contexts to assume that a dollar today will be worth less to people in the future because the economy will grow and people will be richer. So we should discount future costs, even if there were no uncertainty.

16.3  Environmental Protection and Global Warming

331

The second discount rate is different. Here the question is, Should we weight the welfare of people living in the future as heavily as we weight the welfare of people today? In his 2007 report, Stern arrives at his demanding conclusions by applying a discount rate for the well-​being of future people of close to 0. That is, the welfare of people in the future is counted more or less equally to the welfare of people today. (The reason Stern did not propose a discount rate of exactly 0 is disquieting: he assumes that there is a 10 percent chance of extinction within the next one hundred years (2007, p. 47)!) Economists can disagree about both discount rates, but it is important to understand that in doing so they are disagreeing about very different things. Whether or not the economy will grow and people in the future will be better off than current people is an empirical matter. Stern, for example, thinks that climate change may make people in the future worse off than we are; Nordhaus disagrees with Stern’s projections and Nordhaus’s models paint a far rosier future scenario. In contrast, the issue about how to weight the well-​being of those in the future is an ethical one. Most philosophers have argued that it is unethical to assign a greater weight to the welfare of our own generation than to the welfare of others. They argue that ethics demands that we be neutral between generations and across time. Some economists disagree. Whatever one thinks about this issue, it does not turn on empirical facts. Nor does it turn on facts about what people prefer. There is another problem that arises when thinking about distributing the costs of climate change across generations. How we respond today to climate change affects not only the welfare of people in the future: it affects which people there are. If the existence of future people depends upon our failure to address climate change, to what extent can those future people be said to be harmed by our actions? After all, in the absence of our actions, these people would not have existed. Philosophers refer to this problem as the “Non-​Identity-​Problem.” There is a large philosophical literature on this question (Parfit 1984;, Broome 2012). Some have argued that people in the future cannot be harmed by people in the past since it is impossible meaningfully to compare their nonexistence with the well-​being they have, large or small. Others, more plausibly in our view, argue that the nonidentity problem pushes philosophers to move away from thinking about climate change in terms of harms to specific future individuals. Instead, they should think about states of affairs and social positions. People are harmed when their social positions or states of affairs fall below some threshold. On this conception, that a future person is wronged by an action today does not imply that the same

332

Economics and Ethics, Hand in Hand

person is worse off than she would have otherwise been at that time if what was done today had been different. Instead, it implies that she (or anyone in her position) is worse off than she ought to be. Obviously, such an account will need to determine the threshold level of well-​being beneath which no one should fall (see Hanser 1990). In addition to the problems of distributive justice, climate change raises moral issues because its consequences involve great uncertainty and are potentially catastrophic. The IPCC has put together a number of scenarios that represent a range of possible consequences of climate change. In some of these scenarios climate change produces some modest but manageable harms. In others climate change results in extreme harms. Applying a welfare analysis under conditions of uncertainty, where the probability of outcomes is not known, is particularly challenging. Even if one knew the likelihood of various outcomes (i.e., this were a case of risk rather than uncertainty) or at least could specify likelihoods within some probability range, applying a cost benefit or welfare analysis is complicated by the fact that some of the potential harms may be irreversible, and some may be catastrophic. How should a responsible policy maker take into account the possibility of catastrophic outcomes from climate change? One answer that has been proposed is the precautionary principle. It is not much of a “principle.” All it says that when faced with catastrophic risks, no matter how small, one should choose conservatively and take precautions. Policy makers should, if they can, avoid risks of serious harm unless they have safety mechanisms in place. Although vague, the precautionary principle appears very demanding, because it puts the burden on those who propose actions with serious risks. In its strong form, it would require costly regulation or intervention for a great many of the actions a government takes. An event with a one in a million chance of an extremely harmful outcome would require that precautionary countermeasures be taken. The precautionary principle has been incorporated into several international climate treaties, including the Rio Declaration and the Kyoto Protocol. Although widely adopted, the precautionary principle has been subjected to important criticisms. In the first place, it is clear that one cannot be precautionary about everything. Even regulation to address possible bad outcomes itself has possible bad outcomes (Sunstein 2005). Because all actions carry risks –​as does inaction –​the precautionary principle by itself does not offer any guidance about which actions should be taken. In its place, some have proposed a separate principle for dealing with catastrophic and irreversible events. If catastrophic climate change is a

16.3  Environmental Protection and Global Warming

333

possibility, as some have argued, it would make no sense to fail to take preventative, costly action: in that case, the welfare gains from taking such action could be extremely high. One of the debates in the context of climate change concerns whether or not the tail end of likely scenarios is fat or thin: that is, it concerns how likely a catastrophic scenario is to occur (Weitzman 2009). If a long chain of uncertainties with low probability nevertheless leads to catastrophic results, then tail events might be critical in assessing potential damages from climate change. More analysis is needed on the ethical dimensions of catastrophic risk. Other proposals have been offered for dealing with uncertainty rather than risk. In such situations, where probabilities cannot be assigned but catastrophes are possible, some philosophers have advocated adopting something akin to a maximin principle: one should adopt the policy with the least bad outcome. Such a policy would not in general maximize welfare: all it would do is to put a floor under the possible losses. The decision about how to respond to uncertainty and risk clearly has both economic and ethical dimensions. In addition to the values that climate change threatens with respect to humans, there is the question of how to assess its effects on nonhumans. Many features of our planet and of the creatures that live on it are magnificent in themselves, quite apart from the ways in which they may tangibly benefit people. Some people believe that it matters whether animals exist and whether they suffer or thrive, independently of whether people themselves benefit from their flourishing. Some people hold similar attitudes toward historical and cultural monuments and indeed toward cultures as a whole. How should such concerns be understood, and what influence should they have on policy making? These are evidently philosophical questions. Whatever the particular answers, if such things have value independently of their effects on humans, policy makers will need to find some way to measure them. This is obviously a challenging task. Environmental economists have attempted to quantify them in terms of willingness to pay and willingness to accept. If people think that it would be a bad thing if Siberian tigers became extinct or if a shopping mall were built where the Auschwitz crematoria stood, then presumably they would prefer states of the world in which Siberian tigers are not extinct and there are no shopping malls in Auschwitz. And it is surely the case that many people would be willing to pay something in order to preserve species, landscapes, monuments, or cultures. Since there are typically no markets for species, landscapes, monuments, or cultural preservation, one can treat the fact that aesthetic and principled views about the

334

Economics and Ethics, Hand in Hand

environment have little influence on market outcomes as if it involved an ordinary market failure. Cost–​benefit analysts can then use willingness-​ to-​pay information to value environmental preservation and thereby to respect consumer sovereignty. In this way one can attempt to quantify aesthetic and principled concerns to protect species, landscapes, watersheds, historical monuments, or cultures, and by quantifying these concerns one provides a way of trading them off against other values. However, it has been argued that this tactic has conceptual costs (Sagoff 2004). By a sort of definitional alchemy, non-​ self-​ interested concerns with these things have been transmuted into individual interests. Because principles influence preferences, which mainstream normative economists in turn identify with welfare, considerations of principle are treated as if they were individual interests. That many people who do not individually benefit from Siberian tigers would nevertheless be willing to pay to preserve them does not show that they individually benefit from the tigers all the same. Nor does their willingness to pay track the moral importance for them of the principles to which they appeal. Willingness to pay is at best an imperfect measure of individual benefit, and what is at stake with respect to preservation is typically not individual benefit.

16.4 Conclusions We hope to have shown that economics can contribute to the understanding and resolution of problems of human health, the new informal economy, and environmental destruction. Policy makers need to consider trade-​offs and that requires attending to the interests and values at stake. Economics and ethics can together contribute also to solving the hard problems of cooperation that beset efforts to implement solutions to these problems. Obviously there is plenty of work for others, too; and it would be absurd to place the salvation of the world in the hands of economists and moral philosophers. But we hope that the speculative words of this chapter heighten your appreciation of the ways in which economics and ethics can fruitfully inform each other. After working through the arguments and analytical perspectives reviewed in this volume, different readers will arrive at different judgments about practical policy options. We have not defended any single view of ethics, and we have not attempted to assign weights to the

Suggestions for Further Reading

335

differing considerations. Furthermore, policy conclusions obviously depend on judgments concerning the likely consequences of alternatives, and there are many disagreements concerning these matters. But we hope that our discussion encourages readers to reflect on this question:  Should one rest content with evaluating policies solely according to their consequences for satisfying people’s preferences (with perhaps a few adjustments to allow for equity concerns), or should one strive to include in one’s evaluation the consequences for freedom, equality, justice, solidarity, and quality-​of-​life indicators that extend beyond preference satisfaction? Although the former path may keep economics neater and perhaps in appearance more “scientific,” the latter will, we contend, make it more useful and defensible. As we have stressed repeatedly, and as our examples illustrate, the relevance of ethics is not limited to the context of evaluating policy. Ethics enters directly into the development of social welfare theory and into the modeling of strategic interactions as games. It is needed in order to formulate relevant questions for “positive” research. It is implicit in the standard theory of rationality that lies at the core of contemporary economics. It is implicit in theories of how labor markets work and of how social cooperation in general is possible. Economics is certainly not a wayward branch of moral philosophy, but neither is it unrelated to moral philosophy. One enriches both economics and moral philosophy by acknowledging their interdependence.

Suggestions for Further Reading There is a rich literature dealing with the ethical and/​or economic dimensions of important policy questions. On health care, a few places to begin are Daniels (1985, 2007) and Fuchs (2011). For a set of writings that deal with the changing nature of the workforce and the rise of “transactional” rather than organizational arrangements and their implications see Lemann (2012), Harris and Krueger (2015), Hall and Krueger (2015), and Dokko et al. (2015). Climate change has spawned a rich literature taking up the complex aspects involved in policy decisions. On the various proposals for dealing with the distribution of benefits and burdens see Singer (2002), Caney (2010), Chakravarty et al. (2009). Major sources on the nonidentity problem include Parfit (1984), Kavka (1982), Broome (2012, ch. 4), and Boonin (2014).

336

Economics and Ethics, Hand in Hand

Questions for Study and Discussion 1. Of the four prerequisites for addressing the problems discussed in this ­chapter –​recognition, will, knowledge, and cooperation –​which do you think is the main bottleneck for each of the problems? 2. The government can address many of the welfare needs that people have by simply giving them cash or vouchers (such as food stamps), which are like cash, except that they can only be used to purchase certain classes of goods or services. Why is health care different? Why cannot government address health care needs by issuing vouchers that can be used to purchase health care? 3. What is the problem of moral hazard, and why is there no really good solution? 4. What is the problem of adverse selection, and why is it impossible to solve it in a system where individuals freely choose whether to purchase health insurance? 5. Do you think that new forms of work as exemplified by Uber will lead to greater exploitation of workers, or do you think that it will improve the working life of lower-​paid workers and diminish inequality? 6. Have you ever used Uber or Lyft? What was your experience? What do you think of them? What special economic problems do they generate? What moral problems do they cause? What would a utilitarian, a libertarian, or a conventional welfare economist think of them? 7. Why do “gig economy” innovations such as Uber create a stronger case for paternalistic regulation? Do you think that the risks that these innovation pose to the “individual contractors” to which these innovations give consumers access can be met by nudges? 8. Why is it hard to use standard welfare economics to address the problems raised by environmental degradation and global warming? 9. Do you think that future welfare should be discounted, and, if so at what rate? How would you defend your view? 10. What do you see as the main obstacles to an effective response to global warming? What do you think the prospects are for overcoming those obstacles? How much of a contribution can individuals make by driving less, flying less, and turning down their heat?

APPENDIX

How Could Ethics Matter to Economics?

We hope in this book to have shown how knowing moral philosophy can help one to do economics better and how economics can contribute to ethics. The most persuasive way to make this case is the one we pursued in the main text: by describing important aspects of moral philosophy and showing their bearing on economics. Many economists are nevertheless inclined to deny that moral philosophy has anything to contribute to economics. Why? In this Appendix, we explore some of the arguments and reaffirm the central thesis of this book –​that ethics is relevant to economic inquiry. We shall assume that those economists who would deny that moral philosophy is relevant to economics distinguish positive economics, which addresses factual questions, from normative economics, which addresses evaluative questions, and that they recognize that moral philosophy is relevant to normative economics. But they deny that moral philosophy is relevant to positive economics. We respond to three key arguments in support of their position. Section A.1 considers the role of positive economics in policy making and argues that normative questions are almost always implicated in such work. The second section turns to the general question of what it means to say that positive economics is “value free” and then sketches what we call “the standard view” of the relation between positive and normative economics. Section A.3 comments briefly on the exaggerated distinctions sometimes drawn between positive and normative inquiries, criticizing in particular the view that evaluative claims cannot be rationally addressed. The fourth section revisits the standard view and discusses some problems it faces. Having argued that economic work presupposes ethical commitments, that normative inquiries are subject to reasons, and that the standard view is overly strong, we return to the question of whether positive economists have anything to learn from studying ethics. 337

338

How Could Ethics Matter to Economics?

A.1 Objection 1: Economists as Engineers There are economists who deny that ethics is relevant to economics and go on to repudiate normative economics as the bastard offspring of completely distinct inquiries that ought to be completely separate. And yet they do not deny that economics is relevant to policy making. It is relevant –​vitally relevant, they agree –​but in their view only in the way that civil engineering is. Owing to a need for electric power, policy makers might consider building a dam. Civil engineering does not say whether this is a worthy objective. Engineers instead provide information about how difficult it is to build dams in different locations, how much electricity the dams can generate, how much land they will flood, and so forth. Engineers thereby provide answers to some of the “What if…?” questions that policy makers need to answer when they are deciding whether or where to build a dam. In order to accomplish anything, people need knowledge of cause and effect, and engineering can provide such knowledge. The first objection to the view that ethics is relevant to economics thus asserts that the role of economics is exclusively to provide causal information. Ethics helps determine the ends that policy makers pursue, and it constrains the means they may employ. Positive economics clarifies the consequences of alternative policies. Both economics and ethics are crucial to policy making, but neither has anything to do with the other. The following very simple schema might help clarify this view. 1. Our policy ought to achieve goal G and satisfy constraints C1, …., Cn. 2. X satisfies C1, …., Cn and achieves G. 3. Thus our policy should be to do X. Premise 1 comes from ethics or political philosophy while premise 2 comes from positive economics and other relevant bodies of empirical knowledge. The conclusion is a moral judgment, which follows loosely from both the moral premise 1 and the technical premise 2. On the view of economists as engineers, premises 1 and 2 have nothing to do with each other, and ethics has nothing to contribute to economics. This schema is too simple. After all, X may have other desirable or undesirable features, and there may be better alternatives to X. To incorporate these considerations, one might suggest the following schema: 1. Economic policy should be governed by the complete moral or social ranking R. 2. The consequences of X rank higher in R than do the consequences of any other policy under serious consideration. 3. Thus economic policy should be to do X.

A.1  Objection 1: Economists as Engineers

339

The moral premise in this schema says a great deal more than the moral premise in the first schema. It evaluates all the consequences of the various alternative policies. For example, in the case of a proposed law prohibiting arbitrary firings, R would specify the moral importance of at least the following: the benefits and harms of prohibiting arbitrary firings, the harms from the intimidation of workers on the job, the harms of unemployment among the lowest strata of workers, the benefits or harms to overall productivity, and so forth. The economist’s job is then to determine what are the consequences of the alternative policies with respect to these morally relevant aspects. Once all the consequences are known, the policies can be ranked. Economists are seldom able to tell policy makers exactly what the consequences of policies will be, but they may instead be able to determine the probabilities of different outcomes. This information will then make it possible to rank policies via weighting the moral values of their outcomes by the probabilities that they will obtain. As in the simpler schema, the policy conclusion depends on both moral and technical premises that are independent of one another. The second schema, unlike the first, is not too simple. Its drawback is that it is too demanding. No existing moral system is refined enough to provide the needed first premise, and economics is not up to the challenge of providing the second. Real policy making relies instead on partial moral rankings of salient features of policies and their consequences. The moral considerations policy members invoke are influenced by positive information concerning what sort of consequences the policies under consideration have, while the positive study of their consequences is channeled by normative views of what kinds of consequences matter. The view of economists as pure engineers cannot be sustained. We recognize that this view does fit some economic theorizing  –​for example, work devoted to estimating how much revenue would result from changing income-​tax rates. But, even here, economists often need to understand (even if they do not share) the values to which policy makers are committed in order to understand what policy makers want to know and what questions to ask. Furthermore, applying the second schema to the activities economists undertake is often impossible; the political process rarely formulates its economic problems clearly. Fritz Machlup (1969) points out that when economists are called on to give “purely technical” advice about how to accomplish certain ends, they are rarely given purely technical problems. Consider the tasks of economists who are asked to guide the Greek government in managing its debt and restoring economic growth. Without knowledge of the political history, the prevailing values, and the moral and

340

How Could Ethics Matter to Economics?

political constraints on policies, they will not know how to proceed. What range of possibilities should they consider? What options are off the table? It is likely (though not inevitable) that the values to which economists are committed –​including their best judgment about what will be good for the Greek people, whose government they are advising –​will influence what alternatives they consider and what weights they place on the comparative advantages and costs. Even if, in principle, the Greek government could specify all the relevant objectives, constraints, and evaluative considerations, in practice economists will have to rely on a general grasp of ethics to fill in the gaps. Economists need to understand the moral values that shape the set of options to be considered. Evaluative commitments are also bound to be crucial to the choice of problems to investigate and often of the tools of investigation as well (Weber 1919). This is not a criticism of that research: there is no sensible alternative  –​all research must proceed from an evaluation of what it is important to know. It is certainly important that economists be aware of the moral commitments that structure their own research projects, if for no other reason than avoiding bias. These moral commitments may lead them to overlook some feasible alternatives or to overstress the significance of others. Mistakes about how best to respond to economic challenges such as the collapse of the U.S. housing market can diminish the lives of tens of millions. It does not take much moral knowledge to know that widespread financial panics are “important,” but it does take some. In deciding what to study and in thinking about how to apply economics to practical problems, economists must think about ethical matters. They need not do so systematically or self-​consciously, or well, but they cannot avoid ethics altogether. We suspect that engineers often have the same need to invoke values to clarify problems and objectives that economists have. In designing a superhighway, the civil engineer is unlikely to have available a social ranking of all the various combinations of aesthetic appeal, safety, durability, expense, ease of traffic flow, and so on, that combine to make a good highway. Which alternatives then should he consider? Which possible permutations of these parameters need not be presented? Why? It is an oft-​noted fact that much professional practice involves wrestling with situations in which the normative and the positive are deeply entangled.

A.2 Objection 2: Positive Economics Is Value Free: The Standard View Economists cannot avoid an entanglement with ethics if they want to understand the terms of policy debate, to help determine public policies, or

A.2  Objection 2: Positive Economics Is Value Free: The Standard View 341

to select problems for study. But many economists would insist that the role of ethics ends there: ethics can pose or shape the economist’s questions, but it cannot contribute to her answers. The second objection to the view that understanding ethics contributes to doing economics maintains that there is a positive economic science and that this science is “value free.” Let us call this position the standard view. Some further clarification is in order, as follows. (For an insightful alternative interpretation see Mongin 2006.) As mentioned in Chapter 1, most people can readily distinguish questions concerning empirical facts  –​what is  –​from questions concerning what ought or ought not to be the case. This is the so-​called fact–​value distinction. It is difficult to make this distinction precise. For example, in stating, “That was a kind thing to do,” is one describing a fact or morally appraising an action? And some philosophers argue that the fact–​value distinction breaks down altogether. But let us suppose that it is possible roughly to classify questions as either factual or evaluative. The standard view maintains that ethics is concerned with questions of value, while the sciences are concerned with questions of fact. Moreover, the standard view maintains that questions of fact and of value are in two ways independent: (1) although facts may contribute to answering evaluative questions, they cannot settle them, and (2) values have no role whatsoever in answering factual questions. In both ways, there is an inferential gap between facts and values. On the standard view, positive economics can be value free, and, moreover, it should be value free. To speak of a value-​free inquiry may be misleading, because it suggests that the conduct of the inquiry is value free. But the conduct of inquiry cannot possibly be value free. Inquiring involves action, and action is driven by values. As we have already seen, values influence choices of what to study. Values also influence choices of what methods to employ and consequently of what hypotheses to discard or to pursue. It is because of their moral commitment to the goals of science that economists resist “cooking” their data. It is because of their personal morality that economists do not shoot those who disagree with them. The standard view does not deny that values influence the conduct of inquiry. What is meant by a value-​free inquiry is instead (a) an inquiry into a question of fact or a question of logic in which (b) the answers are not influenced by any values except those that are part of scientific activity itself. The standard view recognizes that the results of positive inquiries can be relevant to policy, because those results can show that specific policies facilitate or frustrate the attainment of valued goals. But without some prior evaluative commitment, the findings of positive science settle no questions of policy or value. Thus the study of ideology and of the values of

How Could Ethics Matter to Economics?

342

Table 1.1.  Exaggerated Contrasts between Facts and Values Factual claims

Evaluative claims

Disagreements can be resolved by evidence Relatively little disagreement Descriptive: say how things are True or false Objective Independent of evaluative claims Bear on action, only given motives

No good way to resolve disagreements Relatively little agreement Prescriptive: say how things ought to be Not true or false Subjective Dependent on factual clams Guide action: Logically tied to motives

economists is irrelevant to understanding economics or economic methodology, though it may help one to understand the scientific failings or motivations of particular individuals. We will present some criticisms of the standard view after considering aspects of it more deeply.

A.3 Objection 3: Normative Questions Are Not Subject to Rational Evaluation The standard view draws a sharp distinction between factual claims and evaluative claims. Thus, for example, Milton Friedman maintains with respect to disagreements about fundamental values that “men can ultimately only fight” (1953, p. 5). Reprinted for the reader’s convenience are the exaggerated distinctions of Table 1.1. Sharply distinguishing between facts and values and between positive and normative science does not commit one to these distinctions. For example, the independence of facts and values does not imply that evaluative claims are merely subjective. The possibility of distinguishing between positive and normative economics does not imply the standard view that they should be pursued separately (Weston 1994). At issue in this Appendix is whether positive economics is completely value free, not whether rational arguments concerning values are possible, but even so, we do want to cast doubt on the distinctions listed above. The main point we insist on has, we hope, already been demonstrated in the chapters of this book. It is that normative questions –​questions concerning evaluative claims  –​are subject to rational discussion and resolution. Consider, for example, our discussion of a preference satisfaction theory of well-​being, the adequacy of utilitarianism, or the plausibility of the goal of welfare equality.

A.4  Answering the Standard View

343

One usually addresses such questions with rational arguments rather than experiments (which are, in any case, only rarely possible even in positive economics). Now it is true that rational argument is not sufficient to resolve all disagreements about value. Nevertheless, moral principles can be assessed by examining their implications, and specific judgments can be criticized on the bases of their factual presuppositions and the moral principles upon which they rely. The other contrast we want to contest is that between, on the one hand, the objectivity of factual claims and, on the other hand, the alleged subjectivity and arbitrariness of evaluative claims. Whether evaluative statements can be literally true or false is a difficult question that we will not attempt to answer. Prescriptions cannot be literally true or false, but not all evaluative claims are prescriptions (consider again the statement “that was a kind thing to do”). What is crucial, in our view, is that not all answers to evaluative or specifically ethical questions are equally good. Some are way off the mark, some rely on unreasonable assumptions, while there is good reason to regard others as in some relevant sense correct. If by “objective” one means that claims are correct or incorrect regardless of whether people believe that they are correct, then –​as our discussion shows –​many evaluative claims are objective. Whether a market in pollutants would be beneficial is not determined by whether people think it would be; slavery’s wrongness does not rest on popular opinion. The bottom line is that recognizing that there is an important difference between facts and values and between normative and positive inquiries does not require one to denigrate normative inquiry or to insist that the inquiries be isolated from one another.

A.4 Answering the Standard View: How Knowing Ethics Contributes to Positive Economics Thus far, we have examined some problematic exaggerations of the standard view but we have not criticized it directly. Before taking up this task, it is important to recognize how large a role for values the standard view potentially allows. It is compatible with the recognition that economists may need to understand the concepts and criteria that guide evaluation of economic outcomes and processes, and it can grant that ethics has an important part to play in economic life. Its proponents can even concede that it may sometimes be difficult to be a good economist without knowing some ethics. All the standard view maintains is that positive economics, considered as a body of knowledge, is wholly independent of ethics. We can therefore endorse the standard view without abandoning our project of showing

344

How Could Ethics Matter to Economics?

that ethics has much to contribute to economics. Just as it is important for economists to know some topology and statistics even though there are important parts of economics to which these fields contribute nothing, so it is important for economists to know some ethics even if is not relevant to everything economists do. We do not deny that there is good work in economics that is largely independent of any evaluative concepts and theories. Consider the research that goes into estimating demand elasticities. Yet, as numerous examples in this book show, positive economics is sometimes penetrated by ethical concerns. We showed, for example, that ethical commitments play an important role in discussions of overlapping generations models and of involuntary unemployment. These cases may not be typical. What about inquiries that are in the relevant sense value free? Is the standard view correct to maintain that studying ethics has nothing to contribute to those inquires? We think, on the contrary, that studying ethics can be useful even in circumstances in which the theorist’s own values do not come into play, because the phenomena that economists study are themselves saturated with values. Whether or not the study of economics is value free, the object of that study  –​economic behavior  –​is not. Ethical commitments are among the causal factors that influence people’s economic behavior, and hence they are among the factors with which economists need to be concerned. If people did not generally tell the truth and keep their promises, then economic life would grind to a halt. As theorists who study labor markets have noted, employees and employers have moral beliefs that affect the wage and employment bargains they make. People’s moral dispositions affect economic outcomes. This fact does not imply that economics cannot be value free. Although people’s beliefs and preferences, including their moral beliefs and preferences, obviously have economic consequences, so do facts about their physiology or about the terrain. The mere fact that moral beliefs and preferences are important causal factors does not show that economists need to pay more attention to morality than to biology or geology. However, the importance of knowing some ethics cannot be dismissed this easily, for two related reasons. First, ethical commitments, unlike physiological or geological facts, are not just givens: they themselves often depend on economic institutions and outcomes. Second, their content and this dependence on economic institutions and outcomes are hard to grasp without understanding a good deal about ethics. An example mentioned in Chapter 6 helps clarify these claims.

A.4  Answering the Standard View

345

In The Gift Relationship:  From Human Blood to Social Policy, Richard Titmuss compared different systems by which human blood is collected for the purposes of transfusion, and he argued that the existence of a market in human blood both undermines people’s willingness to supply blood and yields blood of inferior quality to that obtained by an altruistic system. In short, Titmuss argued that in this context the existence of a market causes an efficiency loss. In addition to the statistical evidence, based on comparing different systems of blood supply in the United States, England, and Japan, Titmuss pointed to statements people make about why they donated blood. Donors repeatedly said that they were giving “the gift of life.” Implicit in this discussion and explicit in Peter Singer’s later (1973) essay in defense of Titmuss is the thought that, when a pint of blood can be purchased for fifty dollars, donors may feel that instead of giving something priceless –​the gift of life itself –​they are giving the equivalent of fifty dollars. Hence people might be less willing to donate blood when blood can also be obtained commercially. The bad consequences –​higher costs, shortages, and more hepatitis –​can then be explained. Since fresh blood is perishable, supplies must be regular and blood cannot be stockpiled to accommodate fluctuations in demand. Shortages may result under a commercial system because (a)  people are less willing to donate blood and (b)  the amount of blood available from the small part of the population that is willing to sell blood increases and decreases irregularly. The higher costs are obvious. Finally, in commercial systems, unlike systems involving only voluntary donation, people have an incentive to conceal illnesses such as hepatitis. This case illustrates how people’s moral beliefs and preferences influence economic outcomes and also how economic arrangements (in this case, whether there is a market for human blood) can influence what people feel a moral obligation to do. We do not know whether the explanation offered by Titmuss for the dramatic differences between outcomes in the United States and Britain is correct, but there is no way to understand or assess it without attention to systems of moral beliefs, which (according to Titmuss) explain the choices of individuals. All of this is still arguably sociological rather than ethical knowledge. Economists do not need to make their own moral appraisal of systems of blood donation or of the values of blood donors. What matters are people’s behavioral dispositions and their beliefs about what is right and charitable, not what is “in fact” right or charitable. Just as economists can study the art market without appreciating art or the wine market without a cultivated palate (although they are likely to do a better job if they know something

346

How Could Ethics Matter to Economics?

about art and wine), so can they study the economic causes and consequences of moral beliefs and preferences without having any beliefs of their own concerning what is right or wrong. Positive economists studying blood distribution systems do not need to make moral judgments or to evaluate moral beliefs. All that matters is whether economic agents make those judgments and have those beliefs –​and what their causes and consequences are. But it does not follow that economists have little to gain from studying ethics. To see why, it is worth understanding why Titmuss’s conclusions were so puzzling to economists. While some of his data, such as the lower rate of infected blood in an altruistic system of procurement, make good economic sense, why should providing people with the mere possibility of selling their blood decrease total supply? In a generally sympathetic review, Arrow (1972) can make little sense of this. Part of the problem may be with Titmuss’s specific formulations, as when he says that markets “deprive men of their freedom to choose to give or not to give.” The quotation is puzzling because the option to donate blood still exists, and it might seem that markets in blood increase freedom by providing an additional alternative. Titmuss also seems to assume that altruism is virtually unlimited in this context. Perhaps we are better advised not to expend our limited altruism when it is not needed. As Adam Smith pointed out: But man has almost constant occasion for the help of his brethren, and it is in vain for him to expect it from their benevolence only … It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest. We address ourselves, not to their humanity but to their self-​love and never talk to them of our own necessities but of their advantages. (1776, pp. 26–​27)

Why should the blood donor be different from the butcher or the baker? An economist might claim that it makes sense to call upon limited supplies of altruism only when there is no less costly alternative (Robertson 1956). Apart from the specific problem that results from imperfections in the ability to test blood and the consequent incentive that commercial providers have to lie about their health, it would thus appear that markets in blood –​ like markets in bread, meat, or wine –​can only be for the good. Perhaps this conclusion is premature. Consider Table A.1 that simplified decision matrix sketching the consequences of donating blood in a voluntary or paid system. An economist who is sensitive to the possibility that economic agents care about other things than money and discomfort can apparently explain why individuals are more willing to donate blood when they are not paid

A.4  Answering the Standard View

347

Table A.1.  Donating vs. selling blood Unpaid Donate

Do not donate

Paid

Benefit to recipient Discomfort/​inconvenience Admiration by others and self

Benefit to recipient Discomfort/​inconvenience Lesser admiration by self and others Fee Avoid discomfort/​inconvenience Avoid discomfort/​inconvenience Social censure Lesser social censure

than when they are paid. When people are not paid, they are in effect able to purchase the admiration of others or to face their censure. Altruism does not enter into it. Indeed, it is hard to see how rational altruists (who, as altruists, do not care about the admiration or censure of others) could be more inclined to donate when there is no fee than when there is a fee, since they can always choose to forgo the fee. This diagnosis has testable implications: publicizing who gives blood should increase donations. In drawing this decision table, economists would be challenging the sociologist’s and philosopher’s interpretation and, quite properly, overstepping the bounds of purely positive analysis. This story challenges Titmuss’s view, and it leaves out the fact that donors think of themselves as altruistically giving the gift of life rather than as adhering to a social norm or investing in a reputation. If one takes their altruism at face value, it would seem that an economist would question whether it can be rational for donors to be more reluctant to donate blood when a fee is paid. Are they not making the same kind of cognitive mistake that Kahneman and Tversky identified when they found that people’s views of policies change (as in the Asian disease example) when the policies are “framed” as saving lives or allowing people to die? Would those who are less inclined to donate agree that they are making a mistake owing to how donation is framed, like those who are influenced by framing in Kahneman and Tversky’s example, or would they maintain that there is a real difference between donating blood in a voluntary as opposed to a paid system? To address these questions, economists need to know whether there is a relevant difference between donating blood in a paid system (even if one refuses to accept the payment) and donating blood in an unpaid system. Attempts to explain the phenomena need to attribute motives to individuals, and these motives depend on which differences among outcomes individuals regard as relevant, where moral commitments are among the

348

How Could Ethics Matter to Economics?

factors determining what is relevant. Thinking about motives and their moral appraisal in fact undermines the preceding explanation of donation in terms of seeking approval from others, because the private reputational benefits of donating blood are parasitic on donation’s being an altruistic action. If all donors were motivated by the pursuit of a reputation for self-​ sacrifice to benefit others, then donation would not confer that reputation. Obviously acts of altruism can be costly. It takes time to donate blood, and sometimes one feels sick or dizzy for a while. But altruistic acts carry rewards, too. People take pride in doing something they take to be decent and unselfish. They take pleasure in thinking of the good their blood may do for someone else. They enjoy the good opinion others may form of them (though as just pointed out, the reputational rewards depend upon the act of giving being independently valued). Having given blood once, an agent may be more rather than less likely to give again. If many of one’s neighbors donate, then one may be more rather than less likely to give, even though the need for one’s gift is reduced by the larger supply. Altruism is scarce but it is not in fixed supply, and in some instances its supply increases with its consumption (Hirschman 1985). To appraise and comprehend Titmuss’s explanation, one needs to understand factors such as these. And there is no way to do so if one has no understanding of the ways that moral considerations motivate human beings. Those who are not convinced that this example undermines the standard view might complain that we have shown only that economists sometimes need to know what ordinary people think about morality, not that economists themselves need to think about morality, let alone the sophisticated theories surveyed in this book. Part of this complaint is surely correct, but there is a great deal in ethics that bears on everyday moral thought. Thinking about morality contributes enough to the understanding of how ordinary people are motivated by moral considerations that it can be of great value to economists. There is a further consideration that is relevant to appraising the standard view: the moral convictions of economic agents, unlike causal factors such as rainfall, can be influenced by the way in which they are analyzed and described by economists. For example, learning economic theory may itself change people’s view of benefits and burdens and hence change their moral principles and conduct. Knowledge of economics may, for example, have contributed to the change in attitude toward charging interest on loans of money or to the decline of debtor’s prisons. Furthermore, generalizations about what people rationally do will often influence what people think ought to be done. Even though what

A.5  Conclusions: Ethics into Economics

349

ought to be done does not follow logically from what is done, it often follows psychologically. Modeling human behavior as if it were entirely self-​interested unavoidably legitimizes and fosters self-​interested behavior. Indeed, it seems that learning economics may make people more selfish (Marwell and Ames 1981; Frank, Gilovich, and Regan 1993; Frank 1996). As we argued in Chapter 5, the terms of economic explanations –​like Willie Sutton’s account of why he robbed banks –​can easily carry moral implications. The honesty, trust, and sense of fair play that help economies to function well are not givens that are fortunately abundant or unfortunately scarce. They are not facts of nature, comparable to geological formations or physical laws. They grow or wither depending on the institutions within which people live and the shared understandings they have of those institutions. Their content varies widely from individual to individual and from society to society. Economists can propose policies or make arguments that help to nurture –​or to undermine –​these vital moral resources.

A.5 Conclusions: Ethics into Economics We have argued that economists are not only social engineers contributing to policy in the way that civil engineers contribute technical information about dams and highways. Normative economics attempts to appraise policies, even if usually from a limited point of view, and evaluative thinking is in practice unavoidable in order to formulate well-​ defined questions for positive inquiry. The “standard view” of the relations between positive and normative economics does not deny any of this. It does not deny the existence or legitimacy of normative inquiry. Instead it insists on the possibility of purely positive inquiry and thus on the separability of positive and normative inquiry. Moreover, the standard view recognizes that the conduct of positive science is guided by the values of scientific inquiry and that the questions positive science addresses are influenced by a plethora of specific values, including policy interests. In insisting that positive economics be value free, what is meant is that the positive economist’s own evaluations of economic outcomes, institutions, and processes are not supposed to influence the answers given to the questions under investigation. The defender of the standard view also recognizes that economics does not always live up to the ideal of value-​free inquiry. We argued that thinking about ethics ought to play a larger role within economics than the standard view allows, because the moral commitments

350

How Could Ethics Matter to Economics?

of economic agents are important causes and effects of significant economic phenomena, and that these commitments are influenced by the way that they are described and appraised by economic agents and by economists. Although understanding the moral commitments of agents and the dependencies among them does not require that economists make ethical appraisals, an understanding of human actions requires a good grasp of ethics. As this book has shown, reflection on ethics has an important role to play in both positive and normative economics.

Questions for Study and Discussion 1. What is positive economics? What is normative economics? How do they differ? How are they related? 2. How could normative economics be an entirely positive enterprise? Why is this not a satisfactory view? 3. Could economics be just like civil engineering? Give your reasons. 4. What is the matter with this simple schema? 1. Our policy ought to achieve goal G and satisfy constraints C1, …., Cn. 2. X satisfies C1, …., Cn and achieves G. 3. Thus our policy should be to do X. 5. What is the matter with this alternative schema? 1. Our policy should be governed by the complete moral or social ranking R. 2. The consequences of X rank higher in R than do the consequences of any other policy under serious consideration. 3. Thus our policy should be to do X. 6. Suppose that an economist is asked to design a policy that limits the working hours of Uber drivers. Would economists need to think about values in order to design such a policy? Why or why not? 7. What is the standard view of positive science (or economics) and its relationship to normative science (economics)? What objections to the standard view does the Appendix raise? 8. What do you think of the distinctions drawn in Table 1.1? Are any of these distinctions unclear to you? Which of the distinctions would you defend? 9. What should one say about claims such as “Apple treats its employees well” or “Apple treats its employees badly”? Are they factual claims or evaluative claims? Does it matter?

Questions for Study and Discussion

351

10. Why is it more difficult to have a purely positive science of economics than to have a purely positive science of chemistry or astronomy? 11. What role do you think that values play in current debates about whether governments should do more to stimulate economies or whether they should aim to lower budget deficits? 12. Can you explain why people might be more likely to donate blood when they are not paid to do so than when they are paid? Do you think that you personally would be more likely to donate blood if blood donation were paid or unpaid? Give your reasons. Is there really any difficulty in offering a purely positive economic explanation for the facts about blood donation? Why should blood donors be different than butchers or bakers? 13. How does learning economics influence people’s moral beliefs? Should learning economics have these influences? 14. Can you think of any examples discussed in the book where positive economics is in need of guidance from ethics?

Glossary

adverse selection (Section 16.1) Adverse selection occurs when an offer to sell insurance especially attracts individuals who have reason to believe that they will make a claim. altruism (Sections 5.1, 5.3, and 5.5) Altruism consists of action that aims at the good of someone else. It is not the same as morality. Arrow’s theorem (Section 13.1) This theorem proves that there is no way of constructing social choices or social evaluations on the basis of individual preferences that satisfies a small number of apparently reasonable conditions. asymmetric information (Section 16.1) Asymmetric information refers to the state of affairs in which one party knows more about the goods or services to be exchanged than the other. Asymmetric information can lead to inefficiencies or unfair terms of trade and provides grounds for government regulation or other licensing arrangements. autonomy (Section 10.1) Autonomy can refer to positive liberty, to a capacity to govern oneself, or to a condition of self-​governing. backward induction (Section 14.4) Backward induction consists in reasoning backward from how players would choose at the end of a sequence of moves, to deduce the optimal strategy to employ. Backward induction shows that the uniquely rational strategy to follow in a finite iterated prisoner’s dilemma is to defect on every move. bargaining theory (Section 14.5) Bargaining theory is a branch of game theory in which individuals can improve upon an initial nonagreement point only if they can reach agreement on some distribution of resources. Bayesian theory (Section 4.2) Bayesians offer an interpretation of probability as degree of belief and support the use of expected utility theory in circumstances of uncertainty. capability (Section 8.7) A feature of a person that enables the person to achieve various beings and doings (“functionings”). Two examples of functionings are reading and sight. The corresponding capabilities are literacy and the ability to see. Amartya Sen and Martha Nussbaum have argued that justice, well-​being, and human development should be understood in terms of people’s capabilities and functionings. cardinal utility (Section 4.2) A utility representation in which the utilities of alternatives for an individual can be meaningfully added and subtracted.

353

354

Glossary

commitment (Section 5.2) In Amartya Sen’s view, a person P has a commitment to perform action A if and only if P would perform A in some circumstances even if doing so conflicted with A’s self-​interested and sympathetic preferences. completeness (see Section 4.1) A condition on preferences. A person’s preferences are complete if for all objects of preference x and y, either the person prefers x to y or y to x or the person is indifferent between x and y. consequentialism (Chapter 7) A structure of ethical theories in which the morality of actions and policies depends only upon their results. contractualism (Chapter 12) A range of views concerning justification and motivation for moral theories. The main idea of contractualism is that a moral theory is justified if it would be agreed to by rational individuals under the right circumstances. cost–​benefit analysis (Section 9.4) Cost–​benefit analysis consists of techniques to evaluate policies largely by measuring “willingness to pay”:  how much would those who benefit from policies be willing to pay to institute them and how much would those harmed by policies have to be compensated not to object to instituting them? Although the comparison of these benefits and costs may not by itself wholly decide policy questions, practitioners of cost–​benefit analysis believe that such information is highly relevant. deontological moral principles (Section 7.6) Deontological moral principles are nonconsequentialist principles like the Ten Commandments. In Samuel Scheffler’s terminology, deontological (nonconsequentialist) ethical theories employ both “agent-​centered prerogatives” (they sometimes permit agents to act in a way that does not maximize the good) and “agent-​centered constraints” (they sometimes prohibit agents from acting so as to maximize the good). effective freedom (Section 10.1) The effective freedom of individuals consists of their opportunities  –​the range of alternatives that are meaningfully open to them. Persons can have opportunities that are formally open to them, but have no effective ability to make use of those opportunities. For example, a person may have the formal opportunity to be a professor, but no effective freedom to do so because she is in a permanent coma. efficiency (Chapter 9) In theoretical economics, efficiency is either Pareto efficiency or Kaldor–​Hicks efficiency, which is essentially a form of cost–​benefit analysis. See Pareto efficiency. egalitarianism (Chapter  11) Egalitarianism is a position in political philosophy that maintains that, other things being equal, social arrangements are more just or morally better when there are fewer inequalities in resources, opportunities, well-​ being, status, or the relations among individuals. envy-​free allocations (Section 11.4) An allocation is envy free if nobody prefers anybody else’s allocation to his or her own. equity (Chapter  11) Economists use the term “equity” frequently as a synonym for “justice.” ethics Although some people treat ethics as more theoretical than morality, we take ethics and morality to be synonymous. See morality. expected utility (Section 4.2) The expected utility of an action is the utility of the possible outcomes multiplied by their probabilities. extended sympathy judgments (Section 7.2) An extended sympathy judgment is a preference ranking of “extended alternatives” such as (Jill, y) –​that is, being Jill and

Glossary

355

obtaining some alternative y. Extended sympathy judgments are supposed to provide a basis for making interpersonal comparisons of the extent to which people’s preferences are satisfied. See interpersonal comparisons. extensive form games (Section 14.1) Extensive form games are representations of the possible choice nodes and moves in a game in the form of an inverted tree with the payoffs at the terminal nodes. externality (Sections 2.3 and 6.3) An externality is a consequence of an agent’s action for which there is no market and hence no price. Pollution is a negative externality: when individuals drive their cars, they impose costs on other people, who receive no compensation. Externalities can be negative or positive. While any action can produce an externality if someone disapproves of it, in practice most economists interpret externalities in terms of more or less objective harms. fair allocations (Section 11.4) A fair allocation can be defined in many ways. One influential interpretation of such an allocation holds that it is both Pareto optimal and envy free. Different theories of justice have different implications for which distributions are fair. folk psychology (Chapter 4) Folk psychology is the view that people’s actions are to be explained in terms of their beliefs and their desires. The explanations economists offer of individual choice behavior conform to folk psychology. freedom (Section 10.1) Freedom can be viewed as a three-​place relation among (1) an agent, (2) obstacles of some kind, and (3) some of the agent’s objectives. An agent is free in some regard, when there are no obstacles of a particular kind preventing the agent from doing something. The main kinds of freedom concern (1) the absence of intentional interference from others, (2) the range of alternatives open to one, and (3) the extent to which one is autonomous or self-​determining. happiness (Section 8.6) Happiness is a contested notion. Some take it as a mood –​a quality of experience –​while others take it as an attitude or an appraisal of a life or of some portion of a life. There is an important and rich philosophical literature on whether happiness constitutes welfare. game theory (Chapter 14) Game theory is a mathematical theory concerning interdependencies with applications in several fields. It arose as an attempt to understand strategic interactions among individuals, and its main applications still lie in this domain. gig economy (Section 16.2) The gig economy refers to an organization of the provision of goods or services (but mainly services) whereby firms function as middlemen connecting consumers to individual contractors rather than as employers. Gini coefficient (Section 11.6) A measure of inequality ranging from 0 (perfect equality) to 1 (total inequality). A Lorenz curve plots the cumulative percentages of total income share against the cumulative number of recipients, starting with the poorest individual or household. The Gini coefficient is arrived at by measuring the area between the Lorenz curve and a hypothetical line of absolute equality, expressed as a percentage of the maximum area under the line. independence condition (Section 4.2, Section 4.3) A condition on preferences. An agent’s preferences satisfy the independence condition if the agent’s preferences between two lotteries that differ in only one prize match the agent’s preferences between the differing prizes.

356

Glossary

independence of irrelevant alternatives (Section 13.1) Arrow’s independence of irrelevant alternatives condition says that the social ranking of x and y should depend on nothing except individual rankings of x and y. interpersonal comparisons (Section 7.2) To judge who is worse off or how unequal welfare may be or to consider meaningfully whether policies increase or decrease total well-​being, one needs to be able to make an interpersonal comparison of the level or of the change in well-​being in different people. justice (Chapter 12) Most philosophers take justice to be the most important part of morality. There are many different conceptions of justice, which vary in terms of what rights ought to be respected, what people deserve, and what it means to treat people fairly. libertarian paternalism (Section 10.7) Libertarian paternalism consists in actions designed to benefit individuals by influencing their choices without the use of sanctions by exploiting or counteracting rational foibles that can lead people to make choices that are bad for themselves. libertarianism (Section 10.6) A kind of moral theory that takes freedom to be the most important political value and takes justice to be a matter of protecting individual freedom and respecting individual rights. minimal benevolence (Section 9.1) Minimal benevolence is the thesis that other things being equal, it is morally good that people be better off. moral hazard (Section 16.1) Moral hazard is a side effect of insurance whereby those who are insured take fewer steps to avoid outcomes that generate claims. moral norms (Sections 5.4 and 5.5) Moral norms are distinguished from other norms by their subject matter (interpersonal interaction in which significant benefits and harms are at stake), their weight (they typically override other considerations), and the sanctions (both internal  –​guilt  –​and external  –​blame  –​attach to their violation). morality Morality consists of judgments concerning what things are intrinsically good for people and what actions are permissible, impermissible, and obligatory, and concerning the moral worth of people and their characters. Nash equilibrium (Section 14.1) A Nash equilibrium is a set of strategies in a game that are “best replies” to one another. More precisely, given the strategies of the other players, no player prefers some other strategy to the one he or she has chosen. negative liberty (Section 10.1) Negative liberty is the freedom of individuals from constraints that are intentionally imposed by other people or institutions to do as the individuals choose. normal form games (Section 14.1) Normal form games are tabular representations of the payoffs resulting from all sets of pure strategies in a game. normative vs. positive theories (Sections A.1 and A.2) Positive theories say what is, while normative theories say what ought to be. Positive theories are concerned with facts, while normative theories are concerned with values. As we discuss in this book, the distinction is, however, not easy to draw sharply. norms (Sections 5.4 and 5.5) Norms are prescriptive rules regarding behavior that are shared among a group of people and that are partly sustained by the approval and disapproval of others. nudges (Section 10.7) Nudges are actions that are designed to influence individuals’ behavior, often (though not exclusively) with the objective of benefiting individuals,

Glossary

357

through exploiting common cognitive flaws. This definition is contested, with some authors taking them to be any action that influences choices without imposing significant costs on the chooser. ordinal utility (Section 4.1) An ordinal utility function represents the order of an individual’s preferences in terms of what is more and less preferred but provides no other information about the preferences. paradox of the Paretian libertarian (Section 13.5) Amartya Sen’s paradox of the Paretian libertarian is the proof of an inconsistency between the Pareto principle and a principle of minimal liberty (given some other technical conditions). Pareto improvement (Sections 9.1 and 9.2) A state of affairs S is a Pareto improvement over a state of affairs R if nobody prefers R to S and somebody prefers S to R. Pareto optimality (or Pareto efficiency) (Section 9.2) A state of affairs S is Pareto optimal or Pareto efficient if no states of affairs are Pareto improvements over S –​in other words a state of affairs in which it is not possible to satisfy anyone’s preferences better without frustrating someone else’s preferences. Pareto principle (Chapter 13) The strong Pareto principle says that if nobody prefers x to y and somebody prefers y to x, then x ought to be socially preferred to y. The weak Pareto principle says that if everybody prefers x to y, then x ought to be socially preferred to y. perfect competition (Section 9.1) Perfect competition obtains when there are markets for all goods and services (and thus no externalities), no barriers to entry or exit from any market, and so many traders in every market that no one can influence prices. Perfect competition is also sometimes taken to imply there are no interdependencies among people’s utility functions, Pigou–​Dalton principle (Section 13.5) The Pigou–​Dalton principle holds that a transformation of some distribution that transfers some good from those who had (and continue to have) more to those who have less is, other things being equal, morally preferrable. positive affine transformation (Section 4.2) f ′(x) is a positive affine transformation of f(x) if and only if f ′(x) = af(x) + b, where a and b are real numbers and a > 0. positive liberty (Section 10.1) Positive liberty is rational self-​determination or, in one sense of the term, autonomy. positive vs. normative theories See normative vs. positive theories. potential Pareto improvement (Section 9.4) State S is a potential Pareto improvement or an increase in Kaldor–​Hicks efficiency over state R if it would be possible for those whose preferences are better satisfied in S than in R to compensate those whose preferences are less well satisfied and thus bring about a Pareto improvement. prioritarianism (Section 11.2) Prioritarianism is the view that the interests of those who are worse off should be weighted more heavily than the interests of those who are better off. primary goods (Section 12.2) Primary goods are important building blocks in Rawls’s theory of justice. They are all-​purpose means, things such as income and education that contribute to or do not hinder everyone’s life plans. Rawls argues that primary goods are a better basis for the evaluation of alternative principles of justice than are utilities. prisoners’ dilemma (Section 14.2) A prisoners’ dilemma is a game with a particular pattern of payoffs. All the players have a single dominant strategy, and the result

Glossary

358

of each employing that dominant strategy is suboptimal. The following is a version of a two-​person prisoners’ dilemma, with B a dominant strategy for both J and K: J’s choice K’s choice

A B

A Second best for both Best for K; worst for J

B Best for J; worst for K Third best for both

probability (Section 4.2) Probabilities are real numbers between 0 and 1 that are attached to states of affairs and satisfy the axioms of the mathematical theory of probability. There are currently two main interpretations. “Objective probabilities” may be regarded as the limit of the relative frequency of an occurrence, as 1/​2 is the limit of the frequency with which an unbiased coin will land heads. “Subjective probabilities” are degrees of belief. rationality (Part I) In its loosest sense, rationality is appropriate sensitivity to reasons. Economists define rational preferences as preferences that are complete and transitive and rational choice as choice that it determined by rational preferences. reasons (Chapter 4) Reasons are propositions or the contents of mental states that justify actions or other mental states. republican liberty (Section 10.1) An agent possesses republican liberty to the extent that her negative liberty is secured by appropriate social conditions including rights, laws, and norms. The agent’s freedom is not merely at the pleasure of some other agent. revealed-​preference theory (Section 4.1) The theory of revealed preference attempts to construct preferences out of choice behavior that satisfies consistency conditions. It aims to avoid reference to preference as a subjective state responsible for choice. rights (Sections 10.2–​10.5) Rights involve freedoms for the rights holder and various duties for others. Legal and conventional rights are established by law and convention. To say that agents have a moral or natural right to X implies that the agent ought to have either a legal or a conventional right to X. risk (Section 4.2) One is in a circumstance of risk when the outcome of action is not certain, but all the possible outcomes and their probabilities are known. self-​interest (Sections 4.1, 4.3, 5.2)Self-​interested behavior is behavior that is directed toward what the agent believes to be beneficial to himself or herself. Actions that are governed by people’s preferences are self-​interested only if the preferences themselves are self-​interested. social choice theory (Chapter 13) Social choice theory is the proof and interpretation of theorems concerning the social aggregation of preferences, judgments, and interests. social welfare function (Section 13.1) A social welfare function is a ranking of social states in terms of welfare.

Glossary

359

transitivity (Section 4.1) A condition on preferences. An agent’s preferences are transitive if the agent prefers x to z whenever the agent prefers x to y and y to z for all objects of preference x, y, and z, and similarly for indifference. uncertainty (Section 4.2) One is in a circumstance of uncertainty when some of the possible outcomes of an action or their probabilities are not known. One is usually in circumstances of uncertainty. universal domain condition (Section 13.1) The universal domain condition says that a method of social evaluation or social decision making should work for every profile of individual preferences. It is common in social choice theories, including Arrow’s theorem and Sen’s paradox of the Paretian libertarian. utilitarianism (Chapter 7) Utilitarianism is the best-​known variety of consequentialism. According to utilitarianism, an action or policy is right if and only if it results in no less happiness or no less preference satisfaction than any alternative. utility (Section 4.1) Utility is simply a mathematical device for representing features of preference orderings. It is an index or indicator, not a subjective state like pleasure. Utility is not itself an object of preference or choice. voluntary choice (Section 3.1) For our purposes, a choice is voluntary if and only if (a) it is not forced by a violation of negative liberty and (b) there is at least one acceptable alternative. welfare (Section 5.2, Chapter 8) A person’s welfare is given by how good conditions are for him or her. Economists often measure welfare by the satisfaction of preferences, although this identification is controversial. welfarism (Chapters  7–​9) Welfarism consists in the evaluation of outcomes, institutions, actions, and policies in terms of their effect on individual welfare. well-​being (Section 5.2, Chapter 8) We take well-​being to be synonymous with welfare.

References

Ackerman, Bruce. Social Justice in the Liberal State. New Haven, CT: Yale University Press, 1981. Ackerman, Bruce and Anne Alstott. The Stakeholder Society. New Haven, CT: Yale University Press, 1999. Adler, Matthew. Well-​ Being and Fair Distribution. New  York:  Oxford University Press, 2012.   “Value and Cost-​Benefit Analysis,” in Iwao Hirose and Jonas Olson, eds. The Oxford Handbook of Value Theory. New York: Oxford University Press, 2015. Admati, Anat and Martin Hellwig. The Bankers’ New Clothes:  What’s Wrong with Banking and What to Do about It. Princeton, NJ: Princeton University Press. 2013. Akerlof, George. “Labor Contracts as Partial Gift Exchange.” Quarterly Journal of Economics 84 (1982): 488–​500. Akerlof, George. “Gift Exchange and Efficiency Wage Theory: Four Views.” American Economic Review Proceedings 74 (1984): 79–​83. Akerlof, George and Janet Yellen. “Introduction,” in George Akerlof and Janet Yellen, eds. Efficiency Wage Models of the Labor Market. New York: Cambridge University Press, 1986, pp. 1–​21. Alexander, Sidney. “Social Evaluation through Notional Choice.” Quarterly Journal of Economics 88 (1974): 597–​624. Allais, Maurice. “The Foundations of a Positive Theory of Choice involving Risk and a Criticism of the Postulates and Axioms of the American School” (1952), in Maurice Allais and Otto Hagen, eds. Expected Utility Hypotheses and the Allais Paradox. Dordrecht: Reidel, 1979, pp. 27–​145. Anand, Paul, Graham Hunter, Ian Carter, Keith Dowding, Francesco Guala and Martin Van Hees. “The Development of Capability Indicators.” Journal of Human Development and Capabilities 10 (2009): 125–​52. Anderson, Elizabeth. Value in Ethics and Economics. Cambridge, MA: Harvard University Press, 1993.   “What Is the Point of Equality?” Ethics 109 (1999): 287–​337. Aquinas, Thomas. On Law, Morality and Politics. Indianapolis: Hackett, 2003. Arneson, Richard. “Equality and Equal Opportunity for Welfare.” Philosophical Studies 56 (1989): 77–​93.

361

362

References

  “Liberalism, Distributive Subjectivism, and Equal Opportunity for Welfare.” Philosophy & Public Affairs 19 (1990): 158–​94.   “Luck Egalitarianism and Prioritarianism.” Ethics 110 (2000): 339–​49. Arrow, Kenneth. Social Choice and Individual Values. New  York:  Wiley, 1951 (2nd ed., 1963).   “Values and Collective Decision Making,” (1967). Repr. in Frank Hahn and Martin Hollis, eds. Philosophy and Economic Theory. New York: Oxford University Press, 1979, pp. 110–​26.   “Gifts and Exchanges.” Philosophy & Public Affairs 1 (1972): 343–​62.   “Some Ordinalist-​Utilitarian Notes on Rawls’ Theory of Justice.” Journal of Philosophy 70 (1973): 246–​63.   The Limits of Organization. New York: Norton, 1974.   “Extended Sympathy and the Possibility of Social Choice.” Philosophia 7 (1978): 223–​37.   “A Note on Freedom and Flexibility,” in Kaushik Basu, Prasanta Pattanaik, and Kotaro Suzumura, eds. Choice, Welfare and Development: A Festschrift in Honour of Amartya K. Sen. Oxford: Clarendon Press, 1995, pp. 7–​16. Arrow, Kenneth, Amartya Sen, and Kotaro Suzumura, eds. Social Choice Re-​ Examined:  Proceedings of the IEA Conference Held at Schloss Herstein, Berndorf, Near Vienna, Austria, vol. 1. New York: St. Martin’s Press 1996; vol. 2, 1997. Arrow, Kenneth and Amartya Sen, eds. Handbook of Social Choice and Welfare, vol. 1. Amsterdam: North Holland, 2002. Arrow, Kenneth and Amartya Sen, eds. Handbook of Social Choice and Welfare, vol. 2. Amsterdam: North Holland, 2010. Atkinson, Anthony B. “On the Measurement of Inequality.” Journal of Economic Theory 2 (1970): 244–​63.   Inequality: What Can Be Done? Cambridge, MA: Harvard University Press, 2011. Autor, David H. “Skills, Education, and the Rise of Earnings Inequality among the ‘Other 99 Percent.’’’ Science 344 (2014): 843–​50. Axelrod, Robert. The Evolution of Cooperation. New York: Basic Books, 1984. Bacharach, Michael. Beyond Individual Choice:  Teams and Frames in Game Theory. Edited by Natalie Gold and Robert Sugden. Princeton, NJ:  Princeton University Press, 2005. Baker, C. “The Ideology of the Economic Analysis of Law.” Philosophy and Public Affairs 5 (1975): 3–​48. Barbera, Salvador, Peter Hammond, and Christian Seidl, eds. Handbook of Utility Theory. Dordrecht: Kluwer, 1999. Barry, Brian. The Liberal Theory of Justice. Oxford: Clarendon Press, 1973.   Treatise on Social Justice, vol. 1. Theories of Justice. Berkeley: University of California Press, 1989.   Why Social Justice Matters. London: Polity Press, 2005. Bartels, Larry. Unequal Democracy:  The Political Economy of the New Gilded Age. Princeton, NJ: Princeton University Press, 2008. Batson, C. Daniel. “Experimental Tests for the Existence of Altruism,” in David Hull, Micky Forbes, and Kathleen Okruklik, eds. PSA 1992, vol. 2. East Lansing, MI: Philosophy of Science Association, 1993, pp. 69–​78. Baumol, William. Superfairness:  Applications and Theory. Cambridge, MA:  MIT Press, 1986.

References

363

Becker, Gary. The Economic Approach to Human Behavior. Chicago:  University of Chicago Press, 1976.   A Treatise on the Family. Cambridge, MA: Harvard University Press, 1981. Ben-​ Ner, Avner and Louis Putterman, eds. Economics, Values, and Organization. Cambridge: Cambridge University Press, 1998. Bentham, Jeremy. An Introduction to the Principles of Morals and Legislation. (1789). Edited by W. Harrison. Oxford: Basil Blackwell, 1967. Berg, J., J. Dickhaut, and J. O’Brien. “Preference Reversal and Arbitrage,” in Vernon Smith, ed. Research in Experimental Economics, vol. 3. Greenwich: JAI Press, 1985, pp. 31–​72. Bergson, Abram. “A Reformulation of Certain Aspects of Welfare Economics.” Quarterly Journal of Economics 52 (1938): 30–​4. Berlin, Isaah. “Two Concepts of Liberty,” in Four Essays on Liberty. Oxford:  Oxford University Press, 1969, pp. 118–​72. Bicchieri, Cristina. “Self-​ Refuting Theories of Strategic Interaction:  A  Paradox of Common Knowledge.” Erkenntnis 29 (1988): 69–​85.   “Paradoxes of Rationality.” Midwest Studies in Philosophy 15 (1990): 65–​79.   “Covenants without Swords: Group Identity, Norms and Communication in Social Dilemmas,” Rationality and Society 14 (2002): 187–​222.   The Grammar of Society:  The Nature and Dynamics of Social Norms. Cambridge: Cambridge University Press, 2006. Binmore, Ken. “Modeling Rational Players:  Part I.” Economics and Philosophy 3 (1987): 179–​214.   “Modeling Rational Players: Part II.” Economics and Philosophy 4 (1988): 9–​56.   Fun and Games. New York: D. C. Heath, 1992.   Playing Fair: Game Theory and the Social Contract. Cambridge MA: MIT Press, 1994.   Just Playing. Game Theory and the Social Contract. Vol. II. Economic Learning and Social Evolution. Cambridge, MA: MIT Press, 1998.   Natural Justice. Oxford: Oxford University Press, 2005. Blackorby, Charles and David Donaldson. “Cash versus Kind, Self-​ Selection, and Efficient Transfers.” American Economic Review 78 (1988): 691–​700.   “The Case against the Use of the Sum of Compensating Variations in Cost-​Benefit Analysis.” Canadian Journal of Economics 23 (1990): 471–​94. Blinder, Alan S. After the Music Stopped: The Financial Crisis, the Response, and the Work Ahead. New York: The Penguin Press (2013). Blinder, Alan S. and Don Choi. “A Shred of Evidence on Theories of Wage Stickiness.” Quarterly Journal of Economics 105 (1990): 1003–​15. Block, Walter. “A Free Market in Roads,” in Machan (1982), pp. 164–​83. Blount, Sally. 1995. “When Social Outcomes Aren’t Fair: The Effect of Causal Attributions on Preferences,” Organizational Behavior and Human Decision Processes 63: 131–​44. Boadway, Robin. “Cost-​Benefit Analysis,” in Matthew Adler and Marc Fleurbaey, eds. Oxford Handbook of Well-​Being and Public Policy. New  York:  Oxford University Press, 2016, pp. 47–​81. Bonnano, Giacomo. “The Logic of Rational Play in Games of Perfect Information.” Economics and Philosophy 7 (1991): 37–​65. Boonin, David. The Non-​Identity Problem and the Ethics of Future People. Oxford: Oxford University Press, 2014.

364

References

Bossert W., P. K Pattanaik, and Y. Xu “Similarity of Options and the Measurement of Diversity.” Journal of Theoretical Politics 15 (2003): 405–​21. Boulding, Kenneth. The Economy of Love and Fear:  A  Preface to Grants Economics. Belmont, CA: Wadsworth, 1978. Bowen, William and Michael McPherson. Lesson Plan:  An Agenda for Change in US Higher Education. Princeton, NJ: Princeton University Press, 2016. Bowles, Samuel. “The Production Process in a Competitive Economy: Walrasian, Neo-​ Hobbesian and Marxian Models.” American Economic Review 75 (1985): 16–​36. Bowles, Samuel and Herbert Gintis. “A Political and Economic Case for the Democratic Enterprise.” Economics and Philosophy 9 (1993): 75–​100. Bowles, Samuel, Herbert Gintis, Erik Olin Wright, and Harry Brighouse, eds. Recasting Egalitarianism: New Rules for Communities, States and Markets. London: Verso, 1999. Braithwaite, Richard. The Theory of Games as a Tool for the Moral Philosopher. Cambridge: Cambridge University Press, 1955. Brandt, Richard. A Theory of the Right and the Good. Oxford: Oxford University Press, 1979. Braybrooke, David. Meeting Needs. Princeton, NJ: Princeton University Press, 1987. Brennan, Geoffrey and James Buchanan. The Reason of Rules:  Constitutional Political Economy. New York: Cambridge University Press, 1985. Brennan, Geoffrey and Phillip Pettit. The Economy of Esteem:  An Essay on Civil and Political Society. Oxford: Oxford University Press, 2004. Brennan, Timothy. “A Methodological Assessment of Multiple Utility Frameworks.” Economics and Philosophy 5 (1989): 189–​208. Brighouse, Harry. School Choice and Social Justice. Oxford and New  York:  Oxford University Press, 2000.   Justice. London: Polity Press, 2004. Brighouse, Harry and Adam Swift. Family Values: The Ethics of Parent-​Child Relationships. Princeton, NJ: Princeton University Press, 2014. Brock, Dan. “Ethical Issues in the Use of Cost-​Effectiveness Analysis for the Prioritization of Health Care Resources,” in E. Edejer, R. Baltussen, T. Adam, R. Hutubessy, A. Acharya, E. Evans, and C. Murray, eds. WHO Guide to Cost-​Effectiveness Analysis. Geneva: World Health Organization, 2003, pp. 289–​311. Brock, Gillian, and Harry Brighouse, eds. The Political Theory of Cosmopolitanism. Cambridge: Cambridge University Press, 2005. Broome, John. “Should Social Preferences Be Consistent?” Economics and Philosophy 5 (1989): 7–​18.  “Utility.” Economics and Philosophy 7 (1991a): 1–​12.   Weighing Goods. Oxford: Basil Blackwell, 1991b.   Counting the Cost of Global Warming. Isle of Stroud: White Horse Press, 1992.   Weighing Lives. Oxford: Oxford University Press, 2004.   Climate Matters: Ethics in a Warming World. New York: Norton, 2012. Buchanan, Allen E. Ethics, Efficiency, and the Market. Totowa, NJ:  Rowman & Allanheld, 1985. Buchanan, James. The Limits of Liberty:  Between Anarchy and Leviathan. Chicago: University of Chicago Press, 1975. Calabresi, Guido and Douglas Melamed. “Property Rules, Liability Rules, and Inalienability: One View of the Cathedral.” Harvard Law Review 85 (1972): 1089–​1128.

References

365

Camerer, Colin, George Lowenstein, and Matthew Rabin. Advances in Behavioral Economics. New York: Russell Sage Foundation, 2004. Camerer, Colin, Samuel Issacharoff, George Loewenstein, Ted O’Donoghue, and Matthew Rabin. “Regulation for Conservatives: Behavioral Economics and the Case for Asymmetric Paternalism.” University of Pennsylvania Law Review 151 (2003): 1211–​54. Camerer, Colin and Ernst Fehr. “When Does “Economic Man” Dominate Social Behavior?” Science 311 (2006): 47–​52. Caney, Simon. “Climate Change and the Duties of the Advantaged.” Critical Review of International Social and Political Philosophy 13 (2010): 203–​228. Carter, Ian. A Measure of Freedom. Oxford: Oxford University Press, 1999. Chakravarty, Shoibal et  al. “Sharing Global CO2 Emission Reductions among One Billion High Emitters.” Proceedings of the National Academy of Sciences 106 (2009): 1184–​88. Chetty, Raj, Nathaniel Hendren, and Lawrence F. Katz. “The Effects of Exposure to Better Neighborhoods on Children:  New Evidence from the Moving to Opportunity Experiment.” American Economic Review 106 (2016): 855–​902. Christman, John, ed. The Inner Citadel: Essays on Individual Autonomy. New York: Oxford University Press, 1989. Chu, Y. and R. Chu. “The Subsidence of Preference Reversals in Simplified and Marketlike Experimental Settings: A Note.” American Economic Review 80 (1990): 902–​11. Chubb, John E. and Terry M. Moe. Politics, Markets and America’s Schools. Washington, DC: Brookings Institution Press, 1990. Churchland, Paul M. A Neurocomputational Perspective:  The Nature of Mind and the Structure of Science. Cambridge, MA: MIT Press, 1989. Clayton, Matthew and Andrew Williams. The Ideal of Equality. New York: St. Martin’s Press, 2000. Coase, Ronald. “The Problem of Social Cost.” Journal of Law and Economics 3 (1960): 1–​30. Cohen, G.  A. “Functional Explanation, Consequence Explanation, and Marxism.” Inquiry 25 (1982): 27–​56.   “On the Currency of Egalitarian Justice” Ethics 99 (1989): 906–​44.   “Equality of What? On Welfare, Goods, and Capabilities,” in Nussbaum and Sen (1993), pp. 9–​30.   Rescuing Justice and Equality. Cambridge, MA: Harvard University Press, 2009. Coleman, James. Foundations of Social Theory. Cambridge, MA:  Harvard University Press, 1990. Coleman, Jules. “Economics and the Law: A Critical Review of the Foundations of the Economic Approach to Law.” Ethics 94 (1984): 649–​79. Coleman, Jules L. and Jody Kraus. “Rethinking the Theory of Legal Rights.” Yale Law Review 95 (1986):  1335–​71; rpt. of Markets, Morals and the Law. Cambridge: Cambridge University Press, 1988, pp. 28–​63 Collard, David. Altruism and Economy: A Study in Non-​Selfish Economics. New York: Oxford University Press, 1978. Cooter, Robert. “Rawls’s Lexical Orderings Are Good Economics.” Economics and Philosophy 5 (1989): 47–​54.

366

References

Cowen, Tyler, ed. The Theory of Market Failure:  A  Critical Examination. Fairfax, VA: George Mason University Press, 1988. Daniels, Norman. Reading Rawls. New York: Basic Books, 1976.   Just Health Care. Cambridge: Cambridge University Press, 1985.   Just Health. Cambridge: Cambridge University Press, 2007. Dasgupta, Partha. An Inquiry into Well-​Being and Destitution. Oxford:  Clarendon Press, 1995. d’Aspremont, Claude and Louis Gevers. “Equity and the Informational Basis of Collective Choice.” Review of Economic Studies 44 (1977): 199–​209. Davidson, Donald. “Actions, Reasons, and Causes.” Journal of Philosophy 60 (1963); rpt. in Davidson (1980), pp. 3–​20.   Essays on Actions and Events. New York: Oxford University Press, 1980.   “Judging Interpersonal Interests,” in Jon Elster and Aanund Hylland, eds. Foundations of Social Choice Theory. New York: Cambridge University Press (1986), pp. 195–​211. Dawes, Robyn, Alphons van de Kragt, and John Orbell. “Cooperation for the Benefit of Us–​Not Me, or My Conscience,” in Jane Mansbridge, ed. Beyond Self-​Interest. Chicago: University of Chicago Press, 1990, pp. 97–​111. Deaton, Angus. “What Does the Empirical Evidence Tell Us about the Injustice of Health Inequalities?” In Nir Eyal, Samia Hurst, Ole Norheim, and Daniel Wikler, eds. Inequalities in Health:  Concepts, Measures, and Ethics. New  York, Oxford University Press, 2013, pp. 263–​81. Debreu, Gerard. Theory of Value. New York: Wiley, 1959. Demsetz, Harold. “The Exchange and Enforcement of Property Rights.” Journal of Law and Economics 7 (1964); rpt. in Cowen (1988), pp. 127–​45.   “Toward a Theory of Property Rights.” American Economic Review 57 (1967): 347–​59. Dennett, Daniel. The Intentional Stance. Cambridge, MA: MIT Press, 1987. DeVroey, Michel. Involuntary Unemployment:  The Elusive Quest for a Theory. London: Routledge, 2004. Diamond, Peter. “Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility: Comment.” Journal of Political Economy 75 (1967): 765–​66.   “Wage Determination and Efficiency in Search Equilibrium.” Review of Economic Studies 49 (1982): 217–​27. Dokko, Jane, Megan Mumford, and Diane Whitmore Schanzenbach. “Workers and the Online Gig Economy.” A Hamilton Project Framing Paper (December 2015). Dretske, Fred. Explaining Behavior: Reasons in a World of Causes. Cambridge MA: MIT Press, 1988. Drèze, Jean and Nicholas Stern. 1990. “Shadow Prices and Markets:  Policy Reform, Shadow Prices, and Market Prices.” Rpt. in Layard and Glaister, eds. (1994), pp. 59–​99. Duncan, Greg J. and Richard J. Murnane. Whither Opportunity?:  Rising Inequality, Schools, and Children’s Life Chances. New York: The Russell Sage Foundation (2011).   Restoring Opportunity:  The Crisis of Inequality and the Challenge for American Education. New York: The Russell Sage Foundation. (2013). Dworkin, Gerald. The Theory and Practice of Autonomy. Cambridge:  Cambridge University Press, 1988. Dworkin, Gerald, Gordon Bermant, and Peter Brown, eds., Markets and Morals. New York: HarperCollins, 1977.

References

367

Dworkin, Gerald, Gordon Bermant, and Peter Brown. “What Is Equality? Part 1: Equality of Welfare.” Philosophy & Public Affairs 10 (1981a): 185–​246.   “What is Equality? Part 2: Equality of Resources.” Philosophy & Public Affairs 10 (1981b): 283–​345.   “What is Equality? Part 4: What is Political Equality?” University of San Francisco Law Review, 22 (Fall, 1988):1–​28.   “Foundations of Liberal Equality,” in Grethe Peterson, ed. The Tanner Lectures on Human Values. XI. Salt Lake City: University of Utah Press, 1990, pp. 1–​119.   Sovereign Virtue:  The Theory and Practice of Equality. Cambridge, MA:  Harvard University Press, 2000.  “Sovereign Virtue Revisited.” Ethics 113 (2002): 106–​43. The Economist. “Forget the 1%.” November 8, 2014.” Accessed January 30, 2016 at www .economist.com/​printedition/​2014-​11-​08. Eells, Ellery. Rational Decision and Causality. Cambridge:  Cambridge University Press, 1982. Ellsberg, Daniel. “Risk, Ambiguity, and the Savage Axioms.” Quarterly Journal of Economics 75 (1961): 643–​99. Elster, Jon. Sour Grapes: Studies in the Subversion of Rationality. Cambridge: Cambridge University Press, 1983.   The Cement of Society:  A  Study of Social Order. Cambridge:  Cambridge University Press, 1989a.   “Social Norms and Economic Theory.” Journal of Economic Perspectives 3 (1989b): 99–​117.   Local Justice: How Institutions Allocate Scarce Goods and Necessary Burdens, rev. ed. New York: Russell Sage, 1993.   Closing the Books: Transitional Justice in Historical Perspective. Cambridge: Cambridge University Press, 2004.   Explaining Social Behavior. Cambridge: Cambridge University Press, 2007. Elster, Jon and Aanund Hylland, eds. Foundations of Social Choice Theory. New York: Cambridge University Press, 1986. Elster, Jon and John Roemer, eds. Interpersonal Comparisons of Well-​ Being. Cambridge: Cambridge University Press, 1991. Etzioni, Amitai. The Moral Dimension: Toward a New Economics. New York: Macmillan, 1988. Fankhauser, Samuel, Richard Tol, and David Pearce. “The Aggregation of Climate Change Damages:  A  Welfare Theoretic Approach.” Environmental and Resource Economics 10 (1997): 249–​66. Feinberg, Joel. “Legal Paternalism.” Canadian Journal of Philosophy 1 (1971): 106–​24.   Social Philosophy. Englewood Cliffs, NJ: Prentice-​Hall, 1973.   Harm to Self. Oxford: Oxford University Press, 1986. Fehr, Ernst and Simon Gächter. “Cooperation and Punishment in Public Goods Experiments.” American Economic Review 90 (2000): 980–​94. Fleischacker, Samuel. On Adam Smith’s Wealth of Nations: A Philosophical Companion. Princeton, NJ: Princeton University Press, 2004. Fleming, Marcus. “A Cardinal Concept of Welfare.” Quarterly Journal of Economics 66 (1952): 366–​84. Fleurbaey, Marc. “Equal Opportunity or Equal Social Outcome.” Economics and Philosophy 10 (1995): 25–​55.

368

References

  “Equality of Resources Revisited.” Ethics 113 (2002): 82–​105.   Fairness, Responsibility, and Welfare. New York: Oxford University Press, 2008.   “Equal Opportunity, Reward and Respect for Preferences: Reply to Roemer.” Economics and Philosophy 23 (2012): 201–​16.   “Equality vs. Priority.” Economics and Philosophy 31 (2015): 203–​17. Fleurbaey, Marc and Martin van Hees. “On Rights in Game Forms.” Synthese 123 (2000): 295–​326. Fleurbaey, Marc and François Maniquet. “Fair Social Orderings When Agents Have Unequal Production Skills. Social Choice and Welfare 24 (2005): 93–​127.   A Theory of Fairness and Social Welfare. New York: Cambridge University Press, 2011.   Equality of Opportunity:  The Economics of Responsibility. Hackensack, NJ:  World Scientific, 2012. Fleurbaey, Marc, Kotaro Suzumura and Koichi Tadenuma. “The Informational Basis of the Theory of Fair Allocation.” Social Choice and Welfare 24 (2005): 311–​41. Foley, Duncan. “Resource Allocation and the Public Sector.” Yale Economic Essays Spring (1967): 45–​102. Frank, Robert. Choosing the Right Pond:  Human Behavior and the Quest for Status. Oxford: Oxford University Press, 1987.   Passions within Reason:  The Strategic Role of the Emotions. New  York:  W. W. Norton, 1988.   “The Strategic Role of the Emotions: Reconciling Over –​and Undersocialized Accounts of Behavior.” Rationality and Society 5 (1993): 160–​84.   “What Price the Moral High Ground?” Southern Economic Journal 63 (1996): 1–​17.   What Price the Moral High Ground? Princeton, NJ: Princeton University Press, 2004. Frank, Robert, Thomas Gilovich, and Dennis Regan. “Does Studying Economics Inhibit Cooperation?” Journal of Economic Perspectives 7 (1993): 159–​72. Frankfurt, Harry. “Freedom of Will and the Concept of a Person.” Journal of Philosophy 68 (1971): 5–​20.   On Equality. Princeton, NJ: Princeton University Press, 2015. Frey, Bruno, Felix Oberholzer-​Gee, and Reiner Eichenberger. “The Old Lady Visits Your Backyard: A Tale of Markets and Morals.” Journal of Political Economy 104 (1996): 1297–​313. Frey, Bruno and Felix Oberholzer-​Gee. “The Cost of Price Incentives:  An Empirical Analysis of Motivation Crowding Out,” American Economic Review 87 (1997): 746–​55. Friedman, Milton. “The Methodology of Positive Economics,” in Essays in Positive Economics. Chicago: University of Chicago Press, 1953, pp. 3–​43   Capitalism and Freedom. Chicago: University of Chicago Press, 1962. Frohlich, Norman, Thomas Hunt, Joe Oppenheimer, and R. Harrison Wagner, “Individual Contributions for Collective Goods:  Alternative Models.” Journal of Conflict Resolution 19 (1975): 310–​29. Fuchs, Victor. Who Shall Live:  Health, Economics, and Social Choice. 2nd ed. New York: World Scientific, 2011. Gaertner, Wulf. A Primer in Social Choice Theory, rev. ed. Oxford: Oxford University Press, 2009. Gaertner, Wulf, Prasanta Pattanaik, and Kotaro Suzumura. “Individual Rights Revisited.” Economica 59 (1992): 161–​77.

References

369

Gambetta, Diego, ed. Trust: Making and Breaking Cooperative Relations. Oxford: Basil Blackwell, 1988. Gärdenfors, Peter, “Rights, Games and Social Choice.” Nous 15 (1981): 341–​56. Gärdenfors, Peter and Nils-​Eric Salin, eds. Decision, Probability and Utility:  Selected Readings. Cambridge: Cambridge University Press, 1988. Gaus, Gerald. On Philosophy, Politics, and Economics. Independence, KY: Wadsworth, 2007. Gauthier, David. Morals by Agreement. Oxford: Oxford University Press, 1986.   Moral Dealing:  Contract, Ethics, and Reason. Ithaca, NY:  Cornell University Press, 1990. Gibbard, Allan. “A Pareto-​Consistent Libertarian Claim.” Journal of Economic Theory 7 (1974): 388–​410.   “Interpersonal Comparisons: Preference, Good, and the Intrinsic Reward of a Life,” in Elster and Hylland (1986), pp. 165–​94.   Wise Choices, Apt Feelings: A Theory of Normative Judgment. Cambridge, MA: Harvard University Press, 1990. Gibbard, Allan and William Harper. “Counterfactuals and Two Kinds of Expected Utility,” in C. Hook, J. Leach, and E. McClennen, eds. Foundations and Applications of Decision Theory, vol. 1. Dordrecht: Reidel, 1978, pp. 125–​62. Gilbert, Margaret. “Walking Together:  A  Paradigmatic Social Phenomenon,” in Peter French, Theodore Uehling and Howard Wettstein, eds. Midwest Studies in Philosophy. vol. 15. The Philosophy of the Human Sciences. Notre Dame, IN: University of Notre Dame Press, 1990, pp. 1–​14. Gilens, Martin. Affluence and Influence:  Economic Inequality and Political Power in America. Princeton, NJ: Princeton University Press, 2012. Gintis, Herbert. “The Political Economy of School Choice.” Teachers College Record 96 (1995): 492–​511. Glover, Jonathan. Causing Death and Saving Lives. London: Penguin Books, 1990. Gneezy, Uri and Aldo Rustichini. “A Fine Is a Price,” Journal of Legal Studies 29 (2000): 1–​17. Goldfarb, Robert and William Griffith. “Amending the Economist’s Rational Egoist Model to Include Moral Values and Norms. Part  1. The Problem and Part  2. Alternative Solutions,” K. Koford and J. Miller, eds. Social Norms and Economic Institutions. Ann Arbor: University of Michigan Press, 1991, pp. 39–​84. Goodin, Robert. “Laundering Preferences,” in Elster and Hylland (1986), pp. 75–​101 Gorman, W. “The Intransitivity of Certain Criteria Used in Welfare Economics.” Oxford Economic Papers 7 (1955): 25–​35. Grether, David and Charles Plott. “Economic Theory of Choice and the Preference Reversal Phenomenon.” American Economic Review 69 (1979): 623–​38. Grey, Thomas C. “The Disintegration of Property,” in Pennock and Chapman, eds. 1967, pp. 69–​85. Griffin, James. Well-​ Being:  Its Meaning, Measurement and Moral Importance. Oxford: Clarendon Press, 1986.   Value Judgment: Improving Our Ethical Beliefs. Oxford: Oxford University Press, 1996. Güth, Werner and Eric van Damme. “Information, Strategic Behavior, and Fairness in Ultimatum Bargaining: An Experimental Study.” Journal of Mathematical Psychology 42 (1993): 227–​47.

370

References

Gutmann, Amy. Democratic Education. Princeton, NJ: Princeton University Press, 1987.   “Why Should Schools Care about Civic Education?” in Lorraine McDonnell and P. Michael Timpane, eds. Rediscovering the Democratic Purposes of Education. Lawrence: University of Kansas Press, 2000, pp. 73–​90. Hall, Jonathan and Alan Krueger. “An Analysis of the Labor Market for Uber’s Driver-​ Partners in the United States,” 2015. https://​s3.amazonaws.com/​uber-​static/​ comms/​PDF/​Uber_​Driver-​Partners_​Hall_​Kreuger_​2015.pdf Hamlin, Alan. Ethics, Economics, and the State. New York: St. Martin’s Press, 1986. Hammond, Peter. “Equity, Arrow’s Conditions and Rawls’ Difference Principle.” Econometrica 44 (1976): 793–​804.   “Ex-​Post Optimality as a Dynamically Consistent Objective for Collective Choice under Uncertainty,” in Prasanta Pattanaik and Maurice Salles, eds. Social Choice and Welfare. Amsterdam: North-​Holland, 1983. Hampton Jean. Hobbes and the Social Contract Tradition. Cambridge:  Cambridge University Press, 1986.   “Free-​ Rider Problems in the Production of Collective Goods.” Economics and Philosophy 3 (1987): 245–​73.   The Authority of Reason. Cambridge: Cambridge University Press, 1998. Hanser, Matthew. “Harming Future People,” Philosophy & Public Affairs, 19 (1990): 47–​70. Harberger, Arnold C. “On the Use of Distributional Weights in Social Cost-​Benefit Analysis.” Journal of Political Economy 86 (1978): s87–​s120. Hardin, Russell. Collective Action. Baltimore: Johns Hopkins University Press, 1982.   Morality within the Limits of Reason. Chicago: University of Chicago Press, 1988. Hare, Richard. Moral Thinking: Its Levels, Method, and Point. Oxford: Oxford University Press, 1981. Harper, William. “Ratifiability and Refinements in Two-​ Person Non-​ Cooperative Games,” in Michael Bacharach and Susan Hurley, eds. Foundations of Game Theory: Issues and Advances: Oxford: Blackwell, 1991, pp. 263–​93. Harris, Richard and Olewiler, Nancy. “The Welfare Economics of Ex Post Optimality.” Economica 46 (1979): 137–​47. Harris, Seth and Alan Krueger. “A Proposal for Modernizing Labor Laws for Twentieth-​ Century Work:  The Independent Worker.” 2015. www.hamiltonproject.org/​assets/​ files/​modernizing_​labor_​laws_​for_​twenty_​first_​century_​work_​krueger_​harris.pdf. Harsanyi, John. “Cardinal Welfare, Individualistic Ethics and Interpersonal Comparisons of Utility.” Journal of Political Economy 63 (1955): 309–​21.   “Morality and the Theory of Rational Behavior.” Social Research 44 (1977a): 623–​56; rpt. Sen and Williams (1982), pp. 39–​62.   Rational Behavior and Bargaining Equilibrium in Games and Social Situations. Cambridge: Cambridge University Press, 1977b. Hart, Herbert L. A. “Are There Any Natural Rights?” Philosophical Review 64 (1955): 175–​ 91; rpt. in Waldron (1984), pp. 77–​90. Hausman, Daniel. The Inexact and Separate Science of Economics. Cambridge: Cambridge University Press, 1992.   “The Impossibility of Interpersonal Utility Comparisons.” Mind 104 (1995): 473–​90.   Preference, Value, Choice, and Welfare. New  York:  Cambridge University Press, 2012.

References

371

Hausman, Daniel and Julian Le Grand. “Incentives and Health Policy:  Primary and Secondary Care in the British National Health Service” Social Science and Medicine 49 (1999): 1299–​307. Hausman, Daniel and Michael McPherson, “Taking Ethics Seriously:  Economics and Contemporary Moral Philosophy.” Journal of Economic Literature 31 (1993): 671–​731. Hausman, Daniel and Michael McPherson. “Preference, Belief, and Welfare.” American Economic Review Papers and Proceedings 84 (1994): 396–​400. Hausman, Daniel and Brynn Welch. “Debate: To Nudge or Not to Nudge.” Journal of Political Philosophy 18 (2010): 123–​36. Hayek, Friedrich von. The Road to Serfdom. Chicago: University of Chicago Press, 1944.   “The Use of Knowledge in Society.” American Economic Review 35 (1945): 519–​30.   The Constitution of Liberty. Chicago: University of Chicago Press, 1960.   “The Moral Element in Free Enterprise.” Studies in Philosophy, Politics and Economics. Chicago: University of Chicago Press, 1967, pp. 229–​36.   The Mirage of Social Justice. Chicago: University of Chicago Press, 1976. Herstein, Israel and John W. Milnor. “An Axiomatic Approach to Measurable Utility.” Econometrica 21 (1953): 291–​7. Hicks, John. “The Foundations of Welfare Economics.” Economic Journal 49 (1939): 696–​712. Hirose, Iwao. Egalitarianism. London: Routledge, 2014. Hirsch, Fred. The Social Limits to Growth. Cambridge, MA: Harvard University Press, 1976. Hirschman, Albert. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press, 1970.   “Against Parsimony: Three Easy Ways of Complicating Some Categories of Economic Discourse.” Economics and Philosophy 1 (1985): 7–​22. Hobbes, Thomas. Leviathan, Or the Matter, Forme and Power of a Commonwealth Ecclesiasticall and Civil. 1651. Rpt. New York: Collier Books, 1962. Hochman, Harold and James Rodgers. “Pareto Optimal Redistribution.” American Economic Review 59 (1969): 524–​57. Hohfeld, Wesley. Fundamental Legal Conceptions. New Haven, CT: Yale University Press, 1919. Hollis, Martin and Robert Sugden. “Rationality in Action.” Mind 102 (1993): 1–​35. Hooker, Brad. Ideal Code, Real World:  A  Rule-​Consequentialist Theory of Morality. Oxford: Oxford University Press, 2001. Hooker, Brad, Elinor Mason, and Dale Miller, eds. Morality, Rules, and Consequences. Totowa, NJ: Rowman & Littlefield, 2000. Hornstein, Harvey, Elisha Fisch, and Michael Holmes. “Influence of a Model’s Feelings about His Behavior and His Relevance as a Comparison Other on Observers’ Helping Behavior.” Journal of Personality and Social Psychology 10 (1968): 220–​6. Hume, David. A Treatise of Human Nature (1738). rpt. Oxford: Clarendon Press, 1966.   An Inquiry Concerning the Principles of Morals (1777). Rpt. Indianapolis: Hackett, 1983. Humphreys, Jane. “Cliometrics, Child Labor and the Industrial Revolution.” Critical Review 13 (1999): 269–​83 Hurley, Susan L. Natural Reasons:  Personality and Polity. Oxford:  Oxford University Press, 1989.

372

References

Jones, P. and R. Sugden. “Evaluating Choice.” International Review of Law and Economics 2 (1982): 47–​65. Kagan, Shelly. The Limits of Morality. Oxford: Oxford University Press, 1989.   Normative Ethics. Boulder, CO: Westview Press, 1997. Kahneman, Daniel. 2000a. “Evaluation by Moments: Past and Future.” In Choices, Values and Frames, in D. Kahneman and A. Tversky, eds. New York: Cambridge University Press and the Russell Sage Foundation, ch. 38.  2000b. “Experienced Utility and Objective Happiness:  A  Moment-​ based Approach,” in Choices, Values and Frames, D. Kahneman and A. Tversky, eds. New  York:  Cambridge University Press and the Russell Sage Foundation, pp. 673–​92. Kahneman, Daniel, Paul Slovic, and Aamos Tversky. eds. Judgement under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press, 1982. Kahneman, Daniel, Alan Kreuger, David Schkade, Norbert Schwarz, and Arthur Stone. “Toward National Well-​Being Accounts.” American Economic Review 94 (2004): 429–​34. Kahneman, Daniel and Robert Sugden. “Experienced Utility as a Standard of Policy Evaluation.” Environmental and Resource Economics 32 (2005): 161–​81. Kahneman, Daniel and Alan Krueger. “Developments in the Measurement of Subjective Well Being.” Journal of Economic Perspectives 20 (2006): 3–​24. Kahneman, Daniel and Richard Thaler, 2006. “Utility Maximization and Experienced Utility.” Journal of Economic Perspectives 20 (2006): 221–​34. Kalai, Ehud and Meir Smorodinsky. “Other Solutions to Nash’s Bargaining Problem.” Econometrica 43 (1975): 513–​18. Kaldor, Nicholas. “Welfare Propositions of Economics and Interpersonal Comparisons of Utility.” Economic Journal 49 (1939): 549–​52. Kamm, Frances. Morality, Mortality. Vol. 1. Death and Whom to Save from It. Oxford: Oxford University Press, 1993. Kant, Immanuel. Groundwork of the Metaphysics of Morals. 1785. Trans. H.  Paton. New York: Harper & Row, 1948. Kavka, Gregory. “The Paradox of Future Individuals. Philosophy & Public Affairs 11 (1982): 93–​112.   Hobbesian Moral and Political Theory. Princeton, NJ: Princeton University Press, 1986. Kelman, Steven. What Price Incentives? Boston, MA: Auburn House, 1981.   “A Case for In-​Kind Transfers.” Economics and Philosophy 2 (1986): 53–​74. Kittay, Eva. Love’s Labor:  Essays on Women, Equality and Dependency. London: Routledge, 2013. Knight, Frank. Risk, Uncertainty and Profit. 1921. Rpt. Chicago: University of Chicago Press, 1971. Kolm, Serge-​Christophe. Modern Theories of Justice. Cambridge, MA: MIT Press, 1997.   Justice et Équité. Paris:  Éditions du Centre National de la Recherche Scientifique, 1972; trans. Harold See as Justice and Equity. Cambridge, MA: MIT Press, 2002. Kraut, Richard. What Is Good and Why. Cambridge, MA: Harvard University Press, 2007. Kreps, David M., Paul Milgrom, John Roberts, and Robert Wilson. “Rational Cooperation in the Finitely Repeated Prisoners’ Dilemma.” Journal of Economic Theory 27 (1982): 245–​52. Kreps, David and Robert Wilson. “Sequential Equilibria.” Econometrica 50 (1982): 863–​4.

References

373

Kukathas, Chandran and Philip Pettit. A Theory of Justice and Its Critics. Oxford: Blackwell, 1990. Kymlicka, Will. “Rethinking the Family.” Philosophy and Public Affairs 20 (1991): 77–​97.   Contemporary Political Philosophy, 2nd ed. Oxford: Oxford University Press, 2001. Laffont, Jean-​Jacques. Fundamentals of Public Economics. Trans. John Bonin and Hélène Bonin. Cambridge, MA: MIT Press, 1988. Laslier, Jean-​François, Marc Felurbaey, Nicolas Gravel and Alan Trannoy. Freedom in Economics: New Perspectives in Normative Analysis. London: Routledge, 1998. Layard, Richard. Happiness: Lessons from a New Science. London: Penguin Books, 2006. Layard, Richard and Stephen Glaister, eds. Cost-​Benefit Analysis, 2nd ed. Cambridge: Cambridge University Press, 1994. Le Grand, Julian. Equity and Choice:  An Essay in Economics and Applied Philosophy. New York: Harper-​Collins, 1991. Le Grand, Julian and Bill New. Government Paternalism: Nanny State or Helpful Friend. Princeton, NJ: Princeton University Press, 2014. Lemann, Nicholas. “Transaction Man,” The New Yorker, Oct 1, 2012. www.newyorker .com/​magazine/​2012/​10/​01/​transaction-​man. Lerner, Abba. “Consumption-​Loan Interest and Money.” Journal of Political Economy 67 (1959a): 512–​18.  “Rejoinder.” Journal of Political Economy 67 (1959b): 523–​5. Levi, Isaac. The Enterprise of Knowledge. Cambridge, MA: MIT Press, 1980.   “The Paradoxes of Allais and Ellsberg.” Economics and Philosophy 2 (1986): 23–​53.   “Pareto Unanimity and Consensus.” Journal of Philosophy 89 (1990): 481–​92. Lewis, David. Convention. Cambridge, MA: Harvard University Press, 1969.   “Causal Decision Theory. Australasian Journal of Philosophy 59 (1981): 5–​30. Lichtenstein, Sarah and Paul Slovic. “Reversals of Preference between Bids and Choices in Gambling Decisions.” Journal of Experimental Psychology 89 (1971): 46–​55.   “Response-​Induced Reversals of Preference in Gambling: An Extended Replication in Las Vegas.” Journal of Experimental Psychology 101 (1973): 16–​20.  eds. The Construction of Preferences. Cambridge: Cambridge University Press, 2006. Lipsey, Richard and Kelvin Lancaster. “The General Theory of Second Best.” Review of Economic Studies 24 (1956): 11–​32. List, Christian and Philip Pettit. “Aggregating Sets of Judgments:  An Impossibility Result.” Economics and Philosophy 18 (2002): 89–​110. “Aggregating Sets of Judgments: Two Impossibility Results Compared.” Synthese 140 (2004): 207–​35. Little, Ian. A Critique of Welfare Economics, 2nd ed. Oxford:  Oxford University Press, 1957. Locke, John. Second Treatise on Government (1690), in Two Treatises of Government: A Critical Edition, 2nd ed. by Peter Laslett. Cambridge: Cambridge University Press, 1967. Loewenstein, George. Exotic Preferences: Behavioral Economics and Human Motivation. New York: Oxford University Press, 2008. Lomasky, Loren. Persons, Rights and the Moral Community. New York: Oxford University Press, 1987. Loomes, Graham and Robert Sugden. “Regret Theory: an Alternative Theory of Rational Choice under Uncertainty.” Economic Journal 92 (1982): 805–​24.

374

References

Lucas, Robert E. “Unemployment Policy.” American Economic Review 68 (1978): 353–​7.   2004. “The Industrial Revolution:  Past and Future.” Minneapolis:  Federal Reserve Bank of Minneapolis. Luce, Robert Duncan and Howard Raiffa. Games and Decisions. New York: Wiley, 1957. Lukes, Steven. 1977. Essays in Social Theory. New York: Columbia University Press. MacAskill, William. Doing Good Better. London: Gotham Books, 2015. MacCallum, Gerald. “Negative and Positive Freedom.” Philosophical Review 76 (1967): 312–​34. Machan, Tibor, ed. The Libertarian Reader. Totowa, NJ: Rowman & Littlefied, 1982. Machina, Mark. “Choice under Uncertainty: Problems Solved and Unsolved.” Journal of Economic Perspectives 1 (1987): 121–​54. Machlup, Fritz, “Positive and Normative Economics,” in Robert Heilbroner, ed. Economic Means and Social Ends: Essays in Political Economics. Englewood Cliffs, NJ: Prentice-​Hall, 1969, pp. 99–​124. MacKay, Alfred. Arrow’s Theorem:  The Paradox of Social Choice:  A  Case Study in the Philosophy of Economics. New Haven, CT: Yale University Press, 1980.   “Extended Sympathy and Interpersonal Utility Comparisons.” Journal of Philosophy 83 (1986): 305–​22. Maniquet, François. “Social Orderings and the Evaluation of Public Policy.” Revue d’Économie Politique 117 (2007): 37–​60. Mankiw, N. Gregory. “Defending the One Percent” Journal of Economic Perspectives, 27 (2013): 21–​34. Mansbridge, Jane, ed., Beyond Self-​Interest. Chicago: University of Chicago Press, 1990, pp. 254–​63. Margolis, Howard. Selfishness, Altruism and Rationality. Cambridge:  Cambridge University Press, 1982. Marmot, Michael. The Status Syndrome: How Your Social Standing Directly Affects Your Health and Life Expectancy. London: Bloomsbury, 2004. Martin, Rex. A System of Rights. Oxford: Clarendon Press, 1993. Marwell, Gerald and Ruth Ames. “Economists Free Ride:  Does Anyone Else? Experiments on the Provision of Public Goods. IV.” Journal of Public Economics 15 (1981): 295–​310. Marx, Karl. Capital (1867): New York: Vintage, 1977. Mason, Andrew, ed. Ideals of Equality. Oxford: Blackwell, 1998. Maynard Smith, J. Evolution and the Theory of Games. Cambridge: Cambridge University Press, 1982. McDowell, John. “Are Moral Judgments Hypothetical Imperatives?” Proceedings of the Aristotlian Society Supplementary Volume, 52 (1978): 13–​29. McKean, Roland N. “Economics of Trust, Altruism, and Corporate Responsibility,” in Edmund Phelps, ed. Altruism, Morality and Economic Theory. New York: Russell Sage Foundation, 1975, pp. 29–​44. McClennen, Edward. Rationality and Dynamic Choice:  Foundational Explorations. Cambridge: Cambridge University Press, 1990. McPherson, Michael, “Mill’s Moral Theory and the Problem of Preference Change.” Ethics 92 (1982): 252–​73.   “Efficiency and Liberty in the Productive Enterprise: Recent Work in the Economics of Work Organization.” Philosophy & Public Affairs 12 (1983a): 354–​68.

References

375

  “Want Formation, Morality, and Some ‘Interpretive’ Aspects of Economic Inquiry,” in Norma Haan, Robert N. Bellah, Paul Rabinow, and William M. Sullivan, eds. Social Science as Moral Inquiry. New York: Columbia University Press, 1983b, pp. 96–​124. Meade, James. Equality, Efficiency and the Ownership of Property. 1964. Meckling, William. “An Exact Consumption-​Loan Model of Interest:  A  Comment.” Journal of Political Economy 68 (1960a): 72–​6.  “Rejoinder.” Journal of Political Economy 68 (1960b): 83–​4. Melden, A. I. Free Action. London: Routledge, 1961. Mill, John Stuart. On Liberty. (1859). Edited by Currin V. Shields. New York: Macmillan, 1985.   Utilitarianism (1863). Rpt. in Marshall Cohen, ed. The Philosophy of John Stuart Mill. New York: Modern Library, 1961, pp. 321–​98. Miller, David. “Arguments for Equality.” Midwest Studies in Philosophy 7 (1982): 73–​88. Miller, David and Michael Walzer. Pluralism, Justice, and Equality. Oxford:  Oxford University Press, 1995. Mishan, E. Cost Benefit Analysis: An Introduction. New York: Praeger, 1971.   An Introduction to Normative Economics. Oxford: Oxford University Press, 1981. Mongin, Philippe. “Consistent Bayesian Aggregation.” Journal of Economic Theory 66 (1995): 313–​51.   “The Impartial Observer Theorem of Social Ethics.” Economics and Philosophy 17 (2001): 147–​79.   “Value Judgments and Value Neutrality in Economics.” Economica 73 (2006): 257–​86. Mongin, Philippe and Claude d’Aspremont. “Utility Theory and Ethics,” in Salvador Barberà, Peter Hammond, and Christian Seidl, eds. Handbook of Utility Theory, vol. 1. Dordrecht: Kluwer, 1998, pp. 371–​481 Moulin, Hervé and William Thomson. “Axiomatic Analysis of Resource Allocation Problems.” in Arrow et al. (eds.) (1997), pp. 101–​20. Mueller, Dennis. “The Corporation and the Economist.” International Journal of Industrial Organization 10 (1992): 147–​70; reprinted and cited from D. Hausman, ed. The Philosophy of Economics:  An Anthology, 2nd. ed. New  York:  Cambridge University Press, 1994, pp. 289–​314. Murphy, Liam. “Institutions and the Demands of Justice,” Philosophy & Public Affairs 27 (1999): 251–​91.   Moral Demands in Nonideal Theory. New York: Oxford University Press, 2000. Murphy, Liam and Thomas Nagel. The Myth of Ownership:  Taxes and Justice. New York: Oxford University Press, 2004. Musgrave, Richard. “Maximin, Uncertainty, and the Leisure Trade-​ Off.” Quarterly Journal of Economics 88 (1974): 625–​42. Nagel, Thomas. “The Limits of Objectivity,” in Sterling McMurrin, ed. The Tanner Lectures on Human Values, vol. i. Salt Lake City: University of Utah Press, 1980, pp. 75–​139.   “Libertarianism without Foundations,” rpt. in Jeffrey Paul, ed. Reading Nozick: Essays on Anarchy, State and Utopia. Oxford: Basil Blackwell, 1983, pp. 191–​205.   The View from Nowhere. New York: Oxford University Press, 1986. Narveson, Jan. The Libertarian Idea. Philadelphia: Temple University Press, 1988. Nash, John. “The Bargaining Problem.” Econometrica 18 (1950): 155–​62.   “Non-​Cooperative Games.” Annals of Mathematics. 54 (1951): 286–​95.

376

References

Nehring, Klaus and Clemens Puppe. “A Theory of Diversity.” Econometrica 70 (2002): 1155–​98. Ng, Yew-​Kwang. Welfare Economics: Introduction and Development of Basic Concepts, rev. ed. London: Macmillan, 1983. Nordhaus, William. “A Review of the Stern Report on the Economics of Climate Change.” Journal of Economic Literature 45 (2007): 686–​702. Nozick, Robert. “Newcomb’s Problem and Two Principles of Choice.” in Nicholas Rescher, ed. Essays in Honor of Carl G.  Hempel. Dordrecht:  Kluwer, 1969, pp. 114–​46.   Anarchy, State and Utopia. New York: Basic Books, 1974. Nussbaum, Martha. “ ‘Whether from Reason or Prejudice’ ”: Taking Money for Bodily Services.” Journal of Legal Studies 27 (1998): 693–​724.   Women and Human Development. Cambridge: Cambridge University Press, 2000.   Upheavals of Thought: The Intelligence of Emotions. Cambridge: Cambridge University Press, 2001.   Frontiers of Justice:  Disability, Nationality, Species Membership. Cambridge, MA: Harvard University Press, 2006. Nussbaum, Martha and Amartya Sen, eds. The Quality of Life. Oxford:  Clarendon Press, 1993. OECD. OECD Guidelines on Measuring Subjective Well-​being. OECD. 2013. http://​ dx.doi.org/​10.1787/​9789264191655-​en. Okin, Susan Moller. Justice, Gender and the Family. New York: Basic Books, 1989. Okun, Arthur. Equality and Efficiency:  the Big Tradeoff. Washington, DC:  Brookings Institution, 1975. O’Neill, Martin. “What Should Egalitarians Believe?” Philosophy & Public Affairs 36 (2008): 119–​56. Osborne, Martin. An Introduction to Game Theory. Oxford:  Oxford University Press, 2003. Michael Otsuka. Libertarianism without Inequality. Oxford:  Oxford University Press, 2003. Overvold, Mark. “Morality, Self-​Interest, and Reasons for Being Moral.” Philosophy and Phenomenological Research 44 (1984); 493–​507. Parfit, Derek. “Prudence, Morality and the Prisoner’s Dilemma.” Proceedings of the British Academy, 1979.   Reasons and Persons. Oxford: Oxford University Press, 1984.   “Equality or Priority.” Lindley Lecture, University of Kansas, 1991; rpt. in Clayton and Williams (2000), pp. 81–​125.   On What Matters. Oxford: Oxford University Press, 2011. Pattanaik, Prasanta and Y. Xu. “On Ranking Opportunity Sets in Terms of Freedom of Choice.” Recherches Economiques de Louvain 56 (1990): 383–​90. Paul, Ellen, Fred Miller, Jr., Jeffrey Paul and John Ahrens, eds. The New Social Contract: Essays on Gauthier. Oxford: Blackwell, 1988. Paxton, Will, ed. Equal Shares? Building a Progressive and Coherent Asset-​Based Welfare Policy. London: Institute for Public Policy Research, 2002. Pazner, Elisha A. and David Schmeidler. “A Difficulty in the Concept of Fairness.” Review of Economic Studies 41 (1974): 441–​3.

References  

377

“Egalitarian Equivalent Allocations:  A  New Concept of Economic Equity.” 92 (1978): 671–​87. James Pennock, and John Chapman eds. Property. New  York:  New  York University Press, 1967. Peter, Fabienne. “Choice, Consent, and the Legitimacy of Market Transactions.” Economics and Philosophy 20 (2004): 1–​18. Pettit, Philip. “Virtus Normativa:  Rational Choice Perspectives.” Ethics 100 (1990): 725–​55.   Just Freedom: A Moral Compass for a Complex World. New York: W. W.Norton, 2014. Pettit, Philip and Robert Sugden. “The Backward Induction Paradox.” Journal of Philosophy 86 (1989): 169–​82. Phelps, Edmund S. “Taxation of Wage Income for Economic Justice.” Quarterly Journal of Economics 87 (1973): 332–​54. Piff, Paul, Daniel Stancato, Stéphane Côté, Rodolfo Mendoza-​Denton, and Dacher Keltner. “Higher Social Class Predicts Increased Unethical Behavior.” Proceedings of the National Academy of Sciences 109 (2012): 4086–​91. Pigou, A. C. The Economics of Welfare. London: Macmillan, 1920. Piketty, Thomas. Capital in the Twentieth Century. Trans. Arthur Goldhammer. Cambridge, MA: Harvard University Press, 2014. Pinker, Steven. The Blank Slate: The Modern Denial of Human Nature. New York: Viking Press, 2002. Plamenatz, John. “Diversity of Rights and Kinds of Equality.” in Pennock and Chapman, eds. (1967), pp. 79–​98. Pogge, Thomas. Realizing Rawls. Ithaca, NY: Cornell University Press, 1989.   World Poverty and Human Rights:  Cosmopolitan Responsibilities and Reforms. Cambridge, MA: Polity Press, 2002. Pojman, Lewis and Robert Westmoreland, eds. Equality:  Selected Readings. Oxford: Oxford University Press, 1997. Polanyi, Karl. The Great Transformation. Boston: Beacon Press, 1944. Posner, Richard. “Utilitarianism, Economics, and Legal Theory.” Journal of Legal Studies 8 (1979): 103–​40. Press, William and Freeman Dyson. “Iterated Prisoner’s Dilemma Contains Strategies That Dominate Any Evolutionary Opponent.” Proceedings of the National Academy of Science 109 (2012): 10409–​13. Proulx, Chris, Steven Hsu, Ken Binmore, and Joe Swierzbinski. “Focal Points and Bargaining.” International Journal of Game Theory 22 (1993): 381–​409. Putnam, Hilary. The Collapse of the Fact/​Value Dichotomy and Other Essays. Cambridge, MA: Harvard University Press, 2002. Putnam, Robert. Our Kids:  The American Dream in Crisis. New  York:  Simon & Schuster, 2015. Putterman, Louis. “On Some Recent Explanations of Why Capital Hires Labor.” Economic Inquiry 22 (1984): 171–​87. Radin, Margaret. “Market Inalienability.” Harvard Law Review 100 (1987): 1849–​1937.   Contested Commodities. Cambridge, MA: Harvard University Press, 1996. Railton, Peter. “Alienation, Consequentialism, and the Demands of Morality.” Philosophy and Public Affairs 13 (1984): 134–​71.

378

References

Ramsey, Frank. “Truth and Probability,” in R.  B. Braithwaite, ed. The Foundations of Mathematics and Other Logical Essays. London: Routledge and Kegan Paul, 1926, pp. 156–​98. Rand, Ayn. The Virtue of Selfishness. New York: Signet Books, 1964. Rawls, John. “Two Concepts of Rules.” Philosophical Review 64 (1955): 3–​33.   A Theory of Justice. Cambridge, MA: Harvard University Press, 1971.   “Social Unity and Primary Goods.” in Sen and Williams 1982, pp. 159–​86.   Political Liberalism. New York: Columbia University Press, 1993.   Justice as Fairness. Cambridge, MA: Harvard University Press, 2001a.   Law of Peoples. Cambridge, MA: Harvard University Press, 2001b. Raz, Joseph. The Morality of Freedom. Oxford: Oxford University Press, 1984. Reder, Melvin. “The Place of Ethics in the Theory of Production,” in Michael Boskin, ed. Economics and Human Welfare:  Essays in Honor of Tibor Scitovsky. New York: Academic Press, 1979, pp. 133–​46. Regan, Donald. Utilitarianism and Cooperation. Oxford: Clarendon Press, 1980. Reich, Michael and James Devine, “The Microeconomics of Conflict and Hierarchy in Capitalist Production.” Review of Radical Political Economy 12 (1981): 27–​45. Riley, Jonathan. “Rights to Liberty in Purely Private Matters, Part I.” Economics and Philosophy 5 (1989): 121–​66.   “Rights to Liberty in Purely Private Matters, Part II.” Economics and Philosophy 6 (1990): 27–​64. Risse, Matthias. On Global Justice. Princeton, NJ: Princeton University Press, 2012. Robertson, Dennis. “What Does the Economist Economize?” in Economic Commentaries. London: Staples Press, 1956, pp. 147–​55. Robbins, Lionel. An Essay on the Nature and Significance of Economic Science, 2nd ed. (1st ed. 1932). London: Macmillan, 1935.   “Economics and Political Economy.” American Economic Review, Papers and Proceedings 71 (1981): 1–​10. Roemer, John. “Equality of Talent.” Economics and Philosophy 1 (1985): 151–​88.   “The Mismarriage of Bargaining Theory and Distributive Justice.” Ethics 97 (1986a): 88–​110.   “Equality of Resources Implies Equality of Welfare.” Quarterly Journal of Economics 101 (1986b): 751–​84.   “Egalitarianism, Responsibility, and Information.” Economics and Philosophy 3 (1987): 215–​44.   Free to Lose. Cambridge, MA: Harvard University Press, 1988.   Theories of Distributive Justice. Cambridge, Mass.: Harvard University Press, 1996.   Equality of Opportunity. Cambridge, Mass.: Harvard University Press, 1998.   “Equality of Opportunity:  A  Progress Report.” Social Choice and Welfare 19 (2002): 455–​72.   “On Several Approaches to Equality of Opportunity.” Economics and Philosophy 23 (2012): 165–​200.   “Kantian Optimization: A Microfoundation for Cooperation.” Journal of Public Economics 127 (2015): 45–​57. Rosenberg, Alexander. Microeconomic Laws:  A  Philosophical Analysis. Pittsburgh, University of Pittsburgh Press, 1976.   Philosophy of Social Science. Boulder: Westview Press, 1995.

References

379

Roth, Alvin E. “Repugnance as a Constraint on Markets.” Journal of Economic Perspectives 21 (2007): 37–​58. Roth, Alvin and Michael Malouf. “Game Theoretical Models and the Role of Information in Bargaining.” Psychological Review 86 (1979): 574–​94. Roth, Alvin, Michael Malouf, and J. Keith Murnighan. “Sociological Versus Strategic Factors in Bargaining.” Journal of Economic Behavior and Organizations 2 (1981): 153–​77. Rousseau, Jean-​Jacques. The Social Contract. tr. Maurice Cranston. Harmondsworth, Middlesex: Penguin, 1968. Ryan, Alan. (ed.) The Idea of Freedom: Essays in Honor of Isaiah Berlin. New York: Oxford University Press, 1979. Sagoff, Mark. “Values and Preferences.” Ethics 96 (1986): 301–​16.   Price, Principle, and the Environment. Cambridge:  Cambridge University Press, 2004. Samuelson, Paul. Foundations of Economic Analysis. Cambridge, MA:  Harvard University Press, 1947.   “Evaluation of Real National Income.” Oxford Economic Papers N.S.2 (1950): 1–​29.   “An Exact Consumption-​ Loan Model of Interest with or without the Social Contrivance of Money.” Journal of Political Economy 66 (1958): 467–​82.  “Reply.” Journal of Political Economy 67 (1959): 518–​22.   “Infinity, Unanimity and Singularity:  A  Reply.” Journal of Political Economy 68 (1960): 76–​83. Sandel, Michael. Liberalism and the Limits of Justice. Cambridge: Cambridge University Press, 1982.   What Money Can’t Buy: The Moral Limits of Markets. New York: Farrar, Straus and Giroux, 2013. Satz, Debra. Why Some Things Should Not Be for Sale:  The Moral Limits of Markets. New York: Oxford University Press, 2010. Savage, Leonard. The Foundations of Statistics. New York: Dover, 1972. Scanlon, Thomas. “Preference and Urgency.” Journal of Philosophy 72 (1975): 655–​70.   “Contractualism and Utilitarianism,” in Sen and Williams (1982), pp. 103–​28.   “Equality of Resources and Equality of Welfare:  A  Forced Marriage?” Ethics 97 (1986): 111–​18.   What We Owe to Each Other. Cambridge, MA: Harvard University Press, 1998.   “The Diversity of Objections to Inequality,” in The Difficulty of Tolerance: Essays in Political Philosophy. Cambridge: Cambridge University Press, 2003, pp. 202–​18. Schaffner, Julie. “Attached Farm Labor, Limited Horizons and Servility.” Journal of Development Economics 47 (1995): 241–​70. Scheffler, Samuel. The Rejection of Consequentialism:  A  Philosophical Investigation of the Considerations Underlying Rival Moral Conceptions. Oxford:  Clarendon Press, 1982.  ed. Consequentialism and Its Critics. Oxford: Oxford University Press, 1988.   “What Is Egalitarianism?” Philosophy and Public Affairs 31 (2003): 5–​39. Schelling, Thomas. The Strategy of Conflict. Cambridge, MA: Harvard University Press, 1960.   Micromotives and Macrobehavior. New York: Norton, 1978.   Choice and Consequence. Cambridge, MA: Harvard University Press, 1984.

380

References

Schick, Frederic. “Money Pumps and Dutch Bookies.” Journal of Philosophy 83 (1986): 112–​19. Schmidtz, David. “The Institution of Property.” Social Philosophy and Policy 11 (1994): 42–​62. Schmidtz, David and Jason Brenner. A Brief History of Liberty. New  York: Wiley-​Blackwell, 2010. Schotter, Andrew. The Economic Theory of Social Institutions. Cambridge:  Cambridge University Press, 1981. Schultze, Charles L. The Public Use of Private Interest. Washington, DC:  Brookings Institution Press, 1977. Seidenfeld Teddy, Joseph Kadane, and Mark Schervish. “On the Shared Preferences of Two Bayesian Decision Makers.” Journal of Philosophy 86 (1989): 225–​44. Selten, Reinhart. “Re-​ examination of the Perfectness Concept for Equilibrium in Extensive Games.” International Journal of Game Theory 4 (1975): 22–​25. Sen, Amartya. “Isolation, Assurance, and the Social Rate of Discount.” Quarterly Journal of Economics 81 (1967): 112–​24.   Collective Welfare and Social Choice. San Francisco: Holden-​Day, 1970a.   “The Impossibility of a Paretian Liberal.” Journal of Political Economy 78 (1970b): 152–​57.   “Choice Functions and Revealed Preference.” Review of Economic Studies 38 (1971): 307–​17.   “Welfare Inequalities and Rawsian Axiomatics.” Theory and Decision 7 (1976): 243–​62.   “Rational Fools.” Philosophy and Public Affairs 6 (1977): 317–​44.   “Utilitarianism and Welfarism.” Journal of Philosophy 76 (1979): 463–​88.   Choice, Welfare and Measurement. Cambridge, MA: MIT Press, 1982a.   “Rights and Agency.” Philosophy and Public Affairs 11 (1982b): 3–​39.   “Liberty and Social Choice.” Journal of Philosophy 80 (1983): 5–​28.   Professor Dr. P. Hennipman Lectures in Economics: Theory, Institutions, Policy. Vol. 7. Commodities and Capabilities. Amsterdam: North-​Holland, 1985a.   “Rationality and Uncertainty,” Theory and Decision 18 (1985b): 109–​27.   “Foundations of Social Choice Theory: An Epilogue,” in Elster and Hylland (1986), pp. 213–​48.   On Ethics and Economics. Oxford: Blackwells, 1987a.   “The Standard of Living. Lecture I.  Concepts and Critiques,” (1987b), in Sen et  al. (1987), pp. 1–​19.   “The Standard of Living:  Lecture II, Lives and Capabilities,” (1987c), in Sen et  al. (1987), pp. 20–​38.   “Gender and Cooperative Conflicts,” in Irene Tinker, ed. Persistent Inequalities. London: Oxford University Press, 1990a, pp. 123–​49.   “Welfare, Freedom and Social Choice: A Reply.” Richerches Économiques de Louvain 56 (1990b): 451–​85.   Inequality Reexamined. Cambridge, MA: Harvard University Press, 1992a.   “Minimal Liberty.” Economica 59 (1992b): 139–​59.   “The Possibility of Social Choice.” American Economic Review 89 (1999): 349–​78.   Rationality and Freedom. Cambridge, MA: Harvard University Press, 2002.

References

381

  “Rational Choice:  Discipline, Brand Name, and Substance,” in Fabienne Peter and Hans Bernhard Schmid, eds. Rationality and Commitment. Oxford:  Oxford University Press, 2007, pp. 339–​61.   The Idea of Justice. Cambridge, MA: Harvard University Press, 2009. Sen, Amartya and Bernard Williams, eds. Utilitarianism and Beyond. Cambridge: Cambridge University Press, 1982. Sen, Amartya, John Muellbauer, Ravi Kanbur, Keith Hart, and Bernard Williams. The Standard of Living. Edited by Geoffrey Hawthorn. Cambridge: Cambridge University Press, 1987. Shafer-​Landau, Russ. Whatever Happened to Good and Evil? New  York:  Oxford University Press, 2003. Shapiro, O. and Joseph E. Stiglitz. “Can Unemployment be Involuntary? Reply.” American Economic Review 75 (1985): 1215–​17. Sidgwick, Henry. The Methods of Ethics, 6th ed. London: Macmillan, 1901. Singer, Peter. “Altruism and Commerce: A Defense of Titmuss against Arrow.” Philosophy & Public Affairs 2 (1973): 312–​20.   Animal Liberation: A New Ethics for Our Treatment of Animals. New York: Random House, 1975.   Practical Ethics. Cambridge: Cambridge University Press, 1979.   One World: The Ethics of Globalization. New Haven, CT: Yale University Press, 2002.   The Most Good You Can Do: How Effective Altruism Is Changing Ideas about Living Ethically. New Haven, CT: Yale University Press, 2015. Singer, Peter, ed. Applied Ethics. Oxford: Oxford University Press, 1986. Skidelsky, Robert. John Maynard Keynes:  The Economist as Savior 1920–​ 1937. New York: Penguin Books, 1992, Skyrms, Brian. Causal Necessity: A Pragmatic Investigation of the Necessity of Laws. New Haven, CT: Yale University Press, 1980.   Evolution of the Social Contract. Cambridge: Cambridge University Press, 1996.   The Stag Hunt and the Evolution of Social Structure. Cambridge: Cambridge University Press, 2003.   Social Dynamics. Oxford: Oxford University Press, 2014. Smart, J. J.  C. and Bernard Williams. Utilitarianism:  For and Against. Cambridge: Cambridge University Press. Smith, Adam. The Theory of Moral Sentiments. Edited by D. D. Raphael and A. L. Macfie. Oxford: Clarendon Press, 1776.   An Inquiry into the Nature and Causes of the Wealth of Nations. (1776). Rpt. New York: Random House, 1937. Snowdon, Brian and Howard Vane. Conversations with Leading Economists: Interpreting Modern Macroeconomics. Cheltenham: Edward Elgar, 1999. Solow, Robert. “Wage Bargaining and Unemployment.” American Economic Review 71 (1981): 896–​908.   The Labor Market as a Social Institution. Oxford: Basil Blackwell, 1990. Steiner, Hillel. An Essay on Rights. New York: Wiley-​Blackwell, 1994. Stern, Nicholas. The Economics of Climate Change:  The Stern Review. Cambridge: Cambridge University Press, 2007. Stich, Steven. From Folk Psychology to Cognitive Science:  The Case against Belief. Cambridge, MA: MIT Press, 1983.

382

References

Stiglitz, Joseph. The Price of Inequality. New York: Norton, 2013.   The Great Divide:  Unequal Societies and What We Can Do about Them. New York: Norton, 2015. Stone, Christopher. Should Trees Have Standing, and Other Essays on Law, Morals and the Environment. Dobbs Ferry, NY: Oceana, 1996. Strahilevitz, Lior. “How Changes in Property Regimes Influence Social Norms: Commodifying California’s Carpool Lanes.” Indiana Law Journal 75 (2000): 1231–​96. Strasnick, Steven. “Social Choice and the Derivation of Rawls’s Difference Principle.” Journal of Philosophy 73 (1976): 85–​99. Streeten, Paul. “Appendix:  Recent Controversies,” in Gunnar Myrdal, ed. The Political Element in the Development of Economic Theory. Trans. Paul Streeten. London: Routledge and Kegan Paul, 1953, pp. 208–​17. Sugden, Robert. “Liberty, Preference and Choice.” Economics and Philosophy 2 (1985): 213–​31.   The Economics of Rights, Co-​operation and Welfare. New York: Blackwell, 1986.   “The Metric of Opportunity.” Economics and Philosophy 14 (1998): 307–​37.   “The Logic of Team Reasoning.” Philosophical Explorations 6 (2003): 165–​181. Sugden, Robert and Alan Williams. The Principles of Practical Cost-​Benefit Analysis. New York: Oxford University Press, 1978. Sumner, L. W. The Moral Foundation of Rights. Oxford: Clarendon Press, 1987. Sunstein, Cass. Why Nudge? The Politics of Libertarian Paternalism. New Haven, CT: Yale University Press, 2014.   Laws of Fear: Beyond the Precautionary Principle. New York: Cambridge University Press, 2005.   “Nudging and Choice Architecture:  Ethical Considerations.” Yale Journal on Regulation 32 (2015): Swift, Adam. Political Philosophy: A Beginners’ Guide for Students and Politicians, 3rd ed. Cambridge: Cambridge University Press, 2013. Tan, Kok-​Chor. Justice without Borders: Cosmopolitanism, Nationalism, and Patriotism. New York: Cambridge University Press, 2004. Tawney, R. H. Equality. New York: Harcourt, Brace, 1931. Taylor, Michael C. The Possibility of Cooperation. New  York:  Cambridge University Press, 1987. Taylor, Michael and Hugh Ward. “Chickens, Whales and Lumpy Goods:  Alternative Models of Public Goods Provision.” Political Science 30 (1982): 350–​70. Temkin, Larry. Inequality. Oxford: Oxford University Press, 1993.   “Egalitarianism Defended.” Ethics 113 (2003): 764–​82.   “Equality as Comparative Fairness.” Journal of Applied Philosophy, doi:  10.1111/​ japp.12140, 2015. Thaler, Richard. Misbehaving:  The Making of Behavioral Economics. New  York: Norton, 2015. Thaler, Richard and Cass Sunstein. “Libertarian Paternalism.” American Economic Review, 93 (2003a): 175–​9. Thaler, Richard and Cass Sunstein. “Libertarian Paternalism Is Not an Oxymoron.” University of Chicago Law Review 70 (2003b): 1159–​202.   Nudge. New Haven, CT: Yale University Press, 2008.

References

383

Thiel, H. “The Measurement of Inequality by Components of Income.” Economic Letters, 2 (1979): 197–​9.) Thomson, Garrett. Needs. London: Routledge, 1987. Thomson, Judith. The Realm of Rights. Cambridge, MA: Harvard University Press, 1990. Thomson, William. “Fair Allocation Rules.” In Arrow and Sen (2011), pp. 393–​506. Thomson, William and Terje Lensberg, Axiomatic Theory of Bargaining with a Variable Number of Agents. New York: Cambridge University Press, 1989. Thurow, Lester C. “Cash vs. In-​Kind Redistribution,” in Dworkin, Bermant, Brown, eds. 1977, pp. 85–​106. Titmuss, Richard. The Gift Relationship:  From Human Blood to Social Policy. New York: Random House, 1971. Tomasi, John. 2013. Free Market Fairness. Princeton, NJ: Princeton University Press. Tuomela, Raimo. Cooperation: A Philosophical Study, Philosophical Studies Series 82. Dordrecht: Kluwer, 2000.   “The We-​mode and the I-​mode.” In Schmitt, F. ed., Socializing Metaphysics:  The Nature of Social Reality. Lanham, MD: Rowman & Littlefield, 2003, pp. 93–​128. Tversky, Amos and Daniel Kahneman. “The Framing of Decisions and the Psychology of Choice.” Science 211 (1981): 453–​58. Tversky, Amos, Paul Slovic, and Daniel Kahneman. “The Causes of Preference Reversal.” American Economic Review 80 (1990): 204–​17. Tversky, Amos and Richard Thaler. “Preference Reversals.” Journal of Economic Perspectives 4 (1990): 201–​11. Vanderschraaf, Peter. Learning and Coordination:  Inductive Deliberation, Equilibrium and Convention. New York: Routledge, 2001. van Parijs, Philippe. “The Second Marriage of Justice and Efficiency.” Journal of Social Policy 19 (1990): 1–​25.   “Why Surfers Should Be Fed: The Liberal Case for an Unconditional Basic Income.” Philosophy & Public Affairs 20 (1991): 101–​31.   Real Freedom for All:  What (If Anything) Can Justify Capitalism? Oxford:  Oxford University Press, 1995. Varian, Hal. “Equity, Envy and Efficiency.” Journal of Economic Theory 9 (1974): 63–​91.   “Distributive Justice, Welfare Economics and the Theory of Fairness.” Philosophy & Public Affairs 4 (1975): 223–​47.   “Dworkin on Equality of Resources.” Economics & Philosophy 1 (1985): 110–​27. Vickrey, William, “Measuring Marginal Utility by Reactions to Risk.” Econometrica 13 (1945): 319–​33. Vizard, Timothy and Gareth Rusgys. Personal Well-​Being Survey User Guide 2012–​2013 Dataset. UK Office of National Statistics, 2013. von Mises, Ludwig. Human Actions:  A  Treatise on Economics. New Haven, CT:  Yale University Press, 1941. von Neumann, John and Oskar Morgenstern. Theory of Games and Economic Behavior, 2nd ed. Princeton, NJ: Princeton University Press, 1947. von Wright, Georg. 1971. Explanation and Understanding. Ithaca, NY:  Cornell University Press. Waldron, Jeremy, ed. Theories of Rights. Oxford: Oxford University Press, 1984.   The Right to Private Property. New York: Oxford University Press, 1990.

384

References

Walsh, Vivian. “Rationality as Self-​Interest vs. Rationality as Present Aim.” American Economic Review Papers and Proceedings 84 (1994): 401–​5. Walzer, Michael. Spheres of Justice. New York: Basic Books, 1983. Weber, Max. “Science as a Vocation.” (1919). Trans. H. H. Gerth and C. Wright Mills. In From Max Weber: Essays in Sociology. New York: Oxford University Press, 1946, pp. 129–​56. Weibull, Jörgen. “Persistent Unemployment as Subgame Perfect Equilibrium,” Seminar Paper No. 381 of the Institute for International Economic Studies, Stockholm, May 1987. Weitzman, Martin L. “On Modeling and Interpreting the Economics of Climate Change” Review of Economics and Statistics 91 (2009): 1–​19. Wellman, Carl. A Theory of Rights: Persons under Laws, Institutions and Morals. Totowa, NJ: Rowman & Allenheld, 1985. Wenar, Leif. “The Nature of Rights.” Philosophy & Public Affairs 33 (2005): 223–​53.   “The Nature of Claim Rights.” Ethics 123 (2013): 202–​29. Wertheimer, Alan. Exploitation. Princeton, NJ: Princeton University Press, 1996. Weston, Samuel. “Toward a Better Understanding of the Positive/​Normative Distinction in Economics.” Economics and Philosophy 10 (1994): 1–​18. Weymark, John. “A Reconsideration of the Harsanyi-​Sen Debate on Utilitarianism,” in Elster and Roemer, eds. (1991), pp. 255–​320. White, Brent T. “The Morality of Strategic Default.” UCLA Law Review 58 (2010): 155: 155–​64. Wight, Jonathan. Ethics in Economics: An Introduction to Moral Frameworks. Stanford, CA: Stanford University Press, 2015. Wilkinson, Richard and Kate Pickett. The Spirit Level. London: Bloomsbury Press, 2009. Wilkinson, T. “The Ethics and Economics of the Minimum Wage.” Economics and Philosophy 20 (2004): 351–​74. Williams, Bernard. “The Idea of Equality,” in Peter Laslett and William Runciman, eds. Philosophy, Politics and Society. New York: Barnes & Noble.   “A Critique of Utilitarianism,” in Smart and Williams (1973), pp. 77–​150. Winch, Peter. The Idea of a Social Science. London: Routledge, 1958. Wolff, Robert. Understanding Rawls: A Reconstruction and Critique of a Theory of Justice. Princeton, NJ: Princeton University Press, 1977. Wolff, Jonathan. An Introduction to Political Philosophy. Oxford:  Oxford University Press, 2006. World Bank. Breaking the Conflict Trap:  Civil War and Development Policy. Oxford: Oxford University Press, 2003. Yaari, Menachem and Maja Bar-​Hillel. “On Dividing Justly.” Social Choice and Welfare 1 (1984): 1–​24. Yellen, Janet. “Efficiency Wage Models of Unemployment,” American Economic Review: Papers and Proceedings 74 (1984): 200–​5. Young, H. Peyton. Equity. Princeton, NJ: Princeton University Press, 1995.

Index

abortion, 8–​11 Ackerman, B., 220, 361 Adler, M., xix, 118, 164, 171, 220, 256–​59, 265–​66 Admati, A., 16 adverse selection, 318–​20. See also insurance Affordable Care Act (aka “Obamacare"), 196 aggregation, 251–​52, 255, 265, 358 Agincourt, 131 AirBNB, 320–​21. See also gig economy Akerlof, G., 82–​83, 90, 361. See also partial gift exchange Alexander, S., 231 Allais, M., 65, 67 Allais’s problem, 65t. 4.3. allocations, 149, 285 envy-​free, 210, 245 income fair, 212, 214 Alstott, A., 220 altruism, 75–​77, 84, 90, 101, 186, 285, 346–​48 and blood donation, 100, 104, 345–​48. See also markets: in blood effective altruism, 113 Ames, R., 78, 281, 349 Anand, P., 143 Anderson, E., xix, 99, 104, 121, 204, 215, 220 Anomaly, J., xix anonymity, 257 Aquinas, T., 93, 183 Aristotle, 67, 93, 142 Arneson, R., 143, 202, 204–​05, 208–​09, 220 Arrow, K., 76, 87, 90, 104, 114, 137, 256–​58, 262, 266, 287, 346 impossibility theorem, 247–​54, 256, 263, 265. See also social choice theory Asian disease, 59, 60t. 4.2., 66, 347

assurance game, xiii. See games Atkinson, A., 16, 195, 218–​20 autonomy, 177, 187–​88, 190, 310, 325 condition versus capacity, 177 See also freedom Autor, D., 7 Axelrod, R., 279, 289 Baby Jessica, 121–​23 Bacharach, M., 286 backward induction, 281–​82, 289 Baker, C., 164, 171 Bangladesh, 329 Barbera, S., 266 bargaining theory, 67, 273, 285–​86, 288–​89 Nash solution, 286 and political philosophy, 288 See also justice Bar-​Hillel, M., 285 Barone, E., 159 Barry, B., 225, 241, 289 Bartels, L., 196–​97 Barushok, C., xix Batson, C., 85, 90 Baumol, W., 210–​11 Bayesian theory, 61, 65 Bear Stearns, 4 Becker, G., 90, 119 Ben-​Ner, A., 90 Bentham, J., 59, 110, 123 Bentham’s dictum, 110 on intrinsic good, 127 Berg, J., 73 Bergson, A., 247 Berlin, I., 178, 190 on negative and positive liberty, 177–​78 Beyoncé, 182

385

386 Bicchieri, C., 90, 282, 286, 289 Binmore, K., 67, 241, 288–​89 Blackorby, C., 166, 171 Blinder, A., 16, 83 Block, W., 188 Blount, S., 284 Boadway, R., 171 Bonnano, G., 289 Boonin, D., 335 Bossert, W., 267 Boston Red Sox, 136 Boulding, K., 90 Bowen, W., 301 Bowles, S., 81, 83, 90, 221 Braithwaite, R., 289 Brandt, R., 124, 134 Braybrooke, D., 118 Brennan, G., xix, 86, 90, 228, 237 Brennan, T., 90 Brenner, J., 190 Brighouse, H., 37, 241, 310 Brock, D., 240–​41 Broome, J., 66, 68–​69, 111–​12, 124, 144, 171, 266, 331, 335 principle of the personal good, 150, 247 Buchanan, A., 36 Buchanan, J., 228, 232, 236–​38, 240–​41 Bumble, Mr., 186 burqas (Muslim headscarves), 89 Calabresi, G., 182 Camerer, C., 69, 86, 191, 290 Caney, S., 335 capabilities, 117, 141–​42, 144, 179, 208, 215, 236, 285, 305–​06 and elementary education, 307 versus welfare, 142 See also Nussbaum, M.; Sen, A. Carter, I., 266 causal decision theory, 62 Chakravarty, S., 335 charter schools, 35, 37 Chetty, R., 199 children, 32–​34, 59, 76, 87, 100, 122–​23, 141, 157, 223, 233, 304, 310–​11 child labor, 101–​02 and inequality, 199–​200, 311 and paternalism, 32–​33 picked up late from day care, 100 and school vouchers, 308 See also vouchers (school) Christman, J., 190

Index Chu, R., 73 Chu, Y., 73 Chubb, J., 37 Churchland, P., 89 civil engineering, 338 Clayton, M., 220, 241 climate change, xvi, 72, 131, 168, 314–​15, 328–​33, 335 as an externality, 95 and future generations, 111, 329–​30 and the “Non-​Identity Problem”, 331 and non​human animals, 333 papal encyclical, 154 and welfarism, 328 See also justice: and climate change Coase, R., 96, 153, 171, 185, 191 coercion, 129, 138, 187, 189–​90, 234, 298 and paternalism, 189 Cohen, G. A., 88, 94, 202, 205, 208, 215, 220, 228, 239–​40 on “access to advantage”, 214 Colander, D., 296 Coleman, James, 85–​86, 90 Coleman, Jules, 171, 182 Collard, D., 90 commitment, 76, 87, 226, 230, 282, 285, 293, 310 social, 299 See also morality: moral commitment compensation principle, 212–​14. See also Fleurbaey, M. completeness, 57–​58, 61–​62, 251, 256 conscience, 85, 179 liberty of, 235 consequentialism, 109, 117, 121–​22 versus deontology, 121 See also utilitarianism constitutional choices, 236 constitutional design, 237 constitutionality, 113 versus justice, 9 constraints (agent-​centered), 121 consumer choice theory, 70, 155 continuity, 58, 61–​62, 258 contractualism, 15, 174, 223, 225–​26, 236, 238 coordination, 184, 289 problems of, 241 pure, 276 See also games Cooter, R., 230 cost–​benefit analysis, 37, 152, 159, 161, 165, 167–​70, 302, 304 and justice, 164

Index limits, 163–​64 and policy, 167–​68, 170 and utilitarianism, 165–​66 cost-​minimizing principle, 311 Cowen, T., 171, 191 Cozic, M., xix d’Aspremont, C., 255, 266 Daniels, N., 231–​32, 241, 319, 335 Davidson, D., 71–​72, 89, 171 Dawes, R., 281, 286 De Vroey, M., 43, 51, 83 death, 30, 74, 122, 296 Deaton, A., 232 Debreu, G., 59 debt, 3 Greek debt crisis, 339 defined-​benefit pensions, 321 DeLap, J., xix democracy, xvi, 96, 100, 119, 185, 197, 234, 254 democratic sovereignty, 254 Demsetz, H., 171, 191 Dennett, D., 89 deregulation of financial markets, 2 Devine, J., 90 DeVroey, M., 42 Diamond, P., 258, 300 Dickens, C. Oliver Twist, 186 difference principle, 230, 232, 239–​40, 255, 266 discount rate, 331 discounting, 330 discursive dilemma, 263t. 13.5. distributional weights, 166–​69, 304–​05, 328 distributions contribution fair, 211 income fair, 211 wealth fair, 211–​12 See also allocations distributive concerns, 32–​33, 161, 216, 238, 330–​31 See also allocations; justice: distributive divorce, 223 Dokko, J., 335 Doman, L., xix Donaldson, D., 166, 171 Doolittle, Alfred P., 157n. 2 Dretske, F., 89 Drèze, J., 171 Duncan, G., 6, 16 Dworkin, G., 190 Dworkin, R., 137, 191, 196, 204, 208–​10, 213, 220

387

on the “slavery of the talented,” 212 Dyson, F., 279 economic models, 75, 77 Eells, E., 62 efficiency, 6, 13, 31, 34–​36, 49, 87–​88, 90, 93–​96, 101, 148–​50, 154, 159–​61, 163, 169, 175, 182, 218–​19, 221, 233, 280, 295, 301–​04, 306, 310, 316, 320, 345 See also Pareto efficiency efficiency wages, 81–​83, 90 eflornithine, 30, 200 egalitarianism, 15, 98, 174, 194, 201–​06, 208, 210, 214–​17, 219–​20, 232–​33, 236–​37, 239, 241, 245, 254, 257–​59, 264–​67, 285, 310–​11 basic, 201–​04 egalitarian equivalence, 214 “luck”, 204, 212, 215 relational, 216–​17 resource, 209–​10 welfare, 255 See also equality; justice; prioritarianism Ellsberg, D., 65 Elster, J., 88, 90, 138, 170, 241, 266, 274, 289 employment, 2–​3, 44–​45, 195, 299, 325, 344 and equal opportunity, 216 job security, 35 See also unemployment enemy combatants, 113 environment, 20–​21, 23, 80, 99, 129, 154, 167–​68, 198, 204, 306, 314, 316, 327, 330, 333–​34 as a commodity, 24, 29 and inequality, 302 object of moral concern, 80 and policy, 168, 328 protection, 2 See also climate change; externalities equal respect, 116–​17, 196, 204, 207–​08, 215, 217, 310 as basis for moral theory, 217 and utilitarianism, 116, 225 equality, 15, 22, 133, 146–​47, 169, 173–​74, 197, 201–​03, 206, 208, 212, 214–​18, 295, 304, 309–​10, 316, 335 as basis for moral theory, 196 of capabilities, 215 conditional, 213–​14 democratic, 215 in debates over economic policy, 196 versus economic growth, 219 and fairness, 207, 220

388

Index

equality (cont.) political, 101 of opportunity, 6, 32, 195, 197, 208–​09, 216, 220, 223, 230–​31, 238, 319 of resources, 208, 214 and self-​respect, 207 of wages, 88 of welfare, 208, 342 See also egalitarianism; inequality equilibrium, 42, 50, 176, 283 competitive, 147–​48, 150, 209, 211, 300 in labor market, 42–​43, 83 Nash, 273–​74, 282–​84 perfect, 283 in rational expectations, 47 reflective, 226, 240, 256 equilibrium theory, 47, 93 equity, 150, 164, 166–​67, 170, 210, 289, 303, 328, 335 estate tax, 110, 219 ethics. See morality Etzioni, A., 77, 90 evolutionarily stable strategy, 279–​80 evolutionary psychology, 87, 90 exit and voice, 103 experience machine, 140 exploitation, 28, 37, 208 of irrationality, 189 externalities, 95–​96, 148, 153, 158, 171, 185, 323 negative, 83, 96 pecuniary versus non-​pecuniary, 153n. 1 fact–​value distinction, 12–​14, 51, 89, 341–​43 See also normativity fairness, 4, 111, 126, 137, 158, 164, 208, 210, 241, 258, 280, 284, 288, 295, 303, 307, 310 procedural, 248 Fankhauser, S., 164 Fehr, E., 86, 290 Feinberg, J., 177, 190 financial crisis of 2007-​09, xv, 5, 7, 16 Finland traffic fines, 207 Fisch, E., 84 Fleming, M., 67, 266 Fleurbaey, M., xix, 203, 210, 212, 215, 220, 266–​67 Foell, W., xx Foley, D., 210 folk psychology, 55–​57, 71, 89 401-​Ks, 321

Frank, R., 75, 89, 203, 349 Frankfurt, H., 77, 90, 220 fraternity, 208, 217 freedom, 13, 15, 22, 31, 34, 36, 46–​47, 94–​96, 98, 100, 107, 121, 138, 147, 157–​58, 169, 173, 175–​76, 178–​80, 184–​90, 208, 218–​ 19, 224, 235–​36, 259, 261–​63, 265–​66, 295–​96, 304–​05, 309–​10, 316, 335, 346 and competitive markets, 178 effective , 177–​78, 261–​62, 316 and flexibility, 262 four conceptions summarized, 179 justification, 183 positive and negative, 177–​78, 240, 302 as self-​determination, 177 social conditions of, 215–​16 See also Berlin, I.; libertarianism free-​rider problem, 76 Frey, B., 101, 104 Friedman, M., 31–​32, 37, 175, 183, 342 Frohlich, N., 289 Fuchs, V., 335 functional explanations, 88 functionings, 117, 141–​42, 144, 215. See also capabilities Gaertner, W., 260, 266 Gambetta, D., 90 game theory, 87, 118, 238, 245, 261, 271, 274, 280, 282–​86, 289 cooperative and noncooperative, 254, 273, 282 evolutionary, 271, 279–​80 and norms of fairness, 280 See also equilibrium Game theory, 269 games, 269, 271–​72, 274–​75, 278–​79, 281, 283, 286, 335 assurance (aka “stag hunt"), 277–​78 bargaining, 287 “battle of the sexes”, 276 binary lottery, 287 “chicken”, 275, 278 cooperative and noncooperative, 289 and ethics, 274 extensive form, 269–​71 game forms, 261 “game of life”, 288 “game of morals”, 288 mixed coordination–​conflict, 276, 289 prisoner’s dilemma, 273–​77, 279, 281–​82

Index prisoner’s dilemma (iterated), 279, 281–​82, 288 repeated, 279, 288 of social cooperation, 274 ultimatum, 284 Gärdenfors, P., 68, 260, 266 Gates, B., 199 Gaus, G., 36, 241 Gauthier, D., 67, 69, 131, 143, 236, 240–​41, 290 conception of justice, 236, 288 Gevers, L., 255, 266 Gibbard, A., 54, 62, 134, 171, 260, 266, 288 gift exchange, 82 partial, 82 See also Akerlof, G. gig economy, 320–​22, 324–​27 Gilbert, M., 286 Gilead Parmaceuticals, 200 Gilens, M., 196–​97 Gilovich, T., 349 Gini coefficient, 218–​19 Gintis, H., 37, 81, 83, 90 Glaister, S., 165–​66, 171 global warming, 60, 67, 154, 209, 251, 327–​28. See also climate change; environment globalization, 6 Glover, J., 124 Gneezy, U., 100 Goldfarb, R., 90 Goodin, R., 131, 143 Gorman, W., 162 Great Depression (1930s), 2, 5–​6, 42, 44 Great Recession (2007-​2009), 5. See also financial crisis of 2007-​2009 Grether, D., 72, 89 Griffin, J., 69, 115–​16, 124, 131, 141, 143, 196, 217 conception of utilitarianism, 117–​18 Griffith, W., 90 growth (economic), 2, 16, 48–​49, 102, 110, 198–​99, 219, 252, 299, 339 and the environment, 183, 314 Gutmann, A., 35, 37 Hagen, O., 65 Hall, J., 325, 335 Hamlin, A., 36 Hammond, P., xix, 144, 151, 171, 266 Hampton, J., 69, 241, 275, 278, 289 Hanser, M., 332 happiness, 15, 109–​10, 112, 119, 127, 131, 136–​ 39, 196, 206, 312. See also preferences; utility; welfare

389

Hardin, R., 120, 124, 153, 274, 278, 289 theory of utilitarianism, 118 Hare, R., 120, 124 Harper, W., 62, 283 Harris, R., 144 Harris, S., 324, 335 Harsanyi, J., 63, 67–​68, 115, 117–​18, 124, 137, 171, 228, 236, 240–​41, 255, 257–​58, 265–​66, 289 Hart, H., 191 Hausman, D., xx, 51, 78, 89–​90, 144, 171, 191, 328 Hayek, F. von, 95, 175, 185, 191, 201 health insurance, 316–​20 universal health care, 320 See also vouchers (health care) hedonism, 127, 131, 138–​41 Hellwig, M., 16 Henry V, 131 Herstein, I., 63 Hicks, J., 159–​62, 169 Higgins, Henry, 157n. 2 Hirsch, F., 203 Hirschman, A., 94, 348 on “exit and voice”, 103 Hobbes, T., 236, 241 on life in “state of nature”, 274 on rationality and self-​interest, 225 Hochman, H., 76, 171, 183 Hohfeld, W., 180, 190 Hollis, M., 68 Holmes, M., 84 Homans, G., 82 homelessness, 135, 176, 179 Homo economicus, 79, 92. See also self-​interest Hooker, B., 25, 113, 124 Hornstein, H., 84 housing bubble of 2007, 2. See also financial crisis of 2007-​2009 human development index, 143–​44 Hume, D., 68, 225, 237 on “the circumstances of justice”, 224 against contractualism, 225 Hurley, S., 68–​69, 286 Hylland, A., 266 ideal theory, 238. See also Sen, A.: on justice ideology, 341 immortal souls for sale on Ebay, 93 immunities, 180–​81. See also rights impartiality, 207, 237–​38, 257 as basis of justice, 225

390

Index

in-​kind benefits, 136, 154, 156–​57, 171, 175 and paternalism, 157 independence condition, 61–​62, 66–​67. See also preferences; uncertainty: versus risk independence of irrelevant alternatives, 249–​ 50, 253, 287–​88 indifference curves, 154, 156 individualism (about value), 22 inequality, xv, 147–​48, 150, 173, 194–​96, 199, 201–​02, 206, 208, 214, 217–​19, 277, 310 educational, 320 of educational opportunity, 16 of income, 198 of power, 324 of wealth, xvi, 2, 5–​6, 16, 98, 195, 308. See also Atkinson, A.; Piketty, T.; Stiglitz, J. InstaKart, 320. See also gig economy insurance, 3, 40, 53, 103, 154–​57, 210, 317–​20, 323 hypothetical, 210 unemployment, 324 See also health insurance interpersonal comparisons, 146, 164–​65, 170, 210, 287–​88, 306, 328 of preference intensity, 256 of preference satisfaction, 159 of well-​being, 113–​15, 119, 123, 146, 236 See also sympathy: extended intrinsic versus instrumental goods, 126–​27 irrationality, 29, 57, 66–​69, 73, 118, 177, 282 Jaxen, J., xix Jones, P., 266 justice, 6, 15, 36–​37, 79, 99, 126, 136–​37, 141, 146–​47, 149, 164, 169, 173–​74, 179, 187–​ 88, 197, 215, 223–​24, 227–​41, 288, 295, 311–​12, 335 and climate change, 328–​30, 332 and contractualism, 224–​25, 227, 236, 238 criminal, 93, 99 distributive, 104, 186, 188, 197, 199, 206, 223, 239–​40, 278, 285, 329, 332 and freedom, 185 global, 228 ideal, 239–​40 as mutual advantage, 238 prioritarianism, 203–​04, 206, 208, 215, 220, 258, 266, 328 racial, 199 Rawls’s two principles, 229, 232–​35, 238, 312, 319 as reciprocity, 226

and rights, 185, 188, 224 social, 102, 196 and utilitarianism, 109, 117–​18, 224, 226, 236 versus welfare, 28 See also bargaining theory; Nozick, R.; Rawls, J. Kadane, J., 151 Kagan. S., 124 Kahneman, D., 59, 66, 69, 89, 139, 141, 144, 347 Kalai, E., 287 Kaldor, N., 159–​62, 169 Kamm, F., xix, 122, 124 Kant, I., 183, 234, 241, 274 on justice as impartiality, 225 Kavka, G., 241, 335 Kay, E., xix Keats, J., 115 Kelleher, P., xix Kelman, S., 90, 154, 171 Keynes, J. M., xv, 40–​42, 45–​47, 143 and macroeconomics, 40 on macroeconomics, 42 on unemployment, 47 Kittay, E., 231, 241 Knight, F., 60 Kolm, S., 114, 119, 209, 212, 267 Kraus, J., 182 Kraut, R., 144 Kreps, D., 289 Kreuger, A., 144, 324 Krueger, A., 139, 325, 335 Kuczmarski, M., xix Kukathas, C., 241 Kymlicka, W., 196, 241 labor, 41–​44, 103, 231, 239, 264, 321, 326 agreements, 181 child, 101–​03 division of, 94, 146, 239 labor power, 211 law, 324 reproductive, 97, 104 sweat shop, 1, 208 wages and purchasing power, 42 See also employment; markets: labor; morality: and labor; unemployment Lady Chatterley’s Lover, 260 laissez faire economics, 46 Lancaster, K., 159 Laslier, J.-​F., 266

Index Lawrence, D. H., 260 Layard, R., 144, 165–​66, 171 Le Grand, J., xix, 78, 148, 166, 190–​91, 204, 209 Lehman Brothers, 4 Lemann, N., 321, 335 Lensberg, T., 289 Lerner, A., 49–​50, 300, 302 Levi, I., 57, 65, 67, 152, 171 Lewis, D., 62, 68, 276 liability rules, 182 liberal reward principle, 212–​14. See also Fleurbaey, M. liberalism, 135, 310 minimal, 260 libertarianism, 15, 174–​75, 178, 184–​85, 188, 227, 235, 240, 309, 321 “vulgar”, 191 See also Hayek, R.; Nozick, R. Lichtenstein, S., 69, 72, 89 life expectancy, xvi, 194 disparities, 201, 231 Lipsey, R., 159 List, C., 263, 266 Locke, J., 183, 241 on moral intuition, 183 on the social contract, 224 Lomasky, L., 185, 187, 191 Loomes, G., 68 Lowenstein, G., 69 Lucas, R., 5, 40–​41, 44–​46, 198 Luce, R., 289 Lukes, S., 207 MacAskill, W., 113 MacCallum, G., 176–​77, 179, 190, 261, 296 Machan, T., 185, 191 Machina, M., 68, 144 Machlup, F., 339 MacKay, A., 115, 171, 266 Madison, J., 237 malevolence, 75 Malouf, M., 288 Maniquet, F., 212, 215, 220, 267 Mankiw, G., 199 Mansbridge, J., 90 Margolis, H., 77 markets, 14, 23, 25, 29, 31, 34, 36, 40, 42, 48, 50, 54, 92–​97, 101, 103–​04, 113, 148–​49, 159, 183, 199, 207, 234, 299, 321, 333 in blood, 346 commodity, 41 and corruption, 99 effects on behavior, 101

391

in health care, 319–​20 in human organs, 97–​99 and information asymmetry, 87, 95 labor, 40–​43, 45, 81, 83, 98, 300, 335, 344 moral limits, 8, 30, 36, 92–​100, 102–​04, 153, 200, 300 never completely free, 92 perfectly competitive, 153, 300 in pollution limitation, 23, 154 repugnant, 97–​98 virtues of, 94, 100, 154, 178 See also monopoly Marley, J., 126 Marmot, M., 231–​32 Martin, R., 180, 191 Marwell, G., 78, 281, 349 Marx, K., 94 Mason, A., 113, 220 maximin principle, 230, 333 McClennen, E., 60, 65, 68 McDowell, J., 67–​68 McGuire, T., xix McKean, R., 90 McPherson, M., xx, 61, 90, 134, 144, 328 Meade, J., 241 Mechanical Turk (Amazon), 321, 324, 326. See also gig economy Meckling, W., 49–​50, 300, 302 Melamed, D., 182 Melden, A., 89 mental shoehorn, 115 microeconomics, 45, 71–​72, 83 and folk psychology, 71 Mill, J. S., 96, 110, 116, 123, 127, 144, 181, 189, 234 against paternalism, 96, 189 See also utilitarianism Miller, A., 94 Miller, D., 113, 201–​02, 206–​08, 216 Milnor, J., 63 minimal benevolence, 147–​51, 248 minimum wage, 95, 100, 194, 296, 324–​27 and effective liberty, 325 Mishan, E., 161, 171 Moe, T., 37 money pump argument, 57 Mongin, P., 151, 171, 255, 266, 341 monopoly, 324n3, 159 moral hazard, 3, 5, 16, 317–​19 morality, 1–​2, 6, 8–​9, 11–​12, 14–​15, 26, 53, 70, 74–​75, 77–​81, 88–​89, 92, 110, 113,

392

Index

morality (cont.) 116, 120, 123, 126, 217, 234, 245, 250, 257, 279, 341, 344, 348 “moral mathematics”, 245, 248, 250, 258 blame, 74 dogmatism, 11 and the environment, 327, 329–​33 and freedom, 177–​79 and game theory, 286, 288 guilt, 74, 85 and inequality, 194–​97, 201–​02, 206, 215, 217 and involuntary unemployment, 40, 45, 47, 296–​98 Kantianism, 274 and labor, 324–​27 moral commitment, 77–​78, 88–​89, 173, 235, 276, 285, 288, 337, 340–​41, 344, 347, 349 moral desert, 110, 197, 200–​01, 206 moral judgments, 10, 12, 34, 39, 226, 338 moral norms, 74, 77, 81–​88, 224, 288, 297, 299, 301. See also normativity: influence on behavior moral relativism, 10–​11 and nonhuman animals, 111, 166 and normativity, 53 and preference, 54 and rationality, 53, 74–​81, 118, 245, 254 resentment, 74 rights-​based, 118, 180–​81, 183–​84, 187 shame, 74 and welfare economics, 303 See also contractualism; justice; markets: moral limits; normativity; utilitarianism Morgenstern, O., 63 Moulin, H., 220 Mozart, W. A., 200 Mueller, D., 80 Murdoch, A., xix Murnane, R., 6, 16 Murnighan, J., 288 Murphy, L., 110, 187, 191, 228 Musgrave, R., 231 mutual advantage, 225–​26, 236, 238. See also justice Nagel, T., 67–​68, 124, 136, 184, 187–​88, 191 Narveson, J., 185, 188, 191 Nash equilibrium. See equilibrium: Nash Nash, J., 274, 286, 288–​89. See also bargaining theory; game theory; games Nehring, K., 267

New, B., 191 Newcomb’s problem, 62 Ng, Y.-​K., 171 no-​envy condition, 209 nondictatorship principle, 248 nonidentity problem. See also climate change Nordhaus, W., 331 normativity, 39, 45–​46, 48–​50, 55, 72, 83, 100, 143, 147, 224, 226, 232, 238, 247, 254, 272, 300, 302, 339–​40, 342–​43 versus descriptivity, 53 and economics, 15, 19–​23, 25–​26, 28, 36, 70, 73, 107, 118, 123, 128, 141, 146–​47, 150, 154, 158, 169, 173, 175–​76, 240–​41, 248, 261, 302, 309, 312, 327–​28, 334, 337–​38, 342, 349–​50 influence on behavior, 81–​88, 282 normative economics, 39 and political theory, 252–​53 and rationality, 59, 65 of rationality, 14, 70 and the World Bank memorandum, 303 See also fact–​value distinction; morality Nozick, R., 62, 100, 140, 183, 188, 191, 260, 266 on libertarianism, 185, 188, 309, 321 on rights, 184, 188 See also justice; Newcomb’s problem Nussbaum, M., 85, 99, 119, 144, 241 on capabilities, 142, 215, 231, 305 O’Brien, J., 73 O’Neill, M., 217 Oberholzer-​Gee, F., 101 Occupy Wall Street movement, 5, 195 Okin, S., 241 Okun, A., 51, 171, 216, 218 Olewiler, N., 144 Orbell, J., 281 original position, 227–​28, 234–​35. See also Rawls, J.: theory of justice; veil of ignorance Osborne, M., 289 Otsuka, M., 187 overlapping consensus, 234–​35 overlapping generations, 39, 48, 51, 295, 299, 344 Overvold, M., 132 pain-​minimizing principle, 311 paradox, 280, 282, 285

Index paradox of the Paretian liberal, 259–​60, 266. See also Sen, A. Pareto efficiency, 147–​48, 150, 158, 176, 306 Pareto improvement, 147–​48, 150–​52, 156, 158, 159t. 9.1, 159–​63, 306, 308 potential, 160–​62, 247, 302 Pareto optimality, 264, 286 Pareto optimum, 147–​49, 260 Pareto principle, 150–​52, 164, 171, 247–​48, 250–​51, 253, 255, 257–​60, 264 Pareto, V., 159 Parfit, D., 68, 90, 111, 113n. 1, 118, 124, 132, 134, 141, 202, 220, 276, 289, 331, 335 on “future Tuesday indifference”, 67 and population control, 112 principle of equality, 202 Park, J., xix paternalism, 32, 129–​30, 157, 190–​91, 317 and libertarianism, 100, 321 libertarian, 189 See also children: and paternalism; libertarianism; Mill, J.S. Pattanaik, P., 260, 262, 266 Paul, E, 290 Paxton, W., 220 PayPal, 321. See also gig economy pay-​your-​way principle, 311 Pazner, E., 211 perfect competition, 50, 148–​49, 158, 300. See also Pareto efficiency perfect equilibrium, 283f. 14.10., 283–​84. See equilibrium perjury, 120 Peter, F., 51 Pettit, P., 85–​86, 90, 178, 188, 190, 241, 263, 266, 288–​89 Phelps, E., 231 Pickering, Col., 157n. 2 Pickett, K., 231 Piff, P., 201 Pigou, A. C., 114, 257–​59, 266 Pigou-​Dalton principle, 257–​59, 266 Piketty, T., 5–​7, 16, 195. See also inequality: of wealth Pinker, S., 90 pizza, frozen, 55 Plamenatz, J., 123 Plato, 67, 224 Plott, C., 72, 89 pluralism (reasonable), 234–​35 Pogge, T., 188, 241 Pojman, L., 220

393

Polanyi, K., 94 political theory, 144, 226, 234, 252, 315 pollution, 19–​20, 23–​25, 27–​29, 31, 36, 79–​80, 83, 96, 98, 153–​54, 175, 305–​07, 309, 323, 327–​28 as a commodity, 24 transfers, 21–​23, 25, 28–​29, 302–​03, 305, 307–​11 positional goods, 203–​04 positive affine transformation, 63–​64, 287 Posner, R., 171 poverty, 28–​29, 76, 102, 130, 179, 219, 232, 305, 310 extreme, xvi powers, 181 moral, 229 and rights, 180 separation of, 225 precautionary principle, 332 preference satisfaction, 15, 25, 28–​29, 39, 73, 123, 129–​34, 136, 138–​41, 143, 146–​50, 152, 156–​63, 167, 208, 229, 231, 302–​06, 312, 328, 335. See also welfare constitutive versus evidential views, 128–​31, 133–​38 and welfare, 127–​32, 134, 136–​37, 139, 141, 143, 150–​51, 158–​59, 166, 169, 189, 256, 303–​04, 307, 342 preferences, 31, 56–​59, 61–​66, 68, 71–​73, 75–​77, 79, 88, 90, 94, 109, 115, 119, 123, 128–​29, 131–​35, 139, 141, 143–​44, 147, 151, 155–​56, 163–​64, 168, 170, 177, 179, 209, 212, 214, 227, 229, 231–​32, 236, 247, 249–​56, 259–​60, 262–​64, 270–​72, 274, 280, 282, 284–​85, 304, 306–​08, 317, 334, 344–​46 and abilities, 205 antisocial, 137 complete, 57 and cost–​benefit analysis, 164, 167, 169 counterpreferential choice, 77 expensive, 136–​37, 209, 231 extended, 114–​15 formation of, 138, 304 of future people, 330 and happiness, 110, 127 informed, 118 and (ir)rationality, 55–​56, 59–​60, 64, 67–​68, 70, 74, 131 of Lewd and Prude, 260 and market prices, 25

394

Index

preferences (cont.) metapreferences, 77 nonegoistic, 76, 78, 129 and other people, 136, 177 preference reversal, 72–​73, 89 racist, 137 revealed, 57–​58 self-​regarding, 131–​32, 135 transitive, 56 and utility, 59 of workers, 43–​44 See also morality: preferences; utility theory Preiss, J., xix prerogatives (agent-​centered), 121 Press, W., 279 primary goods, 141–​42, 229–​31, 239, 305–​06 social, 229 See also Rawls, J. principle of personal good, 151–​52 prisoner’s dilemma. See games probability, 60, 62, 78, 151, 156, 229, 270, 279, 287, 332–​33 calculus, 64 distributions, 164 subjective, 61 See also games Proulx, C., 288 prudence, 53, 77 public funding, 31 education, 31–​32 public goods, 83, 90, 101, 187, 233, 275 experiments, 86 Puppe, C., 267 Putnam, H., 89 Putnam, R., 16 Putterman, L., 90 Pygmalion, 157n. 2 Queen Victoria, 115 racism, 138, 298 Radin, M., 99, 104, 216 Raiffa, H., 289 Railton, P., 124 Ramsey, F., 63 Rand, A., 186 rational expectations, 44. See also Lucas, R. rationality, 14, 53–​55, 57, 59–​61, 66–​68, 72, 74–​75, 81, 89, 121, 149, 171, 263, 277, 282, 335 collective, 249–​50, 253, 257–​60 as conceived by economists, 14 and consumer choice, 70

and deontology, 122, 124 formal conditions, 56, 245 and game theory, 271 and markets, 94 of morality, 14, 53 and motivation, 189, 225 reasons versus causes, 71–​72, 89 and self-​interest, 68, 73–​75, 78, 80–​81, 90, 225–​26, 245 theory of rational expectations, 88 and utility theory, 65, 68, 89 See also irrationality; morality; normativity Rawlings, J. K., 199–​200 Rawls, J., 117, 144, 215, 220, 224, 241, 266, 310 on basic liberties, 179 on primary goods, 141–​42, 305 on rules and utility, 120 theory of justice, 226–​40, 255–​56, 288, 303, 307, 312, 319 See also contractualism; difference principle; justice; original position; veil of ignorance Raz, J., 141, 187, 190–​91, 201 Read, C., 143 real income, 128, 161–​62 recession, 2–​3, 16 reciprocity, 117, 231, 238, 241, 276 Reder, M., 87 redistribution (of wealth), 96, 114, 185, 188, 210, 215, 218, 232–​33, 237–​38, 240, 316 Regan, D., 286, 349 Reich, M., 90 representation theorem cardinal, 62–​64 ordinal, 58 responsibility, 204–​06, 213, 298, 322–​23, 329 and autonomy, 177 individual, 209–​10, 214 moral and legal, 323–​24 and paternalism, 32 personal, 47 and prioritarianism, 204 of the state, 311, 316 Rice, A., xix rights, 15, 20, 22, 28, 102, 107, 117, 137, 146–​47, 154, 158, 166, 171, 173–​75, 178, 180–​81, 183–​86, 188, 190–​91, 196–​97, 216–​17, 219, 225, 232, 236, 249, 258–​61, 265–​66, 274, 285, 295, 297, 304, 309–​10, 316, 329–​30 to basic necessities, 183, 186–​87, 312 and consequentialism, 117, 126

Index of employees, 324 human, 183, 252, 328 legal versus moral, 180–​81 of nonhuman animals, 182 parenting, 100 to pollute, 96, 175 positive and negative, 187 property, 77, 92–​93, 95–​96, 100, 118, 153, 175, 182, 184–​88, 191, 224, 238, 240, 278, 312 rectification, 188 as “side constraints”, 184 and welfare, 184 Riley, J., 266 Risse, P., 241 Robbins, L., 146, 170 Robertson, D., 346 Rodgers, J., 76, 183 Roemer, J., xix, xx, 170, 205, 210, 220, 255, 264–​65, 267, 274, 285 Rogers, J., 171 Rosenberg, A., 89 Roth, A., 97, 288 Rousseau, J.-​J., 183, 241, 277 on the social contract, 177, 225 Rusgys, G., 144 Rustichini, A., 100 Ryan, A., xx, 190 Saez, E., 5, 7 Sagoff, M., 135, 334 Sahlin, N.-​E., 68 Samuelson, P., 39, 48–​51, 161, 247, 299–​300, 302 Sandel, M., 99, 104, 241 Satz, D., xx, 37, 92, 104 Savage, L., 63, 65 Scanlon, T., 135, 144, 157, 204, 206–​08, 220, 225–​26, 238, 241 Scheffler, S., 118, 121, 124, 217, 220 Schelling, T., 77, 90, 274, 276, 289 Schick, F., 57 Schmeidler, D., 211, 214 Schmidtz, D., 190–​91 Schon, D., 340 school choice. See vouchers (school) Schotter, A., 241, 289 Schultze, C., 31 Scrooge, E., 126 Seidenfeld, T., 151, 171 Seidl, C., 266

395

self-​imposed constraints, 77 self-​interest, 14, 50, 59, 73–​79, 83, 86, 89–​90, 92, 128–​30, 132, 141, 156, 167–​69, 236–​ 39, 275, 279, 300, 349 and contractarianism, 236 economic, 78, 83–​84, 86 and the “invisible hand”, 148 and libertarianism, 186 Selten, R., 283 Sen, A., 36, 65, 68–​69, 75–​77, 89–​90, 117, 119, 122, 124, 138, 141–​43, 191, 196, 215, 220, 231, 239–​40, 251, 254–​56, 259–​60, 262, 265–​66, 277, 285, 289 on capabilities, 142, 144, 179, 305 on justice, 238 separability condition, 257 service dogs, 322 sexism, 298 Shafer-​Landau, R., 16 Shapiro, C., 40, 45 Shaw, G. B., 157n. 2 Siberian tiger, 328, 333–​34 Sidgwick, H., 110, 116, 123, 127 on utilitarianism and rationality, 75 Singer, P., 112–​13, 124, 166, 330, 335, 345 Skidelsky, R., 47 Skyrms, B., 62, 280, 290 sleeping sickness, 30, 200 Slovic, P., 69, 72, 89 Smart, J. J. C., 124 Smith, A., 79, 173 on enlightened self-​interest, 346 on “the invisible hand”, 148 on the “invisible hand”, 275 on markets and liberty, 92, 94 Smith, J. M., 279 Smorodinsky, M., 287 Snowdon, B., 45, 296 social choice theory, 245, 247, 251, 254, 256, 258–​59, 261, 263, 265–​66 and morality, 254–​55, 265 See also Arrow’s theorem social contract theory. See contractualism Social Security, 49–​50, 301–​02, 324 social welfare function, 118, 165–​67, 169, 184, 220, 247–​54, 256–​59, 264 utilitarian, 253 See also Arrow’s theorem solidarity, 36, 83, 88, 303, 310, 335 undermined by inequality, 207 Solow, R., 45–​46, 83, 90 Solvaldi, 200

396 spleen tumor, 133 Stack, J., xix Stadler, J., xix standard of living, 2, 223 state of nature, 183, 224–​26, 274. See also contractualism Stern, N., 171, 329, 331 Stern Report, 330–​31 Stich, S., 89 Stiglitz, J., 16, 40, 45, 195 Stoehr, D., xix Stone, C., 166 stranger on a train, 132 Strasnick, S., 266 strategies, 269, 279 mixed, 270 pure, 270 tit-​for-​tat, 279, 281 See also game theory subjective degrees of belief, 61 Sugden, R., 68, 139, 144, 171, 241, 260, 263, 266–​67, 278, 286, 289 Summers, L., 19 Sumner, L., 181, 190–​91 Sunstein, C., 189–​91, 332 sure-​thing principle, 61. See also uncertainty: versus risk Sutton, Willie, 71, 79, 349 Suzumura, K., 215, 260, 266 sweatshops. See labor: sweatshop Swift, A., 241, 310 sympathy, 76, 132, 285 extended, 114, 256 Tadenuma, K., 215 Tan, K.-​C., 241 Tawney, R., 207, 217 tax compliance, 2, 79 taxes, 153, 158, 175, 186, 219, 233 cheating, 74, 78 on greenhouse emissions, 154 and incentives, 7, 197 inheritance, 185 and “pay it forward" college financing, 301 property, 311 and public education, 33 Taylor, M., 90, 241, 278–​79, 289 Temkin, L., 202, 206, 208, 218, 220 Ten Commandments, 84 Thaler, R., 69, 89, 139, 144, 189–​91 Thompson, F., xix Thomson, G., 119, 220 Thomson, J., 190–​91 Thomson, W., 215, 220, 289

Index threat point, 286–​87 Thurow, L., 171 Tiny Tim, 208–​09, 214 Titmuss, R., 100, 104, 345–​48 tolerance, 11–​12, 34 Tomasi, J., 191, 241 transitivity, 58, 61–​62, 68, 250 Trump, D. right to Trump Tower, 180 right to vote, 180 Tuomela, R., 286 Tversky, A., 59, 66, 69, 89, 347 “U-​index” (misery index), 139 Uber, 314, 320–​27. See also AirBnB; gig economy uncertainty, 53, 61–​62, 65, 87, 151, 163, 168, 171, 227–​28, 233, 237, 281, 330, 332 versus risk, 60–​61, 67, 333 undominated diversity, 220 unemployment, 2, 42–​44, 47, 51, 83, 88, 184, 233, 298–​99, 339 involuntary, 39–​47, 51, 70, 81, 83, 295–​99, 344 See also employment; morality: and labor United States Steel Corporation, 80 universal domain condition, 249, 257 utilitarianism, 14, 67, 75, 107, 109–​13, 115–​20, 123, 126, 165–​66, 184, 189, 196–​97, 220, 226, 228, 230, 235, 240, 251, 253, 255, 258, 265–​66, 274, 295, 342 versus common moral intuitions, 119–​20 “ex ante" versus “ex post" appraisals, 111 fundamental principle of, 109 rule, 113, 120 versus traditional morality, 113 “weighted”, 258 utility, 54–​55, 59, 62, 67, 71, 110, 112–​14, 117, 119, 123, 128, 134, 148, 162–​63, 198, 213–​ 15, 224, 228, 232, 253, 256, 258, 262, 264, 266, 271, 285–​87, 300, 308 cardinal, 62–​64, 256, 286 diminishing marginal, 114, 166, 198 expected, 63–​68, 152, 228, 236, 255, 259 function, 59, 63–​64, 119, 123, 285–​87 ordinal, 56, 58–​59, 63–​64 as pleasure, 110 theory, 56, 58, 60, 62, 64, 72, 74–​75, 77, 89, 122, 128, 143, 151 value, 39, 165, 173, 176, 182, 200–​01, 204, 208, 214, 218–​19, 223, 230, 235, 240, 252–​53,

Index 299, 303–​05, 309–​10, 312, 316, 329, 334, 337, 339–​43, 349 beyond preference satisfaction, 136 civic, 35, 101 economics as “value free”, 340–​42, 344, 348–​49 of human life, according to economists, 30 individual and social, 248 intrinsic, 109, 200 judgments of, 146, 149, 255 market versus moral, 30–​31, 34, 47, 97 of non-​human life, 112, 329, 333 ultimate, 135 van Heerde, J., xix van Hees, M., 266 van Parijs, P., xix, 220 Vanderschraaf, P., 289 Vane, H., 45, 296 Varian, H., 209–​11 veil of ignorance, 227–​29, 231, 233, 236, 241, 288 and utilitarianism, 255 See also original position; Rawls, J.: theory of justice “veil of uncertainty”, 237 Vickrey, W., 56, 67, 228 Vizard, T., 144 Von Mises, L., 191 von Neumann, J., 63 von Wright, G. H., 89 Voorhoeve, A., xix vote, sale of, 100 vouchers (for health care), 316–​18 vouchers (school), 31–​33, 35–​37, 51, 295, 302–​05, 307–​11 Waldron, J., 191 Walker, S., xix wallet, lost, 84–​85. See also normativity: influence on behavior Walsh, V., 68 Walters, J., xix Walzer, M., 104, 216–​17 Ward, H., 289 Weber, M., 340 Weibull, J., 83 Weitzman, M., 333 Welch, B., 191 welfare, 13–​15, 20, 22, 24–​26, 28, 33, 47, 49–​ 50, 75–​76, 107, 109–​12, 114, 116–​18, 121, 126–​27, 130, 134–​38, 141–​43, 146–​47, 149–​51, 156–​59, 161, 163–​69, 173, 175, 183–​85, 188, 196–​98, 203, 205, 208–​10,

397

212–​15, 218, 230, 235, 241, 247, 251–​58, 262, 264, 285, 288, 304–​10, 312, 316, 320, 328, 331–​35 constituted by preference satisfaction, 25, 28, 123, 127–​28, 130–​35, 139, 305 formal versus substantive theories, 127 of future people, 331 inequalities, 206, 208, 212–​14, 264–​65 of non-​human animals, 111–​12 social, 223, 247–​49, 251, 264, 328 state, 188 total versus average, 110, 112–​14 welfarism, 107, 263 See also preference satisfaction; social welfare function; welfare economics welfare economics, 20–​22, 36, 107–​09, 123, 139, 141–​44, 146–​49, 151, 154, 157–​60, 165, 167, 169, 249, 254, 295, 302–​03, 305, 314, 328 welfare theorem (first), 148 welfare theorem (second), 149–​50 welfarism, 247–​48 Wellman, C., 190 Wenar, L., 191 Westmoreland, R., 220 Weston, S., 342 Weymark, J., 255, 266 White, B., 4 Wight, J., 16 Wilk, L., xix Wilkinson, R., 231 Wilkinson, T., 325 Williams, A., 171, 220, 241 Williams, B., 124, 143, 220 willingness to pay, 25, 29–​30, 72, 160, 163–​64, 166–​70, 303–​05, 308, 328, 333–​34 Wilson, R., 289 Winch, P., 89 Wolff, J., xix, 241 Wolff, R., 241 World Bank, 19–​22, 24–​25, 27, 30, 33, 36–​37, 147, 153, 163, 295, 302–​05, 308–​10 Xu, Y., 266 Yaari, M., 285 Yellen, J., 51, 90 Young, H. P., 240–​41, 289 Zuckerberg, M., 7 Zucman, G., 7