Ageing without Ageism? Conceptual Puzzles and Policy Proposals 0192894099, 9780192894090

Citation preview

Ageing without Ageism?

Ageing without Ageism? Conceptual Puzzles and Policy Proposals Edited by

Greg Bognar & Axel Gosseries

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © the several contributors 2023 The moral rights of the authors have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2022951043 ISBN 978–0–19–289409–0 DOI: 10.1093/oso/9780192894090.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

Acknowledgements We are extremely grateful to colleagues who accepted to serve as referees for the chapters of this volume, including Paula Gobbi, Soren Flinch Mitgaard, Holly Lawford-Smith, Kasper Lippert-Rasmussen, Jeffrey Moriarty, Pedro Pita Barros, Daniel Sabbagh, Nenad Stojanovic, James Taylor, Pierre-Etienne Vandamme, and Fabio Waltenberg. Many thanks also go to Dominic Byatt at Oxford University Press for his support throughout the preparation of this book and to the OUP referees for their very useful insights on our initial proposal. The editors would like to acknowledge the financial support of the Grant Agency of the Czech Academy of Sciences through a project on ‘Taking Age Discrimination Seriously’ (grant ID: 17-26629S, PI: Axel Gosseries) awarded to the Institute of State and Law of the Academy of Sciences of the Czech Republic, Centre for Law and Public Affairs (CeLAPA), created under subsidies for a long-term conceptual development (RVO: 68378122).

Contents List of Figures List of Tables List of Contributors

viii ix x

Introduction

1

Greg Bognar and Axel Gosseries

PART I. CONCEPTUAL PUZZLES 1. Age Discrimination: Is It Special? Is It Wrong?

13

Katharina Berndt Rasmussen

2. Does the Badness of Disability Differ from that of Old Age?

28

Kasper Lippert-Rasmussen

3. In Defence of Age-Differentiated Paternalism

41

Viki Møller Lyngby Pedersen

4. Age and the Social Value of Risk Reduction: Three Perspectives

53

Matthew D. Adler

5. Can Egalitarians Justify Spending More on the Elderly?

71

Paul Bou-Habib

6. Age Limits and the Significance of Entire Lives Egalitarianism

82

Axel Gosseries

7. Age Universalism Will Benefit All (Ages)

94

Simon Birnbaum and Kenneth Nelson

PART II.

POLICY PROPOSALS

8. ‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

115

Anca Gheaus

9. Age and the Voting–Driving Analogy

128

Alexandru Volacu

10. Empowering Future People by Empowering the Young? Tyler M. John

143

Contents

11. COVID-19, Age, and Rationing

vii

159

Greg Bognar

12. Ageism in Assisted Reproduction

172

Francesca Minerva

13. An Education Resource Account for Early School Leavers

184

Andrée-Anne Cormier and Harry Brighouse

14. Differentiating Retirement Age to Compensate for Health and Longevity Inequality?

199

Vincent Vandenberghe

15. Ageing in Place and Autonomy: Is the ‘Age-Friendly’ City Initiative Too Elderly-Friendly?

214

Kim Angell

16. An Age-Based Delayed Housing Wealth Tax

229

Daniel Halliday

17. Two Types of Age-Sensitive Taxation

242

Manuel Sá Valente

18. An Age-Differentiated Tax on Bequests

254

Pierre Pestieau and Gregory Ponthiere

Index

267

List of Figures 4.1 A prioritarian transformation function

57

7.1 Age-related profiles of social insurance income replacement (country group averages 1990–2015)

101

7.2 Age-balance of income replacement in social insurance by type of generational welfare contract in 18 OECD countries, 1990–2015

102

7.3 Income replacement in social insurance for three age-related social risks in Japan, Norway, and Sweden, 1960–2015

103

14.1 Differentiated retirement ages equalizing (expected) ill health across and within countries

206

14.2 Difficulty of differentiating ex post (importance of type-E and type-F errors). The case of low-educated versus highly educated females aged 55–65 in Germany (DEU), France (FRA), Belgium (BEL), and Poland (POL)

209

17.1 Standard and proposed distribution of income (full line) and tax rates (dotted line)

247

List of Tables 1.1 Four forms of discrimination with age-related examples

20

4.1 Current-year incomes of the 35 groups

59

4.2 Utilitarian value of increase ∆p in current-year survival probability (relative to utilitarian value of that increase for 80-year-old, low-income group)

61

4.3 CBA value of increase ∆p in current-year survival probability (relative to CBA value of that increase for 80-year-old, low-income group)

63

4.4 Prioritarian value of increase ∆p in current-year survival probability (relative to prioritarian value of that increase for 80-year-old, low-income group)

65

4.5 COVID-19 infection fatality risk (IFR)

66

4.6 Utilitarian social value of vaccination

67

4.7 CBA social value of vaccination

67

4.8 Prioritarian social value of vaccination

68

7.1 Type of generational profile in age-related social insurance and selected outcomes in 18 OECD countries

104

7.A1 Variables included in the empirical analyses: Childhood risk category by country, averages for the period 1990–2015

107

7.A2 Variables included in the empirical analyses: Working age risk category by country, averages for the period 1990–2015

108

7.A3 Variables included in the empirical analyses: Old age risk category by country, averages for the period 1990–2015

109

11.1 Four patients

162

14.1 Health items: Subjective health

203

14.2 Health items: Objective conditions

204

14.3 Differentiated retirement ages equalizing ill health (international reference = 67): Between- and within-country differentiation

207

List of Contributors Matthew D. Adler, Richard A. Horvitz Professor of Law and Professor of Economics, Philosophy, and Public Policy, Duke University Kim Angell, Department of Philosophy, UiT The Arctic University of Norway Katharina Berndt Rasmussen, Department of Philosophy, Stockholm University, and Institute for Futures Studies Simon Birnbaum, Department of Political Science, Södertörn University Greg Bognar, Department of Philosophy, Stockholm University Paul Bou-Habib, Department of Government, University of Essex Harry Brighouse, Department of Philosophy, University of Wisconsin, Madison Andrée-Anne Cormier, Department of Philosophy, York University, Glendon College Anca Gheaus, Department of Political Science, Central European University Axel Gosseries, Fonds de la Recherche Scientifique (FNRS) and Hoover Chair in Economic and Social Ethics, University of Louvain (UCLouvain)

Kasper Lippert-Rasmussen, Centre for the Experimental-Philosophical Study of Discrimination (CEPDISC), Aarhus University, and Department of Philosophy, UiT The Arctic University of Norway Francesca Minerva, Department of Philosophy, University of Milan Viki Møller Lyngby Pedersen, Centre for the Experimental-Philosophical Study of Discrimination (CEPDISC), Department of Political Science, Aarhus University Kenneth Nelson, Swedish Institute for Social Research (SOFI), Stockholm University Pierre Pestieau, Université de Liège; Center for Operational Research and Econometrics (CORE, UCLouvain); Paris School of Economics Gregory Ponthiere, Hoover Chair in Economic and Social Ethics, University of Louvain (UCLouvain) Manuel Sá Valente, Hoover Chair in Economic and Social Ethics and Superior Institute of Philosophy, University of Louvain (UCLouvain)

Daniel Halliday, School of Historical and Philosophical Studies, University of Melbourne

Vincent Vandenberghe, Institute of Economic and Social Research (IRES), Louvain Institute of Data Analysis and Modeling in Economics and Statistics (LIDAM), University of Louvain (UCLouvain)

Tyler M. John, Department of Philosophy, Rutgers University, New Brunswick

Alexandru Volacu, Faculty of Business and Administration, University of Bucharest

Introduction Greg Bognar and Axel Gosseries

This book aims to contribute to the essential and timely discussion on age, ageism, population ageing, and public policy. It attempts to demonstrate the breadth of the challenges by covering a wide range of policy areas from health care to old-age support, from democratic participation to education, from family to fiscal policy. It bridges the distance between academia and public life by putting into dialogue fresh philosophical analyses and new specific policy proposals. It approaches familiar issues such as age discrimination, justice between age groups, and democratic participation across the ages from novel perspectives. Our societies continue to rely extensively on age criteria, despite the fact that concern for age discrimination is not new. The US Age Discrimination in Employment Act was adopted as far back as 1967. In Europe, age has been increasingly included in anti-discrimination legislation over the past several decades. Legal scholars, sociologists, anthropologists, and other social scientists have long studied how age structures our lives. Children studies and gerontology are vast and well-established fields of scholarship. Yet, with few exceptions, practical philosophers have been less active than researchers from other disciplines on the age front.1 This book aims especially to contribute to filling this gap, in dialogue with other disciplines. Two trends in particular render this a timely exercise. One is the ongoing process of critically scrutinizing our societies through the prism of race, gender, disability, and other categories. This calls for looking at whether age is different—whether it is unique or, as it is sometimes put, ‘special’ from a normative perspective. Can this explain that it tends to get less attention than other social categories? Should we worry less about differential treatment on grounds of age than about differential treatment based on race or gender? And if so, what are the difference-makers that render age special from a normative perspective? The other trend that warrants a closer look at age is the ageing of our societies. Fifteen years ago, fewer than 500 million people were 65 or older. In 2030, there will be more than one billion people over 65, and by 2050, there will be around 1.5 billion.2 During the past decade, the number of older people surpassed the number of children under five for the first time in human history.3 In 2015, Japan was the only country that had more than 30 per cent of its population made up by those over 60; by 2050, this will become the case in all developed countries, including China.⁴ 1 2 3 ⁴

See, e.g. Daniels (1998), McKerlie (2013) or Bidadanure (2021). See National Institute on Aging, National Institutes of Health (US) (2007). See National Institute on Aging, National Institutes of Health (US) and the WHO (2011). See WHO (2015).

Greg Bognar and Axel Gosseries, Introduction. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Greg Bognar and Axel Gosseries (2023). DOI: 10.1093/oso/9780192894090.003.0001

2

Introduction

Population ageing presents enormous challenges. Ageing societies will have to make massive adjustments to their old-age support and healthcare systems, their labour markets, and their social and political institutions. Population ageing will have profound effects on family life, the nature of work, politics, and people’s life plans. With no historical experience to rely on, societies will have to try untested, novel, and creative ways for coping with the challenges of ageing. And they must be able to provide ethical justifications for their choices. Our aim is to provide a multidisciplinary discussion, with contributions especially from philosophy but also inputs from political science, economics, sociology, and other areas. In order to impose some order on a wide-ranging collection of topics, we divided the book into two parts. The chapters in Part I present in-depth discussions of conceptual and normative issues. The chapters in Part II defend specific policy proposals, grounded in explicit normative arguments. Readers interested in conceptual issues can begin at the beginning; readers more interested in policy alternatives can pick one of the chapters from Part II as their entry point. To help orientation, we provide separate overviews of the two parts below.

1. Overview of Part I Part I departs from the fundamental normative question about age. Is unequal treatment on the basis of age permissible? How does it differ, from an ethical point of view, from other forms of differential treatment? Age discrimination has been a neglected area in the literature on wrongful discrimination in philosophy and legal theory. The first three chapters aim to fill some of the gaps by approaching the fundamental normative question from different directions. What do different theories of wrongful discrimination have to say about the wrongness of age discrimination specifically? How is age discrimination connected to disability discrimination? How should we think about the link between paternalism and age? In the opening chapter of Part I (‘Age Discrimination: Is It Special? Is It Wrong?’), Katharina Berndt Rasmussen examines the morality of age discrimination by bringing together philosophical theories of wrongful discrimination and accounts of the ‘specialness’ of age—that is, defences of the claim that there is a moral difference between discrimination on the basis of age and discrimination on other grounds such as gender or race. After providing an overview of considerations that might make age special, Berndt Rasmussen offers a taxonomy of different forms of age discrimination and relates them to three theories of wrongful discrimination. She finds that these three theories differ with respect to their moral assessment of various forms of age discrimination due to the different roles that ‘specialness’ considerations play in each. Rather than arguing for any particular theory, however, Berndt Rasmussen concludes by offering a template for identifying, analysing, and morally evaluating different forms of age discrimination.

Ageing without Ageism?

3

In the second chapter, Kasper Lippert-Rasmussen compares age discrimination and disability discrimination (‘Does the Badness of Disability Differ from that of Old Age?’). He begins with a familiar question from the philosophy of disability: to the extent that being disabled is worse than being non-disabled, is this largely because of factors that are independent of the social environment (the ‘disability as bad difference’ view) or largely because of the (ableist) nature of the social environment (the ‘disability as mere difference’ view)? Correspondingly, we might ask, to the extent that being old is worse than being young, is this largely because of factors that are independent of the social environment or largely because of the (ageist) nature of the social environment? Are the answers to these questions related? If they are, how? Lippert-Rasmussen considers whether, if we are inclined to accept the view that the disadvantages of disability are largely caused by an ableist social environment, we should also accept the view that the disadvantages of old age are largely caused by an ageist social environment. He thinks we should. But Lippert-Rasmussen argues that the view that ageist social environments are primarily responsible for the disadvantages of old age should be rejected. Therefore, we should also reject the view that ableist social environments are primarily responsible for the disadvantages of disability. Yet, while insisting that our views on disability constrain those we adopt on age, Lippert-Rasmussen also stresses that we should not overstate the importance of taking a stance on the mere-difference and bad-difference divide. Rather, we should pay more attention to the specific ways old age and disability cause disadvantage instead of trying to defend broad generalizations about what family of causes predominates. The third chapter examines age discrimination from the perspective of agedifferentiated paternalism. Many of us share the intuition that paternalism is less problematic when applied to children rather than to the elderly. One interest of Viki Møller Lyngby Pedersen’s chapter, ‘In Defence of Age-Differentiated Paternalism’, is that she stresses that we need to consider not only the dimensions of competence and voluntariness but also the magnitude of the good promoted by paternalistic interventions. While this dual account does not generally challenge common-sense intuitions about paternalism, it introduces additional complexity when evaluating paternalistic interventions in a wide range of cases. The chapter considers many examples, from age-differentiated rules for access to sterilization, through age-differentiated fines for not using a helmet, to age-differentiated prices for cigarettes. Readers especially interested in these issues may also want to take a look at Chapter 8 (on paternalism and conceptions of childhood), Chapter 12 (on paternalism in assisted reproduction and social freezing), or Chapter 13 (on anti-paternalism in compulsory education). The remaining chapters in Part I take up issues of distributive justice. They consider the connection between age and different methods of policy evaluation. They examine the role of age in egalitarian theories of distributive justice. They ask whether it is justified that modern welfare states spend more on the elderly than on other age groups, explore the implications of the idea that principles of distributive equality

4

Introduction

should apply to whole lives rather than particular segments (or stages) in life, and defend the view that social insurance systems should be age-balanced, offering similar levels of income replacement across age-related social risks. In ‘Age and the Social Value of Risk Reduction: Three Perspectives’, Matthew D. Adler compares three frameworks of policy analysis from the perspective of fatality risk reduction for different age groups. Do they imply that the value of risk reduction depends on age—and how do they relate age to other factors? The three frameworks are utilitarianism, prioritarianism, and cost–benefit analysis. Adler finds that the value of risk reduction decreases with age but increases with income according to utilitarianism. It decreases even more sharply with age according to prioritarianism, but prioritarianism can also neutralize the effect of income. And for cost–benefit analysis, the value of risk reduction increases with income even more sharply than for utilitarianism, while it first increases and than decreases with increasing age. None of the frameworks, therefore, is neutral with respect to age. They value fatality risk reduction differently depending on a person’s age (and income). Prioritarianism is the only approach that can neutralize the effect of income and put higher value of reducing risks to the young. That may be an attractive feature. Paul Bou-Habib, in his chapter ‘Can Egalitarians Justify Spending More on the Elderly?’, takes an egalitarian approach to age and fair distribution. He argues that the fact that modern welfare states devote a disproportionate amount of their budget to the needs of the elderly raises a puzzle. People who reach old age are often, on the whole, more fortunate than those who don’t because they have enjoyed a longer life. In devoting disproportionate expenditure towards their needs, the welfare state thus appears to be privileging the needs of those who are more fortunate than others. Bou-Habib examines the response to this puzzle provided by relational egalitarians (who hold that we should care not only about how the welfare state distributes resources between persons but also about whether it protects people from mistreatment by others). Relational egalitarians justify disproportionate expenditure on the elderly on the grounds that it is necessary to protect them against domination and marginalization, among other forms of mistreatment. But Bou-Habib finds that the relational egalitarian response does not solve the puzzle. He proposes a different solution, based on two claims. First, suffering is intrinsically bad and should be prevented, even when it is experienced by persons who are more fortunate than others. Second, disproportionate expenditure on the needs of the elderly is a form of insurance that all persons would have purchased in fair circumstances. In his chapter, ‘Age Limits and the Significance of Entire-Lives Egalitarianism’, Axel Gosseries focuses on the claim that principles of distributive justice should be applied to whole lives—that is, to determine what we owe to people as a matter of fair distribution, we need to consider how they fare during their entire lives. Gosseries provides an overview of the entire life view and explores its possible underlying intuitions. He separates a defensive version of the view (which argues that some age limits are not objectionable) from an affirmative one (which argues that some age limits are actually desirable). He concludes that while age limits tend to provide one of the best illustrations of the practical relevance of the ‘entire life’ debate, the latter does not

Ageing without Ageism?

5

necessarily offer us insights that are as significant as expected to defend age limits over their whole range. Part I closes with a more empirically orientated chapter. It starts with the observation that welfare states differ greatly in the extent to which they provide social protection for various age-related social risks. They set different priorities between needs associated with childhood, maturity, and old age. In ‘Age Universalism will Benefit All (Ages)’, Simon Birnbaum and Kenneth Nelson explore and defend the ideal of age universalism in social insurance, according to which the degree of income replacement should be similar across age-related social risks. The argument suggests pragmatic advantages of age-balanced social insurance, showing that it tends to provide higher levels of income replacement for age-related risks throughout the life cycle and achieve more favourable social outcomes in all age groups with respect to poverty rates, trust, and subjective well-being.

2. Overview of Part II The chapters in Part II have a more policy-oriented focus. They cover a range of topics from different perspectives. The topics include political participation, education, health care, retirement, and old-age social services as well as taxation and inheritance. The first cluster of chapters is on political participation and voting rights. The chapters address whether and how disenfranchising the young can be justified on the basis of different conceptions of childhood, whether the voting–driving analogy can justify disenfranchising the old, and whether giving extra weight to the young in political decision-making can be a plausible avenue to addressing concerns about political short-termism. In the first chapter, ‘“Let Them Be Children”? Age Limits in Voting and Conceptions of Childhood’, Anca Gheaus explores alternative views about the nature and value of childhood and their relevance to the issue of children’s voting rights. In particular, she contrasts one view that regards childhood as a mere deficiency and as preparation time for adulthood with a family of views that emphasizes the value of goods unique to childhood, such as playfulness and carefreeness. Defenders of deficiency views tend to assume that the lack of agency is an unqualified bad for children and neglect ways in which childhood allows access to other sources of value. Gheaus maps out how the different accounts bear on arguments for and against enfranchising children. She also explains why children who live in a society in which many adults fail to comply with their duties of intergenerational justice have a weightier interest in voting and hence why the case for children’s enfranchisement is stronger in such circumstances. The next chapter continues to explore political participation by looking at the other end of life. Should there be an age limit such that people over it lose their eligibility to vote? After all, loss of ability is often used to justify restricting people’s freedom. For instance, age-related loss of ability is used to justify the requirement of periodic renewal of driving licences and could result in the loss of driving permit

6

Introduction

for the elderly, limiting their freedom of movement. Can there be an analogous case for voting? In ‘Age and the Voting–Driving Analogy’, Alexandru Volacu asks this question. He examines arguments by analogy in general and formulates such an argument linking driving and voting. He considers different ways the argument could be applied to age-adjusted voting rights. However, in the end, he finds that there are significant dissimilarities between driving and voting. Thus, Volacu concludes that the argument is unsuccessful. In ‘Empowering Future People by Empowering the Young?’, Tyler M. John argues that the state is plagued with problems of political short-termism: excessive priority given to near-term benefits at the expense of benefits further in the future. Political scientists and economists reckon that political leaders rarely look beyond the next 2–5 years, exacerbating problems such as climate change and pandemics. What can be done to counter this? One possible mechanism involves apportioning greater relative political influence to the young. The idea is that younger citizens generally have greater additional life expectancy than older citizens, and thus it looks reasonable to expect that they have preferences that are extended further into the future. If we give greater relative political influence to the young, our political system might exhibit greater concern for the future. But John shows that giving greater political power to the young is unlikely in itself to make states significantly less short-termist: no empirical relationship has been found between age and willingness to support long-termist policies. Instead, he proposes a more promising age-based mechanism. States should develop youth citizens’ assemblies that ensure accountability to future generations through a scheme of retrospective accountability. Policymakers would be rewarded in the future in proportion to the effects of their policies on the long run. This would incentivize them now to choose policies that have the best long-term consequences. The second couple of chapters in Part II are on health care. The first is Greg Bognar’s chapter on ‘COVID-19, Age, and Rationing’. During the COVID-19 pandemic, some hospitals found themselves short of ventilators, intensive care unit (ICU) beds, and qualified medical personnel to take care of patients. Physicians had to make difficult, life-and-death choices. They were aided by various guidelines and recommendations issued by governments and medical associations. Bognar reviews some of these guidelines, looking in particular at the role of age and life expectancy as criteria for the rationing of healthcare resources. He defends the view that the ethical aim of triage should be to maximize benefits and concludes that while neither age nor life expectancy should be used to categorically exclude patients, both may have a role in triage by virtue of their connection to capacity to benefit. Francesca Minerva’s chapter, ‘Ageism in Assisted Reproduction’, begins from the fact that female fertility declines at a much faster rate than male fertility. While, in the past decades, assisted reproduction treatments (ARTs) have dramatically increased women’s chances of getting pregnant over the age of 35, it remains very difficult for women above their mid-40s to get pregnant, given that ART success rates decrease with increasing age. Moreover, many European countries legally prevent women over the age of 45 from accessing ARTs.

Ageing without Ageism?

7

Minerva opposes upper age limits on ARTs for women, rejecting three arguments that are based, respectively, on paternalism (to protect the mother’s health), the link between age and increased risk of abnormalities for the child, and the diminished ability of older parents to take care of children. She also calls for a more pro-active policy, including free access to social freezing and investment in research into ways of delaying menopause. In the following chapter, Andrée-Anne Cormier and Harry Brighouse propose ‘An Education Resource Account for Early School Leavers’. They argue that school should cease to be compulsory at age 16 and that an education resource account (ERA) should be established for students who leave school at that age. The ERA would be sufficient to cover three years of full-time education. It could be linked to inflation and early school leavers could use it in accredited non-profit educational institutions at any later point in their lives. Two sets of arguments support their proposal. The first, building on the empirical literature, focuses on efficiency and highlights the advantages of an ERA with respect to the ‘disruptive’ students issue in particular. The second set of arguments is anti-paternalistic. Cormier and Brighouse distinguish three anti-paternalistic arguments: the view that individuals are the best judges of their own welfare (also discussed in Chapter 3), the idea that autonomous decision-making is a component of well-being, and a respect-based view of what renders anti-paternalism wrong. While they endorse the latter two arguments, their ERA proposal still has a mildly paternalistic dimension since its funds can only be used for education purposes. Vincent Vandenberghe takes up the issue of retirement in ‘Differentiating Retirement Age to Compensate for Health and Longevity Inequality’. As he points out, usually a uniform age is used to proxy work capacity loss and trigger the payment of pensions. Recently, however, some have argued that we need several retirement ages to better match the distribution of work (in)capacity across socio-demographic groups. At first sight, this proposal makes perfect sense. Work capacity declines faster among low-income and low-educated individuals. But there is also a lot of unaccounted heterogeneity even inside narrowly defined socio-economic groups. And this compromises the feasibility and desirability of retirement-age differentiation. Under a regime of systematic retirement-age differentiation, there would be many situations with no retirement for people with serious work restrictions and, simultaneously, numerous cases where entirely healthy people enjoy retirement. An alternative approach would be to stick to a uniform retirement age, backed up by a reinforced disability scheme. Old age also takes centre stage in Kim Angell’s ‘Ageing in Place and Autonomy: Is the “Age-Friendly” City Initiative Too Elderly-Friendly?’ Angell is concerned with the ‘age-friendly cities’ initiative aimed at enhancing people’s opportunity to age in place. He presents an autonomy-based defence of the idea and examines the moral claim that the elderly can make in support of their ability to age in place. He emphasizes, among other considerations, that ageing in place can have cognitive benefits through the routines and habits made possible by familiar environments.

8

Introduction

He argues, however, that the claims of the elderly can come into conflict with the claims of the young. We should not only look at today’s elderly but also anticipate how today’s young will fare when they get old. Angell appeals to the cohort-specific predictions by the Organisation for Economic Co-operation and Development (OECD)— such that, for example, today’s young are expected to be worse off when old than the currently old—to make the case for an ‘all-age-friendly’ (or even ‘young-friendly’) interpretation of the age-friendly cities initiative, while also insisting on the importance of policies benefiting low-income families (regardless of age) and promoting intergenerational housing initiatives. The last cluster of chapters in the book focuses on age and taxation, looking respectively at housing, income, and bequests. In his chapter, ‘An Age-Based Delayed Housing Wealth Tax’, Daniel Halliday considers taxation and housing wealth. Popular narratives around ageing and intergenerational inequality suggest that young people increasingly tend to subsidize older people in spite of enjoying poorer economic prospects. One specific concern is that older and younger birth cohorts are unequally situated with respect to the distribution of housing wealth as well as the distribution of the tax burden. Halliday addresses this concern by proposing an age-based delayed housing wealth tax. The idea is that once homeowners reach a certain age, they are charged some portion of their home’s value on an annual basis, which would eventually be paid to the tax office upon the death of the surviving spouse. This tax can be avoided by downsizing to a home of lesser value and thereby freeing up housing to be purchased by younger people. Retaining a valuable home means, instead, incurring a tax liability that can be used to fund the benefits consumed by retirees. A delayed housing wealth tax can be designed to accommodate variables such as couples who differ in age or single retired homeowners. Halliday argues that his proposal compares favourably with alternatives, such as inheritance taxation, for getting older people to absorb the costs of their care. Just as in the previous chapter, access to housing for the young is a central concern. In his chapter, Manuel Sá Valente distinguishes between ‘Two Types of AgeSensitive Taxation’. One is a form of cumulative income taxation which taxes annual income, taking into account all earlier income years instead of just the last one. The other is an explicitly age-differentiated scheme that taxes annual income adjusted by a rate that depends on the taxpayer’s age. The chapter first presents reasons to support cumulative income taxation and examines how it would affect fiscal obligations across life. Then, it argues that maximin egalitarians—that is, egalitarians who give absolute priority to improving the situation of the least well off—should aim at a hump-shaped tax rate across people’s lives. Such a rate reflects a concern about both early death and poverty in old age, hence focusing on the young and the elderly, not the middle-aged. The chapter questions whether cumulative income taxes can deliver this result without resorting to explicitly age-differentiated taxes. It reaches the conclusion that while cumulative income taxation can benefit the young (including the short-lived among them), age-differentiated taxes are necessary to protect the elderly poor. In the final chapter, Pierre Pestieau and Gregory Ponthiere present four arguments supporting ‘An Age-Differentiated Tax on Bequests’—that is, a tax rate on

Ageing without Ageism?

9

inheritance that varies with the age of the deceased. The arguments are based on different ethical foundations and lead to an inheritance tax that can either increase or decrease with the age of the deceased. Pestieau and Ponthiere make the case for an age-differentiated tax based on the idea of compensating unlucky prematurely dead persons. Their view supports a bequest tax that increases with the age of the deceased. Along with Chapters 5 and 17, their chapter illustrates the many normative problems that differential longevity raises. ∗∗∗ Together, these chapters provide us with a sense of the complexity of the issues at stake. The account we accept about the wrongness of discrimination makes a difference to which age-based policies can be defended. So does the view of paternalism we take and, more generally, the background theory of justice we endorse. It matters whether we consider differential longevity unfair, whether we are concerned with equality between entire lives or parts of lives, whether we hold that differential treatment by age is relevantly similar to unequal treatment by race, gender, or disability. The justification of age-based policies can be affected by a multitude of seemingly remote normative commitments and ideas. In addition, it is influenced by empirical assumptions about age and ageing. Several chapters in this book have critically discussed such assumptions, including the connection of age to specific abilities (for instance, working capacity or political competence), characteristics (for instance, fertility) or dispositions (for instance, long-termist preferences). All this suggests that it is probably wise to renounce the quest for a unified view of the normative relevance of age. The role of age is likely to remain different in different policy areas and in different policies within those areas. This means that we are left with several tasks. We should keep critically investigating the degree to which specific age-based policies can be justified. When we develop and debate new policies, we should continuously keep in mind how they are affected by age and how they would affect different age groups. We should keep unveiling common patterns of justification that operate across several domains and that oppose or support age-based practices. Behind all that, there is the question whether there should be a society on the horizon in which age loses its structuring function: where people can work before studying, retire before starting to work, or have children well after having started their professional career. We hope that this book will contribute to helping the reader decide whether such a society would not only be feasible but also, and more importantly, desirable.

References Bidadanure, Juliana Uhuru. 2021. Justice Across Ages: Treating Young and Old as Equals. Oxford: Oxford University Press. Daniels, Norman. 1998. Am I My Parents’ Keeper? An Essay on Justice between the Young and the Old. Oxford: Oxford University Press.

10

Introduction

McKerlie, Dennis. 2013. Justice between the Young and the Old. Oxford: Oxford University Press. National Institute on Aging, National Institutes of Health (US). 2007. Why Population Aging Matters: A Global Perspective. Baltimore, MA: National Institute on Aging, National Institutes of Health (US). National Institute on Aging, National Institutes of Health (US) and the WHO (World Health Organization). 2011. Global Health and Ageing. Baltimore, MA: National Institute on Aging, National Institutes of Health (US) and WHO. World Health Organization (WHO). 2015. World Report on Ageing and Health. Geneva: World Health Organization.

PART I

CONCEPTUAL PUZZLES

1 Age Discrimination Is It Special? Is It Wrong? Katharina Berndt Rasmussen

1. Introduction Imagine Dana and Eli, two applicants for a vacant position. While both are qualified in all relevant respects, neither is called to be interviewed for the job simply because of their gender (Dana is female) and race (Eli is black), respectively. These cases are intuitively clear instances of discrimination and intuitively morally wrong. They are classified as group discrimination also by the following definition:1 An agent (say, an employer) group discriminates against someone (say, an applicant) on grounds of property P, by doing something (e.g., disregarding their application) if and only if: (i) in disregarding the application, the employer treats the applicant worse than she would have treated him, had he not had P, or had she not believed him to have P, (ii) it is because the applicant has P or because the employer believes that he has P, that she treats him worse, and (iii) P is the property of being a member of a socially salient group, i.e., a group perceived membership of which is important to the structure of social interactions across a wide range of social contexts.

Being female and being black, respectively, are socially salient properties in many societies. Clearly, not being called to be interviewed for a job that one applied for just because of such a property—when one is qualified and would have been called had one lacked the property—amounts to being treated worse in at least one of the following senses: (a) being made worse off (e.g. deprived of an opportunity), or 1 Cf. Lippert-Rasmussen (2014); Berndt Rasmussen (2019); and many of the entries in LippertRasmussen (2017). In legal terms, this definition captures (possibly legally acceptable) differential treatment as well as (unlawful) discrimination. On the moral status of these phenomena, see section 4.

Katharina Berndt Rasmussen, Age Discrimination. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Katharina Berndt Rasmussen (2023). DOI: 10.1093/oso/9780192894090.003.0002

14

Age Discrimination: Is It Special? Is It Wrong?

(b) being treated as inferior (e.g. considered not worthy of equal consideration with other—male or white—applicants).2 And arguably, treating someone worse in at least one of these senses just because of their socially salient property is prima facie morally wrong.3 Hence, ceteris paribus, Dana and Eli are wrongfully discriminated against. Now, imagine Alex, Billie, and Charlie, three equally qualified applicants for a vacant position. They belong to three different birth cohorts (say, three consecutive generations) and thus to different age groups: Alex is 60+, Billie is 40, and Charlie is under 18. While Billie is called back to be interviewed, the other two are not—simply because of their old and young age, respectively. Clearly, being old (60+) and being young (under 18) are socially salient properties in many societies and, in this sense, comparable to being female and being black. Moreover, in not being called back, Alex and Charlie suffer the same form of worse treatment as Dana and Eli. Hence, even their cases should be classified as discrimination by the above definition and count as morally wrong for the same reasons. Writers on age discrimination, however, suggest that age is ‘special’, that is, relevantly different from other grounds of discrimination such as gender or race. Such specialness, in turn, might have moral ramifications, possibly making age discrimination less severe, or more justifiable, than these other forms. The specialness of age may thus translate into lesser moral seriousness of age discrimination. Our practices and intuitions seem to support this idea: in many societies, it is common and commonly accepted that, for example, the right to vote and to run for office is not granted to minors, that car or life insurance premiums are age-adjusted, or that retirement at a certain age becomes mandatory. On the other hand, there might be good reasons to change these practices and discount these intuitions. This chapter examines the moral status of age discrimination by bringing together accounts of the wrongness of discrimination with accounts of the specialness of age. Section 2 summarizes the special features of age and their role in different proposals to justify the use of age criteria and suggests a template within which these specialness considerations become relevant: the argument from specialness. Section 3 explores different forms of age discrimination. Section 4 presents three influential accounts of the wrongness of discrimination and shows that different forms of age discrimination are covered by different wrongness accounts. Moreover, I return to the proposals justifying the use of age criteria, based on the specialness of age, and explore the roles they can play under these different accounts. Section 5 concludes.

2. Specialness of age and age-based treatment The debate on age discrimination largely focuses on chronological age, defined as the number of years from a person’s birth to the given date. Chronological age 2 See Berndt Rasmussen (2019). 3 For ease of exposition, I switch between talking about an action’s being prima facie (morally) wrong and there being a prima facie (moral) reason against the action.

Ageing without Ageism?

15

(henceforth without the qualifier) has three special features, as compared to, for example, gender or race: (a) perfect passage-of-time correlation (by definition); (b) high divisibility: a normal life span can be partitioned into many alternative age segments of different sizes; and (c) multiple belonging: a person with a normal life span will belong to a number of such age segments as time passes.⁴ A key idea is that, because of these special features, relying on age criteria (e.g. age limits) may sometimes be justified in two different ways. First, means–end efficiency: age is a good proxy (statistical indicator) for certain given target variables and, as such, a tool for more efficient decision-making. Second, overall betterness: using age criteria leads to better overall outcomes, in some sense. I will discuss these two proposals in sections 2.1 and 2.2, indicating which specialness features (a–c) become relevant at different points in the arguments.

2.1 Means–end efficiency: Age as reliable and precise proxy In typical employment cases, applicant age is ‘significantly correlated’ (Gosseries 2014: 63f.) to, for example, job qualifications and expected productivity. This is mainly due to specialness feature (a), passage-of-time correlation. To illustrate: Charlie, who is under 18, most likely does not have a college or university degree. This has to do with the way our societies organize their education systems: age of entry, sequence of educational levels, number of years required for a degree, etc. Also, under-18-year-olds are likely not to have the cognitive capacities required for jobs where risk management is central. This has to do with typical neurological/brain development. Alex, who is 60+, is likely to lack some of the physical and cognitive abilities of an applicant in their forties, like Billie, but also likely to have more experience-based competences. This has to do with processes of physical and mental decline and with the time dimension of learning curves, respectively. In a medical context, given statistical facts about life expectancy, age can work as a proxy for remaining life years (Cupit 2013; Bognar 2015). Thus, age can be one relevant factor for allocating scarce health resources, as exemplified by the concepts of quality-adjusted life year (QALY), and disability-adjusted life year (DALY).⁵ Thus, for a variety of reasons, age can be a reliable proxy for relevant individual target variables. Of course, for any given applicant or patient, an employer or medical provider might make a more accurate assessment of such target variables by conducting an individual assessment. Yet, since the latter tend to be more time-consuming, ⁴ For discussions of these features, see Macnicol (2006); Cupit (2013); Gosseries (2014); Bidadanure (2016). ⁵ See Bickenbach (2016).

16

Age Discrimination: Is It Special? Is It Wrong?

complicated, or morally problematic, relying on age criteria can facilitate decisionmaking. Thus, to promote means–end efficiency, there may be reasons to use age criteria. Justifying age criteria by reference to means–end efficiency also relies on specialness feature (b), high divisibility. Since a normal lifespan is divisible into any set of age segments that are relevant in a given context, using age criteria potentially allows for higher levels of ‘precision’ (Gosseries 2014: 63f.; cf. Bidadanure 2016) in approximating the relevant target variables, compared to, for example, gender and race. Using more precise age criteria allows for increased means–end efficiency (at least, if information is not too costly). However, there is a general objection against this efficiency justification of age criteria. In openly sexist or racist societies, where women or people of colour are denied higher education, gender and race can also be reliable—and sufficiently precise— proxies. Still, we would not consider the resulting discriminatory hiring decisions as morally less severe or more justifiable than their counterparts in our (less openly sexist or racist) societies due to these different social facts. Whether something is a good proxy is contingent on, for example, physical, biological, and psychological facts but also social facts that depend on (the aggregate of) our choices, such as how we choose to set up our education system. When social facts are invoked to justify the use of a proxy, the justificatory burden shifts to these social facts and underlying choices. This means that if we seek to justify the use of age criteria by appealing to the fact that age is a good proxy, yet this latter fact is explained by the fact that our society is organized around age criteria, our argument becomes circular unless we can independently justify society’s organization around age criteria (Gosseries 2014: 65f.; Lippert-Rasmussen 2014: 283–299). The analogy with gender and race thus serves to illuminate the point that establishing age as a reliable and precise proxy for certain target variables is not sufficient for an overall justification of age criteria. Moreover, we should pause to note that this, indeed, is not necessary either. One could reject the idea that age is a reliable, sufficiently precise proxy for specific individual target variables and still concede that relying on age criteria allows for better overall outcomes on a collective level. For example, if the justification of certain age limits on alcohol consumption appeals to epidemiological evidence of their correlation to reduced youth criminality or disease rates, such criteria may be used primarily as a means for achieving a better collective outcome, even though correlation with specific individual target properties may be weak. The concern for better overall outcomes is thus distinct from the concern for means–end efficiency. And overall outcomes need to be considered even in the final analysis of the justificatory force of means–end efficiency considerations.

2.2 Better overall outcomes: Utility or fairness/equality Attempts to justify age criteria from better overall outcomes are mainly concerned either with utility or with fairness/equality. One key consideration under the headline

Ageing without Ageism?

17

of utility is collective ‘sequence efficiency’ (Gosseries 2014: 70f.): organizing certain aspects of society around age criteria promotes total utility as it makes society overall more functional. The education system is an obvious example. There are efficiency reasons for making people go through stages of education, starting at a young, learning-conducive age and consecutively taking higher levels, prior to entering the labour market. Again, however, one could object that there might be overall efficiency gains from race or gender discrimination—say, in a caste-like hierarchy—which we would not grant any justificatory force. With age, though, there is a crucial difference. In the education system example, collective sequence efficiency originates from individual sequence efficiency.⁶ In certain contexts, ordering someone’s activities in chronological sequences promotes their achievement of the context-given objectives due to the specialness feature (a), perfect correlation with the passage of time, which, in turn, correlates with other factors. For example, learning how to spell prior to signing up for creative writing classes yields a better outcome for the individual than the reverse. Individual sequence efficiency is thus a further specialness consideration—and one which can appeal even to non-utilitarians. The trade-offs it is based on are not interpersonal, as in the sexist or racist caste system, where some people’s extra burdens generate benefits for others. Rather, the trade-offs are intrapersonal: collective gains result from a more efficient distribution of burdens and benefits over individuals’ lifetimes, as allowed by the specialness feature (c), multiple belonging. Moreover, specialness feature (b), high divisibility, facilitates the tailoring of age segments to improve intrapersonal distributions—to the overall benefit of both the individuals and the collective whole. From a perspective of fairness or equality, age criteria seem problematic as they imply worse treatment (in at least one of the above senses) of some—and, hence, inequality. Old Alex and young Charlie are denied an employment opportunity, which is granted to middle-aged Billie. The key idea with justifying such treatment, nevertheless, from a fairness or equality perspective is that, in the case of age, the inequality is mitigated once we consider whole lives. According to ‘complete life neutrality’ considerations (Gosseries 2014: 66ff), for a fixed set of age criteria, ceteris paribus, there will be no inequality in treatment over complete lives. The ceteris paribus clause ensures that—due to specialness feature (c) (i.e. multiple belonging)—each age criterion for resource allocation impacts equally on everyone, if only at different points in time, thus treating each equally over their lifetime.⁷ However, this idea presupposes a very demanding ceteris paribus condition, where everyone has the same life length as well as the same lifetime profile of needs, goals, and desires. Yet, realistically, a medical treatment offered to all those under 70 affects those who happen to need it at 65 or 75, respectively, very differently; a pensions ⁶ Bidadanure (2016: 247) calls this ‘lifespan efficiency’. Alternatively, one could ask which resource allocation pattern over their lifetime a rational chooser would prefer from behind a Rawlsian veil of ignorance (Cupit 2013; cf. Daniels 1988; Bognar 2008, 2015). ⁷ Some egalitarians may object that relational inequalities at any specific time matter and do not disappear because of reversed inequalities at some other time (Bidadanure 2016).

18

Age Discrimination: Is It Special? Is It Wrong?

plan for all and only those above 65 benefits only whoever lives that long (Cupit 2013). Moreover, changes in these age criteria—or in the environment in which these criteria receive significance—may lead to complete life inequalities between birth cohorts. For example, if the required age for a benefit is raised at some point in time, all birth cohorts reaching the threshold age after that point will be worse off, over their lifetime, than previous ones.⁸ According to ‘affirmative egalitarian’ considerations (Gosseries 2014: 70ff), age criteria may effectively reduce existing social inequalities over complete lives. Consider, for example, mandatory retirement at age 67. Assume that there is a number of consecutive birth cohorts, each of which has a certain proportion of unemployed individuals. As one cohort reaches the threshold, their jobs are made vacant and filled with members of the subsequent cohorts, shifting employment benefits to some previously unemployed. Later on, these cohorts, of course, have to do the same for their successors. Then, people’s life trajectories are equalized in the sense that fewer will go through stretches of unemployment prior to, while all are retired after, this age threshold.⁹ Another example is the allocation of life-saving treatment: fairness may require that we give it to a young person, who has not yet reached the ‘fair innings’ threshold of a ‘complete or full life’, rather than to someone beyond this threshold (Bognar 2008). Similarly, McKerlie (1992) suggests that, due to differences in life span, we should discriminate in favour of the young to concentrate resources to life stages through which more will live. Acknowledging inequalities over whole lives (e.g. in employment opportunities or life years) combined with special feature (c), multiple belonging, may thus provide affirmative egalitarian support for age-based treatment.

2.3 Argument from specialness To pinpoint the role of specialness considerations for the overall moral assessment of age discrimination, I propose the following argument from specialness: (1) There is a prima facie reason against group discrimination. (2) Age-based treatment is a form of group discrimination (like gender- or racebased treatment). (3) In some contexts, specialness considerations concerning means–end efficiency or overall better outcomes provide reasons for age-based treatment that outweigh the prima facie reason against it. (4) Hence, in such contexts, we have overall reason for age-based treatment. ⁸ The ‘diversification’ approach (Gosseries 2014: 63) highlights that complete life inequalities, resulting from age criteria, are moderated due to age’s specialness: criteria tend to vary across contexts (due to (b)), and resulting disadvantages tend to spread out over an individual’s lifetime (due to (c)). ⁹ See Wedeking (1990); Arneson (2006: 797f.). For dissenting views, see Overall (2006); Nussbaum and Levmore (2017).

Ageing without Ageism?

19

The specialness considerations of premise (3) have just been explored. In section 3, I briefly address premise (2) and consider four ways in which age-based treatment constitutes group discrimination. In section 4, I consider three influential accounts of the prima facie wrongness of group discrimination concerning premise (1).

3. Forms of age discrimination Whenever an agent’s age-based treatment of another individual relies on a statistical correlation between age and some agent-relevant end, aiming at means–end efficiency, we are dealing with statistical discrimination. We can capture the statistical version of age discrimination by modifying condition (ii) of our definition of group discrimination: (ii∗ ) it is because the employer believes that the applicant has property P, and that P is statistically correlated to a target variable that is relevant for the employer, that she treats him worse. (See Lippert-Rasmussen 2014: 81; Schauer 2017: 42f.)

The general idea is that an agent believes that some socially salient property (say, old age) is statistically correlated to a relevant end (such as productivity) and therefore treats old applicants worse in the sense of making them worse off or treating them as inferior, for example, by straightforwardly dismissing their application. Note that agents and subjects could also be collectives rather than individuals: say, a society enacting policies that affect age groups differently on the basis of widely shared or publicly communicated beliefs concerning such statistical correlations. Statistical discrimination is typically considered as an instance of direct discrimination (Lippert-Rasmussen 2014: 88). Direct discrimination is often characterized in terms of an agent’s intention to treat people differently—and indirect discrimination in terms of a ‘facially neutral’ policy that treats all the same but, non-intentionally, has disparate impact.1⁰ This orthodox dichotomy, however, confounds two distinctions. One is between differential treatment (treating older and younger people differently) and disparate impact (treating all the same when this affects older and younger people differently). The other distinction is between the intentional and non-intentional use of (agerelated) criteria. These two distinctions result in four, rather than two, distinct forms of discrimination (Berndt Rasmussen 2020). Age discrimination, understood as statistical discrimination according to condition (ii∗ ), can fall under either differential treatment or disparate impact on the intentional dimension. For example (Table 1.1: box I), an employer who sets a minimum hiring age of 25 due to neurological evidence that under-25-year-olds tend 1⁰ See Macnicol (2006: 21); Altman (2016: §3.1); Lippert-Rasmussen (2017: 3); Moreau (2017: 166–167).

20

Age Discrimination: Is It Special? Is It Wrong?

Table 1.1 Four forms of discrimination with age-related examples Differential treatment

Disparate impact of the same treatment

Intentional

(I) Intentionally using age as a criterion for differential treatment: setting a minimum hiring age of 25

Non-intentional

(III) Treating differently without intention: perceiving equally qualified older applicants as less qualified, due to implicit age bias11

(II) Employing a facially neutral criterion, ‘overqualified’, while intending its differential impact on older applicants (IV) Employing a facially neutral criterion, ‘overqualified’, without intending its differential impact on older applicants

to have underdeveloped brain regions for risk assessment, statistically discriminates against those under 25 in the differential treatment sense. (Had the employer set this age limit because of her dislike of younger people, this would not be statistical discrimination.) Or (box II), an employer who uses the criterion ‘overqualified’ because she prefers younger employees can be said to statistically discriminate against older applicants, in the disparate impact sense, if this preference is due to her belief that younger age correlates with ‘more grit’. (Had her preference instead reflected a fear of ageing, this would not be statistical discrimination.) Even collective agents can be said to statistically discriminate (against individuals or groups), for example, when a society publicly justifies a minimum employment age with reference to statistical correlations between such a policy and overall education levels. (Had this policy instead been publicly justified by reference to religious doctrine, this would not have been statistical discrimination.) According to this analysis, all statistical age discrimination is intentional (in an individual or collective sense). But not all forms of intentional age discrimination fall under the label of statistical discrimination, for example, those due to dislike, fear, or religious doctrine. And, as Table 1.1 suggests, there are also forms of age discrimination that constitute non-intentional discrimination, for example, due to implicit age bias (box III) or due to failing to consider that one’s criteria may affect different age groups differently (box IV). A further upshot is that, by teasing apart the two conflated distinctions underlying the orthodox dichotomy of direct and indirect discrimination, we get equipped with a wider, yet still unifying lens. The lens is wider in helping us capture a wider variety of phenomena that qualify as age discrimination. It is unifying in providing a coherent template for the further evaluation of this variety. Using this new lens means that we refrain from using the conflated direct/indirect distinction and instead try to state clearly which of the four forms of discrimination we are interested in, or dealing with in a specific case. 11 For empirical studies, see Cortina et al. (2013); Malinen and Johnston (2013); Derous and Decoster (2017).

Ageing without Ageism?

21

4. Wrongness of age discrimination This new taxonomy has repercussions for the moral status of discrimination. I will sketch three prominent accounts of the prima facie wrongness of discrimination and apply them to the four forms of age discrimination. For each of these three accounts, moreover, I consider whether reasons derived from age’s specialness, as identified in section 2, may outweigh such prima facie reasons.

4.1 Intention-focused accounts According to one influential account, an instance of discrimination is prima facie morally wrong when and because it is ‘premised on the belief that some types of people are morally worthier than others’, in the sense of meriting greater moral concern (Alexander 1992: 161). According to another such account, such an instance is prima facie wrong when and because it is ‘done from unwarranted animus or prejudice against persons of [a socially salient] type’ (Arneson 2006: 787f.). These accounts thus focus on the agent’s intentions: the mental states motivating their act. To illustrate: if Alex, who is 60+ is not called back for an interview, and this is due to the employer’s belief that older applicants are just not worth the trouble or to an aversion against older people (Table 1.1: box I), this constitutes prima facie wrongful age discrimination. The same holds if Alex’s application instead fails some facially neutral criterion (say, ‘overqualified’) that the employer has posited because of such mental states (box II). Now, consider some variations. For example, the employer may be motivated by affirmative egalitarian considerations: they want to give an equalizing push to younger applicants, who are not as firmly established in the job market as their older competitors. Or they go for means–end efficiency and uses age as a proxy, thus engaging in statistical discrimination. Then, the employer’s treatment of Alex turns out as not prima facie wrong since it is motivated not by the above objectionable intentions but rather by specialness considerations. As sketched so far, intentionfocused accounts provide intuitively correct judgements of a variety of cases and, moreover, can take specialness considerations into account. However, now suppose that the employer acts—intentionally—from religious doctrine that does not imply any assumptions of inferiority, aversion, or prejudice. For example, they believe that people above 60 should be given time to rest or worship instead of having to work (Table 1.1: box I). Turning down Alex’s application would then not qualify as prima facie wrong.12 Moreover, all non-intentional forms of age-based treatment—due to implicit age bias (box III) or mere failure to grasp the hiring criteria’s disparate impact (box IV)—get a moral free pass. Since they are not 12 See Arneson (2006: 801f.).

22

Age Discrimination: Is It Special? Is It Wrong?

motivated by the objectionable intentions, according to these accounts, there is no prima facie reason against them. These verdicts seem intuitively less plausible. Proponents of intention-focused accounts must either bite these bullets or considerably revise their accounts to include even ‘benign’ intentions and non-intentional forms of discrimination (e.g. framing implicit biases as aversions or deriving the failure to consider the elderlies’ vulnerabilities from beliefs about their inferiority). Now, consider only cases which do qualify as prima facie wrongful age discrimination according to intention-focused accounts. Proponents have the further option of accepting the argument from specialness such that the prima facie reasons against the treatment in question can be outweighed by other reasons due to age’s specialness. For example, while the employer’s decision not to call Alex back because of ageist prejudice is prima facie wrong, affirmative egalitarian considerations may provide an outweighing reason in favour of the decision, all things considered. However, proponents of intention-focused accounts will presumably object that bad intentions cannot be counterbalanced by good outcomes. They then deny premise (3), thus restricting specialness considerations to the prima facie level. Intention-focused accounts thus manage to capture our intuitions in some cases of intentional age-based treatment but fail in others. Moreover, they cannot differentiate between non-intentional cases but rather give them all a moral free pass. Finally, while they can account for specialness considerations when these constitute exculpating mental content on the part of the agent, they will presumably deny them the weightier status of outweighing moral reasons.

4.2 Social meaning accounts While the previous accounts focus on the intentions behind the act, the present ones focus on its meaning for others. Here, an instance of discrimination is prima facie wrong when and because it is objectively demeaning, in the sense of expressing ‘a lack of respect for the equal humanity of the other’, from a position of power that gives force to the expression (Hellman 2008: 36).13 To assess whether these conditions are met, we need to interpret the act within its social context by examining conventional methods of expressing respect and disrespect, the agent’s social status, etc. Such an assessment will take certain mental states of people—which may include the agent but also the victim, the general public, possibly a hypothetical impartial observer— into account: beliefs and attitudes concerning, for example, socially salient groups, power relations, and social norms.1⁴ But, in contrast to the previous accounts, the mental states of the agent are not the only decisive criteria here. Social meaning accounts have the potential to assess all four forms of age discrimination. For example, regardless of whether the employer’s decision to disregard Alex’s 13 See Shin (2009); Bidadanure (2016). 1⁴ According to Hellman (2008: 75), demeaningness is modestly objective: neither entirely independent of people’s beliefs and practices (strongly objective) nor totally dependent on what the majority thinks (minimally objective).

Ageing without Ageism?

23

application is motivated by blatant ageist prejudice (Table 1.1: boxes (I) or (II)) or due to unnoticeable implicit bias (box III), it is prima facie wrong if it expresses disrespect. Or, if the employer uses the facially neutral criterion ‘overqualified’, being genuinely unaware of how this affects older applicants (box IV), their action may still express disrespect if it signals that older people’s interests may safely be neglected. Note that even seemingly benign actions may be rendered objectionable if they become tainted by disrespect. For example, if a volunteer helps the elderly with an air of superiority or a patronizing demeanour, this may then render their treatment demeaning, via the uptake of others, and, hence, prima facie wrong. In the abovementioned cases, social meaning accounts deliver intuitively plausible verdicts. On the other hand, some intuitively objectionable age-based treatment might be exculpated. Consider an employer who immediately dismisses an older, clearly qualified applicant as unqualified. In a social context where everyone is aware of the ubiquity of implicit age bias, the action may then be exculpated as a merely unfortunate consequence of the employer’s all-too-human psychology rather than an expression of disrespect. This seems more problematic. The idea that the moral status of age-based treatment hinges on its social meaning within a given context has further troubling implications. Imagine a social context with commonly shared ageist beliefs, where age-based treatment is routinely seen as a ‘natural’ consequence of age-based properties. No such treatment may then even register as prima facie wrong since it never expresses disrespect. In the extreme, in a fully age-segregated society, there may be no instance of wrongful age discrimination at all. This seems intuitively questionable, to put it mildly (consider equivalent examples involving gender or age). Now, consider only cases which do qualify as prima facie wrongful age discrimination according to social meaning accounts. Proponents may, again, accept or reject that prima facie reasons against age discrimination can be outweighed by specialness considerations, as proposed in premise (3) of the argument from specialness. Here, it seems rather plausible that they should accept the argument. Specifically, premise (3) captures the dynamics involved in changes to the social meaning—and thus arguably the moral status—of age-based treatment. Consider the employer’s decision not to call back Charlie, who is under 18 but clearly qualified. In a social context, where this expresses a lack of respect for youngsters, this is then prima facie wrongful age discrimination. Now, assume that public debate arises around the social benefits of a minimum employment age of 18. Advocates point out that this would ensure that youngsters finish their basic education prior to entering the labour market, thus referring to collective sequence efficiency considerations. This means that specialness considerations are employed as outweighing reasons in favour of an age limit, along the lines of premise (3). If this debate eventually yields widespread agreement on this age limit, the prima facie status of the type of act changes: the employer’s next decision not to call back a qualified under-18-year-old applicant just because of their age will adhere to changed social norms and thus no longer express disrespect. This reasoning along the lines of the argument from specialness brings out that social meaning accounts have the advantage of being responsive to changing social

24

Age Discrimination: Is It Special? Is It Wrong?

contexts, thus keeping their verdicts closer to common intuitions (within these contexts). The flipside is that these accounts risk losing the critical distance to the social context they are supposed to evaluate—as seen in their counterintuitive assessments within the above extremely age-segregated society.

4.3 Harm accounts Both previous accounts of the wrongness of discrimination pick up on one sense of ‘being treated worse’: (b) being treated as inferior. The present account picks up on the other: (a) being made worse off. Here, an instance of discrimination is prima facie wrong when and because it harms the victim, that is, makes them fare worse as compared to what they would have had they not had socially salient property P (Lippert-Rasmussen 2014; Berndt Rasmussen 2019). Thus, harm accounts compare the actual and counterfactual outcome, relative property P, in terms of the subject’s welfare levels. Harm accounts also have the potential to assess all four forms of discrimination. Consider Alex, who is 60+ and not called back. Harm accounts are insensitive to whether the decision is motivated by open ageism (Table 1.1: box I) or implicit bias (box III), whether a facially neutral criterion is used in a premediated way (box II) or due to lack of better understanding (box IV). What counts is only the resulting harm to Alex, that she (in the relevant sense) fares worse. But does she, really? As the evidence from middle-aged Billie indicates, had Alex been middle-aged, she would have been called back. Has she thus been made worse off compared to a counterfactual state in which she had been middle-aged, rather than 60+? This may seem obvious at first. Yet, in the light of the above complete life neutrality account, there is room for doubt. Arguably, had Alex been middle-aged at this point in time (like Billie), she would have been 60+ at some later point—and, ceteris paribus, not been called for an interview then instead. Maybe we could say that Alex is made worse off compared to had she been middle-aged (like Billie) at any time of her life. Yet, this is nomologically impossible and hence cannot be a relevant counterfactual here. There is still another sense in which we could claim that Alex is made worse off, viz., compared to the counterfactual state in which no age criteria are used at all. Assessing this claim is a difficult empirical matter. Specialness considerations pertaining to individual sequence efficiency give us reason to doubt that such a claim would eventually bear out. Thus, under strict ceteris paribus conditions with individually (complete life) beneficial age criteria, the employer’s decision not to call Alex back due to her age does not make Alex worse off and is thus not prima facie morally wrong. In many cases, however, such demanding ceteris paribus conditions do not hold (e.g. employment opportunities vary over birth cohorts) or age criteria do not achieve individual sequence efficiency. Then, the employer’s decision may make Alex worse off and thus be prima facie morally wrong.

Ageing without Ageism?

25

Thus, for harm accounts, some specialness considerations—individual sequence efficiency, complete life neutrality—are relevant on the prima facie level. Others are not: for example, the affirmative egalitarian claim that mandatory retirement rules reduce inequalities between people will not matter for the assessment of whether Alex is harmed. The same goes for collective sequence considerations (that focus on collective benefits) or means–end considerations (that focus on the employer’s benefit of using age as a proxy). Thus, while all specialness considerations can be relevant at the prima facie level for intention-focused accounts (if they figure in the agent’s mental content) and for social meaning accounts (if they shape the meaning of the act), harm accounts are more selective at this level. However, even here, all specialness considerations may re-enter the picture in the role specified by premise (3) of the specialness argument, generating outweighing reasons for age-based treatment that is prima facie wrong. For example, the employer’s decision not to call Alex back may be harmful to Alex and, hence, prima facie wrong but nevertheless overall justified due to the promotion of utility (through collective sequence efficiency) or equality (through affirmative egalitarian effects). One final note: the central focus of harm accounts is the victim and the discriminating act’s effects on how they fare in life, which is intuitively appealing. Some might worry that harm accounts nevertheless miss other important factors, as covered by the two rival accounts, respectively. Still, depending on how the concept of harm is spelled out, (part of) the victim’s harm may result from, for example, their perception of the action as socially demeaning or of the agent’s intentions. The harm account may thus be able to integrate some of these factors as well.

5. Conclusion In this chapter, I have brought together considerations of the specialness of age with three influential accounts of the moral wrongness of discrimination. I have evaluated these accounts’ moral verdicts for age-based treatments and explored how they can take specialness considerations into account. In doing so, I have proposed a template for recognizing, understanding, and morally evaluating different forms of age discrimination. As a final concession, the chapter does not give determinate answers to its two title questions. My hope is that this proposed template may be useful for others interested in further pursuing these questions.

Acknowledgements Many thanks to the participants of the Prague (2017) and Madrid (2019) ‘Taking Age Discrimination Seriously’ (TADS) workshops and to an anonymous reviewer for valuable comments. Special thanks to Greg Bognar and Axel Gosseries for editorial

26

Age Discrimination: Is It Special? Is It Wrong?

support, constructive comments, and never-ending patience. Financial support by the Swedish Research Council, Riksbankens Jubileumsfond, and the Marianne and Marcus Wallenberg Foundation is gratefully acknowledged.

References Alexander, Larry. 1992. ‘What Makes Wrongful Discrimination: Wrong Biases, Preferences, Stereotypes, and Proxies’. University of Pennsylvania Law Review 141 (1), pp. 149–219. Altman, Andrew. 2016. ‘Discrimination’. In The Stanford Encyclopedia of Philosophy (Winter edn), ed. Edward N. Zalta. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/win2016/entries/discrimination. Arneson, Richard. 2006. ‘What Is Wrongful Discrimination?’ San Diego Law Review 43 (4), pp. 777–806. Berndt Rasmussen, Katharina. 2019. ‘Harm and Discrimination’. Ethical Theory and Moral Practice 22 (4), pp. 873-891. Berndt Rasmussen, Katharina. 2020. ‘Implicit Bias and Discrimination’. Theoria 86 (6), pp. 727–748. Bickenbach, Jerome. 2016. ‘Disability and Health Care Rationing’. In The Stanford Encyclopedia of Philosophy (Spring edn), ed. Edward N. Zalta. Metaphysics Research Lab, Stanford University. https://stanford.edu/entries/disability-care-rationing. Bidadanure, Juliana. 2016. ‘Making Sense of Age-Group Justice: A Time for Relational Equality?’ Politics, Philosophy & Economics 15 (3), pp. 234–260. Bognar, Greg. 2008. ‘Age-Weighting’. Economics & Philosophy 24 (2), pp. 167–189. Bognar, Greg. 2015. ‘Fair Innings’. Bioethics 29 (4), pp. 251–261. Cortina, Lilia M., Dana Kabat-Farr, Emily A. Leskinen, Marisela Huerta, and Vicki J. Magley. 2013. ‘Selective Incivility as Modern Discrimination in Organizations: Evidence and Impact’. Journal of Management 39 (6), pp. 1579–1605. Cupit, Geoffrey. 2013. ‘Age Discrimination’. In International Encyclopedia of Ethics, ed. Hugh LaFollette. https://doi.org/10.1002/9781444367072.wbiee284 Daniels, Norman. 1988. Am I My Parents’ Keeper?: An Essay on Justice between the Young and the Old. Oxford: Oxford University Press. Derous, Eva, and Jeroen Decoster. 2017. ‘Implicit Age Cues in Resumes: Subtle Effects on Hiring Discrimination’. Frontiers in Psychology 8. https://doi.org/10.3389/fpsyg.2017. 01321 Gosseries, Axel. 2014. ‘What Makes Age Discrimination Special? A Philosophical Look at the ECJ Case Law’. Netherlands Journal of Legal Philosophy 43 (1), pp. 59–80. Hellman, Deborah. 2008. When is Discrimination Wrong? Cambridge, MA: Harvard University Press. Lippert-Rasmussen, Kasper. 2014. Born Free and Equal?: A Philosophical Inquiry into the Nature of Discrimination. Oxford: Oxford University Press.

Ageing without Ageism?

27

Lippert-Rasmussen, Kasper. 2017. ‘The Philosophy of Discrimination: An Introduction’. In The Routledge Handbook of the Ethics of Discrimination, ed. Kasper LippertRasmussen. London/New York: Routledge, pp. 1–16. Macnicol, John. 2006. Age Discrimination: An Historical and Contemporary Analysis. Cambridge: Cambridge University Press. Malinen, Sanna, and Lucy Johnston. 2013. ‘Workplace Ageism: Discovering Hidden Bias’. Experimental Aging Research 39 (4), pp. 445–465. McKerlie, Dennis. 1992. ‘Equality between Age-Groups’. Philosophy & Public Affairs 21 (3), pp. 275–295. Moreau, Sophia. 2017. ‘Discrimination and Freedom’. In The Routledge Handbook of the Ethics of Discrimination, ed. Kasper Lippert-Rasmussen. London/New York: Routledge, pp. 164–173. Nussbaum, Martha C., and Saul Levmore. 2017. Aging Thoughtfully: Conversations about Retirement, Romance, Wrinkles, and Regret. Oxford: Oxford University Press. Overall, Christine. 2006. ‘Old Age and Ageism, Impairment and Ableism: Exploring the Conceptual and Material Connections’. National Women’s Studies Association Journal 18 (1), pp. 126–137. Schauer, Frederick. 2017. ‘Statistical (And Non-Statistical) Discrimination’. The Routledge Handbook of the Ethics of Discrimination, ed. Kasper Lippert-Rasmussen. London/New York: Routledge, pp. 42–53. Shin, Patrick S. 2009. ‘The Substantive Principle of Equal Treatment’. Legal Theory 15 (2), pp. 149–172. Wedeking, Gary A. 1990. ‘Is Mandatory Retirement Unfair Age Discrimination?’ Canadian Journal of Philosophy 20 (3), pp. 321–334.

2 Does the Badness of Disability Differ from that of Old Age? Kasper Lippert-Rasmussen

1. Two views of disability Here is a story that some people tell about disability, the disadvantages involved in being disabled, and our views on those matters. Some decades ago, a lot of people accepted: The bad-difference view of disability: Being disabled is worse for a person than being able-bodied. To the extent that it is worse, this is largely independent of the social environment in which disabled people live.

On this view, we should first and foremost regard disabilities as bodily or mental dysfunctions. Perhaps, so the story continues, people back then recognized that our social environment also matters to the disadvantages experienced by disabled people. However, they thought that for most disabilities, the disadvantages they involve would not be significantly different under different social circumstances.1 We cannot significantly mitigate the disadvantages involved in disabilities such as, say, chronic obstructive pulmonary disease (COPD), unipolar major depression, and advanced dementia simply by reforming our social world. Things might be slightly different when it comes to disabilities like deafness and being a wheelchair user, but even these disabilities would be quite significantly disadvantageous under more disability-friendly circumstances. Or so the general view was. One of the natural corollaries of this view is that the right way to address the disadvantages of disability is to mitigate or eliminate them through medical interventions.2 1 An obvious question here is, ‘How much more accommodating?’ Since my main interest at this point is to describe a story that is often told in relation to disability and then see if it can be told about old age as well (I do so shortly), it is of less importance here to offer precise answers to this question. However, that if by ‘more accommodating’ we mean ‘maximally accommodating’, we might well run into the problem that being maximally accommodating of disabled people requires an unjustly large part of resources that could otherwise be spent covering the needs of able-bodied people. 2 By saying that this is a ‘natural corollary’, I do not mean to give the impression that it actually follows. There is nothing incoherent, or even implausible, about saying, for example, that our ideals of beauty are oppressive and yet the state should address the problems suffered by some people who diverge significantly Kasper Lippert-Rasmussen, Does the Badness of Disability Differ from that of Old Age?. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Kasper Lippert-Rasmussen (2023). DOI: 10.1093/oso/9780192894090.003.0003

Ageing without Ageism?

29

Ideally, such interventions can transform disabled persons into able-bodied ones and even prevent the coming into existence of future disabled persons. Fortunately, so the story continues, fewer people nowadays accept this view. In its stead, the following alternative view has become increasingly popular: The mere-difference view of disability: Being disabled is worse for a person than being able-bodied. To the extent that it is worse, this is largely because of the ableist nature of the social environment in which disabled people live.³

People who subscribe to this view think that being disabled is a disadvantage in most societies across the world. However, they would explain this fact first and foremost by appeal to facts about ableist social structures rather than by appeal to bodily and mental impairments involved in disability (Barnes 2014). Specifically, they would think of being disabled in analogy with the disadvantages of being black in racist society. In itself, being black is in no way bad for you. However, it is bad for you if you live in a social world which stereotypes, excludes, and oppresses black people. In such a society, something like the mere-difference view of being black seems obviously correct. Friends of the mere-difference view of disability claim that the same is true of their view; that is, we should essentially regard disabilities and the disadvantages that they involve as the manifestation of an ableist social environment—one which matches the needs and abilities of able-bodied people much better than the needs and abilities of disabled people. An example of this is the until recently widespread absence of wheelchair access to public buildings (Moreau 2020: 56–57). If wheelchair access were universal, being a wheelchair user would be much less of a disadvantage and, given this and other accommodations, it might even not be regarded a disability for the same reasons that being unable to run 100 metres in less than 13 seconds is not regarded as a disability. Another example of ableism is the still common thought that providing wheelchair access involves ‘extra costs’, as if providing access to able-bodied people is the baseline relative to which extra costs are identified. To support this point, friends of the meredifference view of disability might point to the public provision of sports facilities for able-bodied people (e.g. large football pitches in inner cities), which, in some ways, might be more expensive than sports facilities needed by wheelchair users. Generally, such provision is not considered expensive accommodation of able-bodied people. Why not? Because our implicit baseline is set by the needs of able-bodied people. Or so the argument goes, correctly or not. Given the mere-difference view of disability, it is natural to infer that the morally appropriate way to address disability is to reform our social environment to make from those norms by offering them cosmetic surgery, for example, because of how difficult it is to change ideals of beauty in ways that remove the relevant socially constructed misfortune. 3 For a defence of a view closely related to the mere difference view, see Barnes (2014, 2016). For a reply exhibiting a commitment to the bad difference view, see Kahane and Savulescu (2016). Some theorists favour a stronger version of the mere difference view according to which being disabled is solely disadvantageous as a result of an ableist environment.

30

Does the Badness of Disability Differ from that of Old Age?

it more inclusive and, thus, less discriminatory against disabled people (though see fn 2) (Oliver 1990). Specifically, from this perspective, trying to ‘cure’ people of their disabilities, or even preventing the coming into existence of new disabled people, would appear to add insult to injury.

2. A similar pair of views about old age I now want to shift our focus from disability to old age for a moment. A pair of views about old age and the disadvantages involved in old age similar to the bad- and meredifference views of disability are easily conceivable. Consider first: The bad-difference view of old age: Being old is worse for a person than being young. To the extent that it is worse, this is largely independent of the social environment in which old people live.

Just as friends of the bad-difference view of disability regard disability as essentially being a matter of bodily or mental dysfunctions, friends of the bad-difference view of old age would say that, essentially, becoming old is a matter of increasing degrees of bodily and mental dysfunction. As Hauskeller puts it: ‘In ageing, we experience the gradual disintegration of our bodies and, often enough, our minds. Taken by itself this really is bad for us, that is, your ageing is bad for you and mine for me’ (Hauskeller 2011: 29–30). Friends of the bad-difference view of old age need not deny that the degree to which the social environment is accommodating of old people makes a difference to the degree of disadvantage that they suffer. What they deny is that the disadvantages of old age would be significantly lessened under social circumstances more accommodating of old people in just the same way that friends of the baddifference view of disability deny that the disadvantages involved in most disabilities would be significantly reduced under more accommodating social circumstances. In defence of their view, they can point to the fact that most people of, say, 80 and older have significant impairments in, for example, their vision and hearing, mobility, memory and learning ability, and in their ability to adapt to new situations; they can then stress the difficulty of imagining a social world in which such impairments are not disadvantageous (Burke and Barnes 2006).⁴ Just as friends of the bad-difference view of disability are inclined to regard disability as a medical problem to be dealt with through medical interventions, friends of the bad-difference view of old age hold that the disadvantages of old age should be mitigated (to the extent possible) through healthy lifestyles; surgery; medicine; and— ideally, in the long term—a scientific understanding of the biological processes of ageing that will enable us to slow down, or even reverse, these processes such that, in the future, people might avoid the misfortune of becoming old rather than continue to exist without bodily and mental decline. ⁴ For data on how the likelihood of disability increases with age, see the BURDIS Network Project (2004).

Ageing without Ageism?

31

In parallel to how friends of the mere-difference view of disability typically infer— validly or not—that disabled people have a complaint about their situation in ableist societies, friends of the bad-difference view of old age typically think that old people have no complaint about many, or even most, of the disadvantages that come with old age. After all, there is no one against whom this complaint can be directed since it is only natural and, to a large extent, unavoidable that, unfortunately, old people are disadvantaged in the ways that old people typically are disadvantaged.⁵ I suspect that many readers are sympathetic to the bad-difference view of old age as I have set it out here or in some not very different form. In fact, I suspect that a good portion of those readers are also sympathetic to the mere-difference view of disability and, more generally, that many more people are sympathetic to the baddifference view of old age than to the bad-difference view of disability. I suspect this because, in my view, our thinking about old age has not undergone a partial transition similar to the one from the bad-difference view of disability to the mere-difference view posited in the story set out in the previous section (Oldman 2002). However, the following view about old age is structurally identical to the mere-difference view of disability: The mere-difference view of old age: Being old is worse for a person than being young. To the extent that it is worse, this is largely because of the (anti-old people) ageist nature of the social environment in which old people live.

On this view, the bodies and minds of old people are different from that of young people. That fact disadvantages them, so, on this matter, they agree with friends of the bad-difference view, but it does so largely because of the fact that we live in social environments that, by and large, are geared towards meeting the needs of people who are not old. In a non-ageist social world (or, for that matter, in an ageist, pro-elderly social world), old age would not be significantly worse for people—if worse at all— than youth or middle-age. As with the mere-difference view of disability, a natural implication of this view is that, primarily, we should not think of the disadvantages of old age as a medical problem—let alone of ageing as something that, ideally, we should stop or reverse. Rather, we should think of it as a social problem, where the proper response to the disadvantages faced by old people should be mitigated, or even largely eliminated, through the reform of those unjust, ageist structures of the social world that disadvantage them. I am not claiming that we can easily divide people into those who accept the bad-difference and those who accept the mere-difference view of old age. Indeed, it probably has never occurred to many that there is such a view as the mere-difference ⁵ This asymmetry might be overdetermined in that people tend to think of disability as something that only an unfortunate minority is inflicted with whereas the disadvantages of old age occur to everyone who is fortunate enough not to die young. Accordingly, while the former clashes with equality across people in terms of the quality of their whole lives, the latter actually reduces this inequality. I set aside this potential source of the relevant asymmetry here.

32

Does the Badness of Disability Differ from that of Old Age?

view of old age. But because there is such a view and because, in recent decades, there has been a trend away from the bad-difference view of disability in the direction of the mere-difference view of disability, it is natural to ask whether the two pairs of view go hand in hand such that if one accepts the mere-difference view of disability, then one must also accept the mere-difference view of old age. In section 3, I shall argue that this is indeed the case. If I am right about this, we can then ask whether we should accept the mere-difference view of old age as I suspect most people, who hold a view on this matter, would not (section 4). Since I myself am one of those persons, I then ask whether this forces us, modus tollens, to reject the mere-difference view of disability (section 5). I shall argue that it does. However, drawing on the distinction between universal and generic generalizations, I also argue that, for many purposes, answering questions such as ‘Is the mere-difference(/bad-difference) view of disability (or old age) true?’ is less important than is often assumed (section 6). I consider this message to be the most important one in this chapter.

3. Why the mere-difference view of old age is true if the mere-difference view of disability is In this section, I defend the following claim: The conditional: If the mere-difference view of disability is true, then so is the mere-difference view of old age.

One thought behind this claim is that old age generally involves developing disabilities or conditions similar to disabilities (WHO 2011; Barlow and Walker 2015; Jecker 2021). One central disadvantage of becoming old is the increased probability that one will become disabled. The World Health Organization (WHO) estimates that while 10 per cent of the world’s population has some form of disability, the same is true of 50 per cent of those over age 85 (OECD 2004: 14; Chappell and Cooke 2010: 2). In light of this, it is hard to see how the bad-difference view of disability could be true if the mere-difference view of old age is true.⁶ After all, if disabilities form a major portion of the disadvantages of old age, if old age involves no comparable counterbalancing benefits, and if the disadvantages involved in old age all come down to an unaccommodating social environment, then the disadvantages involved in disability cannot obtain independently of social environment.⁷ ⁶ This would suggest that, generally, old people encounter all the disadvantages that disabled people encounter. However, in addition to those disadvantages, they might encounter other disadvantages that follow directly from being old (see the last paragraph of this section). Also, in principle, there could be interactional effects such that the state of being disabled involves fewer (or, for that matter, more) disadvantages when one is old than when one is not. I thank the editors for drawing my attention to this fact. ⁷ As I explain in section 6, the expression ‘disadvantages involved in disability’ is a problematic expression. However, set these problems aside for the moment being.

Ageing without Ageism?

33

Consider the next conditions different from but similar to disabilities. Take reduced hearing ability. Deafness is generally considered a disability in part because it prevents one from being able to take part in many social interactions. But the same is true, albeit to a lesser extent, of reduced hearing. Generally, older people have worse hearing than young people. If the disadvantages experienced by deaf people are due to an ableist social world, it is difficult to see why the same is not true of older people with impaired hearing. Similarly, if the disadvantages involved in being disabled in a way that involves wheelchair use should be conceived in accordance with the mere-difference view, it is difficult to see why we should not similarly adopt a mere-difference view of the reduced mobility of older people. Similar claims are true of many other conditions that are characteristic of old age, including slower reflexes, impaired memory, decreased muscle strength, poor balance, and low oxygen consumption. Admittedly, even if many of the disadvantages involved in old age and disability are either the same or at least of the same kind, this does not entail that the conditional is true. After all, old age might be bad in ways that neither involves disabilities nor are much affected by social environment. If so, the bad-difference view of old age might be true even if the mere-difference view of disability is true. There are manifest differences between being old and being disabled. First, unless one dies prematurely, one will eventually become old. The same is not true of being disabled, though hardly any of us escapes some form of disability if we grow sufficiently old. Second, one might live through a temporary disability. However, one cannot live through transitory old age, so that one ceases to be old.⁸ In this sense, old age is irreversible, while disability is not. However, these differences do not imply that the two bad-difference views are not tied together. For these claims are not about whether old age and disability come to most, or all, of us or are reversible. They are about the badness of old age and disability. Hence, differences in terms of (near-)universality and reversibility do not undermine the conditional. Second, it might be maintained that many of the disadvantages experienced by older people are, in a certain sense, unavoidable and, perhaps for that reason, not a disadvantage to be regretted. Perhaps this is so, but some disabilities, such as severe depression and complete paralysis, seem equally to be conditions that involve unavoidable disadvantages. Even if old age and disability differ in this respect, it is unclear why unavoidability prevents something from being regrettable, and even if it did, the distinction between being and not being unavoidable cuts across the distinction between something being bad because of the social circumstances under which it occurs and something being bad largely independently of the social circumstances under which it occurs. Some specific social conditions might be unavoidable in the relevant sense. Again, the present consideration gives us no reason to reject the conditional. Finally, it might be suggested that one way in which being disabled is bad for you lies in the fact that disabled people are subjected to all sorts of harmful and offensive ⁸ At any rate, not in a chronological sense of ‘old’. Biological age is different.

34

Does the Badness of Disability Differ from that of Old Age?

stereotypes. These clearly would be non-existent in a non-ableist social world. The disadvantages involved in being old do not similarly derive from being the victim of stereotypes. In response, it should be observed that it is not as if there are no harmful stereotypes of older people. A PEW Research Center study: finds a sizable gap between the expectations that young and middle-aged adults have about old age and the actual experiences reported by older Americans themselves. These disparities come into sharpest focus when survey respondents are asked about a series of negative benchmarks often associated with aging, such as illness, memory loss, an inability to drive, an end to sexual activity, a struggle with loneliness and depression, and difficulty paying bills. In every instance, older adults report experiencing them at (often far) lower levels than younger adults report expecting to encounter them when they grow old. (Taylor et al. 2009: 2)

Notably, there is little reason to suspect that old people are subjected to those stereotypes only if they also develop disabilities or conditions similar to disabilities. I conclude that, unless we can offer better reasons than those I have discussed here, it does indeed seem that the conditional is true, that is, that if the mere-difference view of disability is true, then so is the mere-difference view of old age.

4. Is the bad-difference view of old age true? If we find my arguments in section 3 sound and if we find the mere-difference view of disability appealing, we might suggest that we should reject the bad-difference view of old age. Should we do so? Unfortunately, the mere-difference view of old age looks quite implausible. Suppose you were offered a pill at no cost at age of 25 that would stop your body and mind ageing—meaning, for example, that your risk of Alzheimer’s would not double every five years of age after age 64 (WHO 2011: 14). The pill would not affect the length of your life. Would you decline to take it on the ground that doing so would be pointless as it would not affect you for better or worse? I suspect that almost no one would, even if they were confident that the social world of the future would be a non-ageist one in which people with slower reflexes, impaired memory, and so on enjoy maximal feasible social accommodation. If so, this suggests that being old is not a disadvantage largely because of the ageist nature of social environment in which we live.⁹ It might be conceded that most of us probably would choose to take the anti-ageing pill but insist that this provides little support for the bad-difference view of old age. ⁹ Thought experiments involving anti-disability pills have been offered by several theorists against the mere difference view of disability. Many of the responses and counterresponses to these uses apply, mutatis mutandis, here as well. My main point is that similar thought experiments bear on the mere-difference view of old age.

Ageing without Ageism?

35

The choice, it might be said, simply reflects a mid-life bias which makes us underestimate the pleasures and benefits of old age and exaggerate the troubles it involves (Jecker 2020).1⁰ Specifically, we have the idea of a ‘well-rounded life’—one that naturally arcs from childhood, through adolescence and maturity, and reaches old age. Old age is merely different from other phases of life. It plays a different, but equally important, role in relation to a life that is good as a whole, and in its absence a life is poorer. On this view, old age should not be assessed in isolation but in terms of its contribution to the greater narrative whole of which it is a part, and that contribution might be very significant even if the segment of one’s life in which one is old, considered in isolation, is worse than the segment in which one is young—also considered in isolation from the rest of one’s life. I am sympathetic to the view that, in part at least, the quality of any given segment of one’s life should be assessed holistically in view of how it fits into one’s life as a whole. Nevertheless, I find the argument above problematic in view of the fact that the idea that the value of a life, as well as segments thereof, is tied to the narrative structure, which is the life form, is consistent with life being better still if the last stage involves the agility and sharpness of mind of a 25 year-old. More generally, one could live a more well-rounded life if one were free of many of the impairments and restrictions that old age typically visits on us and which, regrettably, often interrupt the narrative of our lives unfolding in a coherent manner by rendering us unable to complete projects, fulfil intentions, and continue relationships. I conclude that the bad-difference view of old age does, indeed, seem plausible.

5. Rejecting the mere-difference view of disability? If the conditional is true and if the mere-difference view of old age is false, then, modus tollens, we must reject the mere-difference view of disability. This inference might seem appealing in the light of some of the objections that have been raised to the social model of disability. As Tom Shakespeare puts it: What would it mean to create a barrier free utopia for people with learning difficulties? Reading and writing and other cognitive abilities are required for full participation in many areas of contemporary life in developed nations. What about people on the autistic spectrum, who may find social contact difficult to cope with […] With many solutions to the disability problem, the concept of addressing special needs seems more coherent than the concept of a barrier free utopia. (Shakespeare 2017: 2019; see also Harris 2001; Singer 2005; Bognar 2016)

However, in rejecting the mere-difference view of disability, we need not go as far as accepting the bad-difference view of disability. One thing the social model of disability has taught us is that, for a wide range of disabilities, including deafness and 1⁰ See also the quote from the PEW study in section 3 (Taylor et al. 2009).

36

Does the Badness of Disability Differ from that of Old Age?

paraplegia, the level and type of disadvantage conferred by disability is significantly affected by how others react to disabilities and how unaccommodating our social environments are of these conditions. Hence, we might suggest instead: The moderate bad-difference view of disability: Being disabled is worse for a person than being able-bodied. To the extent that it is worse, this is to a significant degree dependent of the social environment in which disabled people live but also to a significant degree independent of the social environment in which they live.

This view is distinct from both the mere- and the bad-difference views of disability. Admittedly, it is imprecise to the extent that it is unclear what exactly qualifies as ‘a significant degree’. But, at least for reasonable ways of specifying that notion, the view has the attraction of being more plausible than both of the two alternative views about disability that I introduced in section 1.11 Essentially, this is so because it acknowledges that, generally, the disadvantages that disabilities involve are the result of quite a complex interaction of social and non-social factors, where none of these play an insignificant role. Still, there are two reasons why one might find the moderate bad-difference view of disability deficient. The first one is that ‘disability’ is an extremely broad category. The impairments constituting disability are many and varied, and, indeed, one prominent theme in the disability literature is whether there are any unifying features of the conditions normally considered disabilities (Goering 2015: 137). Muscular dystrophy, orthopaedic, speech and hearing impairments, visual impairments, epilepsy, cerebral palsy, mental retardation, HIV, specific learning disabilities, and diabetes are all considered disabilities for the purpose of the Americans with Disability Act.12 These involve conspicuously different disadvantages and to assess the truth of something like the moderate bad-difference view of disability (and, for that matter, the two views about disability that I introduced in section 1), we have to develop some way of getting from specific assessments of how social environment affects the disadvantages of each of these conditions to an overall judgement about how the social environment affects the disadvantages of disability in general. My main point here is not to argue that this cannot be done (though I don’t think anyone has, in fact, done anything coming even close to it). My main point is that doing so might not be particularly interesting in light of the fact that specific disabilities might be very different from disability in general, however you bridge the inferential gap between knowledge about the source of disadvantages of individual disabilities and knowledge about the source the disadvantages of disabilities in general. It might be that when it comes to deafness and paraplegia, it is, indeed, the case that, within feasible social environments, being deaf or paraplegic involve virtually no disadvantages overall compared to not having these disabilities but that when it comes to COPD, 11 The same could be said about the analogous view of the badness of old age. 12 See https://www.upcounsel.com/list-of-disabilities-covered-under-ada.

Ageing without Ageism?

37

unipolar major depression, and advanced dementia, these would be major disadvantages in any feasible social environment (Andrić and Wündisch 2015: 9; Bognar 2016: 46). If we adopt the perspective of a social planner, we are better served by knowledge of the sources of the disadvantages involved in specific disabilities rather than by knowledge about the truth of the moderate bad-difference view of disability. In fact, something similar could be said about the analogous moderate bad-difference view of old age in light of the fact that the condition of the average 65-year-old is quite different from that of the average 95-year-old. Specifically, it is more plausible to put the age-related disadvantages endured by people in their sixties largely down to social circumstances than it is to claim the same about age-related disadvantages experienced by people in their nineties—for instance, because of the frequency of Alzheimer’s among people in their nineties. Moreover, the specific conditions affecting 65- and 95-year-olds and the disadvantages these conditions involve are bound to be quite different. Specifically, some, but far from all, conditions affecting these people, especially the latter group of people, are bound to involve significant disadvantages whatever the social environment. The second reason why one might find the moderate bad-difference view of disability deficient is that, like the other views on disability and old age that I have presented, it can either be read as involving a universal or a generic generalization. On the former reading, the view is true if, and only if, of any disabled person it is true that this person’s life would have been better without disability and that this is, to a significant degree, true in virtue of this person’s social environment. On the latter reading, the view is true if it is true of a sufficiently high number of disabled people (short, and perhaps far short, of all disabled people) that their lives would have been better without disability and that this is, to a significant degree, true in virtue their social environment. On this reading, we understand the moderate baddifference view of disability in the same way we understand the statement, ‘Ticks are dangerous.’ This claim is true even if some—even most—ticks are not carriers of Lyme’s disease and, thus, not dangerous (Leslie and Lerner 2016). Now, the moderate bad-difference view of disability is much more likely to be true when read as a generic generalization—it can be true even if there are many disabled people who are not worse off for being disabled and even if, for many of those who are, this fact is, to a significant degree, dependent of their social environment. But this also means that the truth of the moderate bad-difference view of disability is much less significant. Because true and false generic generalizations abstract from features of the specific conditions over which they generalize, we need to assess whether these specific conditions are bad in themselves or not. Suppose that we adopt the perspective of a disability activist. From that perspective, it is not particularly interesting if, say, the moderate bad-difference view of disability is true since that is consistent with the fact that a lot of the disadvantages involved in disability are, to varying degrees, determined by modifiable social structures. For a similar reason, it is not particularly interesting that the view is false since that is consistent with there being many disadvantages involved in major forms of

38

Does the Badness of Disability Differ from that of Old Age?

disability, which are, indeed, largely socially determined and for that reason likely to be disadvantages that we can mitigate or even completely eliminate by eliminating ableism (cf. Barnes 2014; Kahane and Savulescu 2016). Other disadvantages cannot be significantly affected, let alone eliminated, by changing social environments and, accordingly, it would be wrong to see such disadvantages as a result of ableism. Hence, it is less important to resolve whether generic generalizations like the moderate bad-difference view of disability are true and more important to analyse the specifics of each disadvantage involved in particular forms of disability.13 If we broaden the perspective, we could say something similar about old age. This is not to deny that, for many practical purposes, generic generalizations are extremely useful, for instance, when we need to rely on heuristics and make quick decisions. Nor is it to deny that generic generalizations might serve important political ends. Generic generalizations about disability not being bad in themselves have had a large, essentially beneficial influence on societal discourse about disability. Perhaps the same benefits would arise from similar reform of societal discourse on old age. Still, social planners and activists need, and should not forget about, the specific disadvantages involved in each particular disadvantage of disability and old age.

6. Conclusion In this chapter, I have defended three main claims. First, I have argued that how we think about the disadvantages involved in disability constrains how we should think about the disadvantages of old age and vice versa. Specifically, I have argued that if the mere-difference view of disability is true, then so is the mere-difference view of old age. However, and that is my second main claim, the mere-difference view of old age is false. Old age is bad for us and is so, to a significant degree, independently of our social environment. If so, we must also give up on the mere-difference view of disability. In its stead, I have proposed the moderate bad-difference view of disability. My third, main point is that while this view might be true, it might also not be very significant whether broad-scoped, generic generalizations like this—and those involved in the other mere/bad-difference views discussed in this chapter—are true since they ignore the great variation across different individuals and different conditions. From the perspective of social planners and activists, those variations should be crucial. Also, I suggested that it might be a good thing if, in future discussions among theorists, less focus is given to the truth or falsity of generic, broad-scoped generalizations about disability and old age and more attention is paid to the disadvantages of specific conditions affecting disabled and elderly people (or, for that matter, all of us). In fact, this might be by far the most important claim of this chapter. I believe it goes quite far towards meeting the real concerns motivating many of the friends of the mere-difference views of disability and old age even if, for some, it does not 13 Other generic generalizations like this one include the mere/bad difference views of disability/old age that I introduced in section 1.

Ageing without Ageism?

39

go far enough, since it is consistent with the—as I see it, plausible—view that some disadvantages involved in some disabilities and in old age will be significant even in social environments which are very accommodating of disabled and old people.

Acknowledgements I thank Greg Bognar and Axel Gosseries for very helpful editorial and substantive comments. This chapter draws on and extends arguments developed in Kasper Lippert-Rasmussen. 2022. ‘A Puzzle about Disability and Old Age’. Journal of Applied Philosophy 39 (1): 103–116. I thank DRNF144 for financial support in relation to work on this chapter.

References Andrić, Vuko, and Joachim Wündisch. 2015. ‘Is It Bad to Be Disabled? Adjudicating between the Mere-Difference and the Bad-Difference Views of Disability’. Journal of Ethics and Social Philosophy 9 (3): 1–17. Barlow, Fiona K., and Nicole Walker. 2015. ‘Disability and Ageing’. In Encyclopedia of Geropsychology, ed. Nancy Pachana. Singapore: Springer: 674–680. Barnes, Elizabeth. 2014. ‘Valuing Disability, Causing Disability’. Ethics 125 (1): 88–113. Barnes, Elizabeth. 2016. The Minority Body. Oxford: Oxford University Press. Bognar, Greg. 2016. ‘Is Disability Mere Difference?’ Journal of Medical Ethics 42 (1): 46–49. BURDIS (Burden of Disease) Network Project. 2004. ‘Disability in Old Age. Final Report. Conclusions and Recommendations’. Finnish Centre for Interdisciplinary Gerontology, University of Jyväskylä Finland. https://ju.se/download/18.3783220012d8f123ca 58000115/1520578695703/DISABILITY%20IN%20OLD%20AGE.pdf. Burke, Sara, and Carol Barnes. 2006. ‘Neural Plasticity in the Ageing Brain’. Nature Reviews Neuroscience 7: 30–40. Chappell, Neena L., and Heather A. Cooke. 2010. ‘Age Related Disabilities—Aging and Quality of Life’. International Encyclopedia of Rehabilitation 56: 1–3. Goering, Sara. 2015. ‘Rethinking Disability: The Social Model of Disability and Chronic Disease’. Current Reviews in Musculoskeletal Medicine 8: 134–138. Harris, Joshua. 2001. ‘One Principle and Three Fallacies of Disability Studies’. Journal of Medical Ethics 27 (6): 383–387. Hauskeller, Michael. 2011. ‘Is Ageing Bad for Us?’ Ethics and Medicine: An International Journal of Bioethics 27 (1): 25–32. Jecker, Nancy S. 2020. Ending Midlife Bias. New York: Oxford University Press. Jecker, Nancy S. 2021. ‘The Time of One’s Life’. History and Philosophy of the Life Sciences 43.

40

Does the Badness of Disability Differ from that of Old Age?

Kahane, Guy, and Julian Savulescu. 2016. ‘Disability and Mere Difference’. Ethics 126 (3): 774–788. Leslie, Sarah, and Adam Lerner. 2016. ‘Generic Generalizations’. In Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/generics/. Moreau, Sophia. 2020. Faces of Inequality. Oxford: Oxford University Press. OECD (Organisation for Economic Co-operation and Development). 2004. ‘Dementia Care in 9 OECD Countries: A Comparative Analysis’. OECD Health Working Papers. Paris: OECD. Oldman, Christine. 2002. ‘Later Life and the Social Model of Disability: A Comfortable Partnership?’ Ageing & Society 22: 791–806. Oliver, Michael. 1990. The Politics of Disablement. London: Palgrave Macmillan. Shakespeare, Tom. 2017. ‘The Social Model of Disability’. In The Disability Studies Reader, ed. Lennard J. Davis. New York: Routledge, pp. 195–203. Singer, Peter. 2005. ‘Ethics and Disability: A Response to Koch’. Journal of Disability Policy Studies 16 (2): 130–133. Taylor, Paul, Rich Morin, Kim Parker, D’Vera Cohn, and Wendy Wang. 2009. ‘Growing Old in America: Expectations vs. Reality’. PEW Studies, 29 June. https://www. pewresearch.org/wp-content/uploads/sites/3/2010/10/Getting-Old-in-America.pdf. WHO (World Health Organization). 2011. Global Health and Ageing. Bethesda, MD: World Health Organization.

3 In Defence of Age-Differentiated Paternalism Viki Møller Lyngby Pedersen

1. Introduction In many countries, children and adolescents are not allowed to buy tobacco and spirits, to get a tattoo, or to use sunbeds. Other examples of regulations targeting this age group are compulsory education and child labour laws. Such restrictions appear paternalistic as they restrict the liberty of minors to act imprudently in self-regarding matters. However, almost everyone agrees that paternalism of this kind is justifiable. According to Mill, ‘[t]hose who are still in a state to require being taken care of by others, must be protected against their own actions’ (Mill 1859 [2011]: 9). Paternalism towards adults, on the other hand, is much more controversial (Feinberg 1986: 23–24; Anderson 1999: 301–302; Shiffrin 2000). Examples include laws against gambling, prostitution, smoking, extreme sports, or driving without a seatbelt. There is agreement that paternalism is easier to justify when targeting young people. Exceptions, however, occur at both ends of the age spectrum: certain paternalistic procedures seem permissible in elder care, and sometimes one should avoid interfering with a child’s imprudent actions. An important purpose of this chapter is to shed light on different ways in which age affects the justification of paternalism. Specifically, I discuss two potentially important factors for the justification of paternalism: competence and the good promoted by paternalistic policies and actions. By systematically showing how each factor is sensitive to age, I defend age-differentiated paternalism. Central to the discussion of when paternalism is justified is the debate between hard and soft paternalists. Hard paternalists accept certain interventions in the voluntary choices of competent individuals. Soft paternalists only accept interfering either when people are insufficiently competent or, if sufficiently competent, when their actions are insufficiently voluntary (Grill 2020: 4). Accordingly, a person’s competence or the degree of voluntariness associated with an action, according to soft paternalists, is crucial to assessing whether paternalism towards that person is justified. Children are not yet competent. This contributes to explaining why many paternalistic procedures towards children seem unproblematic. Some adults, including some elderly people, are also, to some extent, incapable of deciding for themselves Viki Møller Lyngby Pedersen, In Defence of Age-Differentiated Paternalism. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Viki Møller Lyngby Pedersen (2023). DOI: 10.1093/oso/9780192894090.003.0004

42

In Defence of Age-Differentiated Paternalism

and find themselves ‘in a state to require being taken care of by others’ (Mill 1859 [2011]: 9). However, I argue that soft paternalism towards children will more often be warranted than towards older adults. The good promoted by paternalistic policies or actions (including the bad that would have occurred in their absence) represents a second factor that plays a role in assessing whether such policies or actions are justified. For example, many proponents of hard paternalism hold that we are justified in interfering with competent people’s voluntary actions if the stakes are profound and the infringement involved is tolerable.1 I present arguments to the effect that the good promoted by paternalism is graduated by age. Specifically, I argue that the number of (good) life years at risk of being lost and the number of years lived are central to assessing the potential harm involved in a given imprudent activity. In his 1972 article, Gerald Dworkin understands paternalism as ‘interference with a person’s liberty of action justified by reasons referring exclusively to the welfare, good, happiness, needs, interests or values of the person being coerced’ (1972: 65). Since then, many have made efforts to refine the definition of paternalism (e.g. Shiffrin 2000: 218; Grill 2012; Dworkin 2017). An important adjustment has consisted in broadening the scope of the concept and including interferences with a person’s autonomy that are not necessarily coercive (Dworkin 1988: 107). Hereinafter, I shall discuss policies that satisfy the conditions of this broader understanding of paternalism.

2. Competence In his classic piece On Liberty, Mill formulates and defends his critique of paternalism. However, in Mill’s view, we may justifiably prevent people from harming themselves if they are not acting voluntarily. He illustrates this point with the case of a person who is about to cross a ramshackle bridge. If we doubt that the person knows about the condition of the bridge, we are justified in holding the person back until we are certain that the person is acting voluntarily (Mill 1859 [2011]: 93). This is the germ of what Feinberg later labelled ‘soft paternalism’. Soft paternalists accept interfering with a competent person’s imprudent behaviour when the behaviour is not ‘voluntary enough’ (Feinberg 1986: 118) ‘or when temporary intervention is necessary to establish whether it is voluntary or not’ (1986: 12). They also accept interfering with an insufficiently competent person’s plans. According to Feinberg, soft paternalism amounts to a qualified form of antipaternalism (1986: 12). When people are not competent or when their choice is not voluntary, it is, in a sense, not their own choice but something foreign to them. Hence, interfering with it to avert harm from befalling them does not constitute an infringement of their autonomy (1986: 12–14). Whether or not one accepts the soft paternalistic principle of never interfering with voluntary actions performed 1 See, among others, Dworkin (1988: 127) and Pedersen and Midtgaard (2018: 775).

Ageing without Ageism?

43

by competent agents, soft paternalistic restrictions generally seem easier to justify than hard paternalistic ones. The latter interfere with competent people’s voluntary behaviour. When is a choice sufficiently voluntary? Feinberg suggests that a ‘perfectly voluntary’ choice is free from coercion or duress, (subtle) manipulation, ignorance, mistaken belief, and circumstances that are temporarily distorting (1986: 115, diagram 20–5).2 However, few, if any, choices require that all these elements be satisfied fully to qualify as sufficiently voluntary (1986: 115). Both voluntariness and competence are matters of degree, and the degree of non-voluntariness or incompetence required to justify paternalism depends on factors relating to the activity in question such as the risk and seriousness of harm and its irrevocability (1986: 117–124). The concern for ensuring voluntariness seems to be reflected in the Danish sterilization law. To be allowed to undergo sterilization in Denmark, you must be at least 18 years old. If you are between 18 and 25 years old, you must wait at least six months from deciding to undergo the surgery until the surgery is actually performed. This reflection period can be seen as a way of establishing that the decision to sterilize is sufficiently voluntary. Competence, according to Buchanan and Brock (1986), is ‘to be understood as decision-making capacity’ (1986: 22). This includes capacities that enable the individual to receive information, express wishes, and understand potential consequences (1986: 22–26). Moreover, in order to evaluate outcomes, the person must have a reasonably consistent value set or conception of the good (1986: 26). The incapacities typical of childhood explain why soft paternalists consider paternalism warranted towards persons under a certain age (Feinberg 1986: 183). Children have not adequately developed their own dispositions and reflective capacities. Thus, they tend (like the man on the unsafe bridge) to act without knowing or understanding (the consequences of) what they are doing. In this way, children are incompetent relative to many important decisions, which is why others (most often the parents) must decide on their behalf (Feinberg 1986, 183; Buchanan and Brock 1989: 229). Paternalism becomes more problematic as the child grows older and develops the capacities needed for competence.3 A similar argument can be made for some older people. The issue of paternalism towards the elderly can be a sensitive one. Ageing is frequently associated with stereotypes, and restrictions targeting the elderly should obviously not be based on prejudices. On the other hand, it is important that we understand and respond to the special needs that often arise with age. This is especially true if we see paternalism not only as warranted under certain conditions but also as a duty in some cases (see, e.g. Pedersen and Midtgaard 2018: 775; Pedersen 2021). For far too many, old age is associated with some significant deterioration in cognitive capacities, including certain forms of dementia. Moreover, chronic diseases are common among the elderly, and 2 Feinberg also includes lack of competence as a voluntariness-vitiating factor (Feinberg 1986: 115, diagram 20–5). However, as it seems that an incompetent person can be acting voluntarily, I consider voluntariness and competence as separate matters. 3 I exclude here instances where people never develop the relevant capacities.

44

In Defence of Age-Differentiated Paternalism

many experience impaired cognitive functions in the final stages of such disorders (Buchanan and Brock 1986: 17–18). Hence, according to Feinberg’s soft paternalistic principle, we may (have to) protect the incompetent elderly from some of their selfregarding conduct. For example, relatives may be justified (or even morally required) in forcing incompetent adults who, in various ways, pose a risk to themselves into care homes.⁴ Still, age seems to be a more reliable proxy for competence at a young age than at an old age. Due to biological circumstances, all children (at least below a certain age) lack decision-making capacities. As the child grows older, the relevant capacities will usually develop and our soft paternalistic reasons will vanish. At a certain later age, the capacities needed for competence will be substantially reduced again for some. If a person’s competence is reduced to below the sufficiency threshold (corresponding to the activity being performed), there will again be a case for soft paternalism. Whether this happens, however, varies greatly and depends on several factors such as illness or accidents. Hence, in contrast to the case of children, soft paternalism towards older adults cannot be implemented as a general rule. It must be adopted on a single-case basis in order to establish that people interfered with are, in fact, insufficiently competent. The upshot is that soft paternalism based on insufficient competence seems easier to justify as a general approach towards children than towards older adults.⁵

3. Competence-focused objections to paternalism: Best judgement and development Focusing on competence, we have just seen how the justification of soft paternalism varies with people’s age. In this section, I will show how age affects the relevance of two competence-focused objections to paternalism: the best-judge argument and the development argument. The arguments are competence-focused in the sense that they are both objections to drawing implications in favour of paternalism from the fact that people have difficulties judging what is good for them. In brief, the bestjudge argument denies that others should be more competent on the matter, and the development argument claims that paternalism is a further source of people’s incompetence. According to Mill’s best-judge argument, paternalistic regulations will do more harm than good because people themselves are the best judge of what will promote their own interests and well-being. There is a predominant risk that paternalists will ⁴ The incompetent elderly and children should plausibly not be subject to the same paternalistic policies. Those who are properly considered to be incompetent in some domains may have sufficient decisionmaking capacities in other areas (Buchanan and Brock 1989: 218, 280). For example, some Alzheimer’s patients may pose a risk to themselves without assisted living facilities but still be competent enough to vote in parliamentary elections. For more on the right to vote by people with cognitive impairments, see Beckman (2014). ⁵ Here, I leave the potential relation between age and voluntariness aside. However, age may affect the propensity of competent individuals to make voluntary choices in different ways. One example is if adolescents are more vulnerable to peer pressure than other age groups.

Ageing without Ageism?

45

not succeed in promoting other people’s good because they do not have the necessary insights to do so (Mill 1859 [2011]: 74, 81). In response to this objection, many have expressed scepticism about the claim that an individual is always the best judge of what is in his or her best interest. Moreover, even when this is actually the case, people often fail to act in ways that promote their interests (Thaler and Sunstein 2008: 9; Quong 2011: 97; Hanna 2018: 44). There are many findings about human errors. We are biased, spontaneous, and unrealistically optimistic; we make short-termist decisions; we procrastinate; we are influenced by default settings; and we are bad at predicting consequences (Thaler and Sunstein 2008). Sometimes, reasoning failures may impair the quality of our choices to the effect that they can no longer be characterized as sufficiently voluntary. In such cases, policies seeking to improve our choices may be defended with appeal to soft paternalistic reasons. The prevalence of human errors also seems relevant to the justification of hard paternalism. Even when reasoning failures do not lead to our choices being insufficiently voluntary, they challenge Mill’s argument to the effect that hard paternalistic policies would typically produce more bad than good for people (see also Dworkin 1972: 70–74). Still, from the fact that people have difficulties judging or achieving what is in their best interest, it does not follow that others are more competent on the matter; and the relevance of this concern may increase with people’s age. As John Kleinig puts it: Although elderly people do not always perceive what is in their interests as well as they might (a failing to which we are all subject), they, unlike young children, nevertheless confront the world with a wealth of experience and a rich set of attitudes, understandings, and interests, and more often than not, are likely to be better positioned with respect to determining their good than those who would otherwise determine it for them. (1983: 167–168)

Whether people actually become better at ‘determining their good’ with age is an empirical issue. In support of Kleinig’s view, it has been shown that older adults tend to have more ‘regular’ preferences than young adults do (‘[a] decision-maker is “irregular” if she would choose B from {A, B, C} but not from {A, B}’ [Tentori et al. 2001: 87]). Such findings seem to entail that both soft and hard paternalism becomes harder to justify the older the people get. Soft paternalism becomes less relevant as competence and voluntariness increase. As for hard paternalism, it seems less likely that the good promoted by unwanted restrictions substantially exceeds the bad inflicted by it when people are adept at determining their own interests. Moreover, as regards the sterilization example, studies show ‘a consistent inverse relationship between women’s age at sterilization and their likelihood of regretting having had the procedure’ (Curtis et al. 2006: 205). This suggests that women become better at determining their interests with respect to sterilization with age. That a person regrets her choice may, but need not, be an indication that she was insufficiently

46

In Defence of Age-Differentiated Paternalism

competent or acted insufficiently voluntarily at the time of her decision. This is important for soft paternalists. In their view, the state or others should not interfere with a competent person who makes a voluntary decision to be sterilized on the basis that she is likely to regret her choice. This latter justification would only be available to hard paternalists. However, in some areas, people’s level of experience does not seem to follow their age. For example, many elderly people struggle with technology that young people find easy to deal with. This can lead to older people being more vulnerable to internet fraud such as phishing scams. Thus, even though older people have more life experience and therefore may ‘be better positioned with respect to determining their good’ (Kleinig 1983: 167–168), they sometimes seem to have a harder time acting in ways that actually promote their good. Hence, the best-judge objection to paternalism does not always fare better at older ages than at younger ages. According to the development argument, paternalism hampers people’s development. Freedom of choice has educative value (Mill 1859 [2011]: 56). We learn from our mistakes. This argument seems sensitive to age as there seems to be a relevant difference between young people’s and adults’ interest in personal development. As argued in section 2 on soft paternalism and competence, children have not yet acquired the capacities and knowledge that seem important to formulating and pursuing their own perception of the good life. The process of acquiring such capacities plausibly requires some freedom of choice, including the freedom to make bad choices (Buchanan and Brock 1989: 230). The development argument seems amplified when related to a certain child liberationist argument for equal rights for children. According to Godwin (2011), children’s capacities to exercise their liberty rights are plausibly impaired by the way in which children are already prevented from exercising such rights (2011: 263, 268). If it is true that people’s development is hindered by paternalism (which, in many societies, is frequently used towards children), it may be argued that some of the incapacities we associate with childhood are something we impose on children, that is, they do not arise from biological circumstances. The interest in development seems essential for children; however, the freedom to learn from mistakes also seems important to those adolescents who can no longer be said to be incompetent relative to their imprudent decisions (and towards whom restrictions would be hard paternalistic). Older adults, of course, also have an interest in personal development, but they are not in the process of building the foundation for their future exercise of autonomy to the same extent. While the best-judge argument tends to render paternalism more difficult to justify towards older adults, the development argument might lead to the reverse conclusion. There are, however, different ways in which the development argument may be contested. First, it might be challenged that the developmental value of liberty should lead to age-differentiated policies. For example, Kleinig argues that we should take care not to restrict the freedom or autonomy of the elderly at the cost of their ability to form a stimulating existence for themselves. He writes, ‘[w]hat the retired person needs is not moth-balling but access to new fields and spheres of activity in which his

Ageing without Ageism?

47

or her experience and creativity can be expressed and nurtured’ (Kleinig 1983: 168). Although it takes different forms depending on age, there is plausibly room and need for personal development in all stages of life. Second, sometimes the lessons we learn from our imprudent choices are prohibitively expensive (Hanna 2018: 53). A person may have learned a lot from having sold a bodily organ, but the value of this lesson may be limited since the choice has also resulted in a significant loss of future options, including opportunities for further development (Hughes 1998: 90). Moreover, some addictive activities (e.g. smoking) provide immediate pleasure but potentially lead to harm at a later stage. This raises doubt as to whether easy access to cigarettes will tend to provide the relevant learning in the youngest years of life when the interest in personal development seems most important.⁶ Accordingly, the developmental value of making mistakes often seems insufficient to oppose paternalism. Finally, paternalism may sometimes promote, rather than hinder, learning. For example, some acts within typical educational systems, especially in the education of young people, seem to involve paternalism (Birnbacher 2015). It is likely that few would deny that there is educational value in making young people attend their classes. I will end this section by summarizing the results. The conclusions from the bestjudge argument and the development argument point in different directions. The best-judge argument seems to provide some reason to interfere less with people the older they are. People’s level of experience typically increases with age, and the more life experience they gain, the more problematic it might be to assess, on their behalf, what would be in their best interest. However, there are exceptions. In some areas, older people seem to have more difficulties than younger people in actually acting in ways that advance their interests. As regards the development argument, young people seem to have a stronger general interest in keeping a wide range of options open in order for them to develop and pursue their own perception of the good life. When paternalism limits these opportunities for development, this provides reason for interfering less with people the younger they are. Nevertheless, the relationship between paternalism and development seems to be highly dynamic. Sometimes, paternalism plausibly protects, or even fosters, people’s development.

4. Good promoted In what follows, I argue that the good promoted (including the bad prevented) by paternalism tends to diminish with people’s age. The argument bears on the fact that, in general, the additional life expectancy of young people is larger than the additional life expectancy of older people. Hence, compared to older adults, young people risk missing out on advantages for a longer period of time when they act imprudently. ⁶ For an in-depth normative analysis of the smoking case, see Goodin (1989).

48

In Defence of Age-Differentiated Paternalism

The argument is not related to competence or voluntariness. If two people with identical degrees of competence and voluntariness are behaving equally imprudently, the argument implies that paternalism towards the youngest person will be more warranted than paternalism towards the older one. As we shall see, this applies to paternalism in both its soft and its hard version. Imagine the following two scenarios: Young motorcyclist: A competent 18-year-old person is voluntarily riding a motorcycle without wearing a crash helmet. The motorcyclist crashes and incurs serious, irreversible brain damage. Old motorcyclist: An equally competent 90-year-old person is just as voluntarily riding a motorcycle without wearing a crash helmet. The motorcyclist crashes and incurs equally serious, irreversible brain damage.

In which scenario would the good promoted (including bad prevented) by a law requiring motorcyclists to wear helmets be greatest? I take it that the case of the young motorcyclist is significantly worse. What explains this intuition? There are different ways in which it can be bad if the risk of a person’s imprudent conduct materializes. First, imprudent activities may lead to conditions that are intrinsically bad, for example, pain (Kagan 2012: 210). A person’s imprudent activities may also lead to conditions that are bad for others, for example, if other people incur economic or emotional costs (Feinberg 1986: 139–141). However, in this chapter, I am interested in the relevance of people’s age for the paternalistic justification, which traditionally appeals to the good promoted for the person interfered with (not the good promoted for others). For the same reason, it is also irrelevant from the perspective of age-differentiated paternalism that older adults tend to have more dependents than younger people (even though it may be relevant from the perspective of an all-things-considered assessment). Moreover, if we hold all these intrinsic and other-regarding types of badness constant across the two scenarios, the young motorcyclist case still seems worse. The following argument provides one plausible explanation of this judgement. The crash of the young motorcyclist seems worse for the young motorcyclist than the crash of the old motorcyclist would be for the old motorcyclist because of differences in additional life expectancy.⁷ First, if the crash reduces additional life expectancy, it deprives the young motorcyclist of more years of living. Second, even if the crash does not affect additional life expectancy, it deprives the young motorcyclist of more years of good quality.⁸ ⁷ Here, I am inspired by Kagan (2012: 211). ⁸ Age also seems relevant to a narrower hard paternalistic justification. Many argue that we are sometimes justified in interfering with a person’s short-term autonomy in order to ensure the person’s long-term autonomy (see, e.g. Mill 1859 [2011]: 99–100). However, when a person’s additional life expectancy is short, it seems less justifiable to interfere in order to ensure the person’s future autonomy.

Ageing without Ageism?

49

The young motorcyclist could reasonably expect to live on without brain damage for many years to come.⁹ The old motorcyclist could perhaps also expect to live on without brain damage for some years to come. However, the older the person gets, the less reasonable it seems to regret missing out on any more (non-brain-damaged) years of living.1⁰ In marginal cases, we may know for certain that the young person would not have a greater additional life expectancy than the older person. In such cases, the paternalistic policy would not promote the young person’s good to the same extent as under common circumstances. However, even when there are no differences in additional life expectancy, it still seems that the 18-year-old motorcyclist who would have lived to only 25 years old loses more in a crash than the 90-year-old who could have lived until the age of 97. One plausible explanation, which I shall briefly outline, appeals to differences in life-year opportunities.11 Compared to the older motorcyclist, the younger motorcyclist has had fewer life-years to pursue life plans (e.g. less time for education, career, family life, other self-realization, etc.). For the same reason, it seems more detrimental to a person’s quality of life to lose additional years to pursue life plans the fewer life-years for such the person has already had. Additional life-year opportunities seem to have a greater value for people when added earlier in life.12 A variant of this argument may be applied in the sterilization case as well. Young women who choose to be sterilized at an early age have had fewer years at a fertile age, and thus fewer life-year opportunities to have children, than women who choose to be sterilized later (possibly after they have already had one or more children). If a paternalistic policy (e.g. one that forces women to wait at least six months before being sterilized) protects people from making an irreversible decision they will later regret, the argument implies that the policy is more beneficial to their quality of life the younger they are.

5. Relevance of age for soft and hard paternalism What are the implications of the arguments from section 4 for the importance of age in relation to the justification of hard paternalism and soft paternalism, respectively? Many proponents of hard paternalism think that we are justified in interfering with a competent person’s voluntary choice if the good promoted for the person interfered ⁹ I assume that the helmetless 18-year-old motorcyclist is the same person as, say, the brain-damaged 40-year-old. However, the argument is complicated if we accept the ‘complex view’, according to which the 40-year-old is ‘a later self, one that has been injured by the actions of an earlier self ’ (Kleinig 1983: 46). In that case, the analysis would arguably presuppose an evaluation of the psychological connectedness between the respective selves (see Parfit 1984: 206). 1⁰ I write ‘less reasonable’ and not ‘unreasonable’ because one can still regret that we will all die. 11 Here, I am inspired by Bognar (2015: 259) and Nielsen (2021, 231–232). 12 There are plausibly other relevant ways in which the good promoted varies with age, for example, if elderly people generally incur more serious and irrevocable harms by, say, a crash than young people (who are physically better able to recover).

50

In Defence of Age-Differentiated Paternalism

with substantially exceeds the bad inflicted by the policy. If it is the case that the good promoted diminishes with people’s age (roughly speaking), then the good promoted might be more likely to substantially exceed the bad inflicted the younger people are. A possible policy implication may be that young people should receive a higher fine than older people if they are caught by the police driving without a crash helmet (given that the amount of the fine reflects its effect in discouraging helmetless driving). Similarly, the state could let the price of cigarettes decrease with age. While soft paternalists are primarily concerned about protecting the insufficiently competent or insufficiently voluntary (and refuse to protect competent people against voluntarily inflicted self-harm), their position still takes the good promoted into account. According to soft paternalists, the degrees of incompetence and nonvoluntariness required to justify paternalistic policies depend on the seriousness and risk of harm and its irrevocability (Feinberg 1986: 117–124). This implies that we can set a high bar for how voluntary an action must be if the action potentially causes self-harm of a great magnitude. In view of this, the argument of this section implies that policymakers should pay more attention to ensuring voluntariness the younger people are. The Danish policy on sterilization seems to reflect this consideration as the reflection period of six months applies only to young people.

6. Conclusion Most paternalistic policies can be introduced in age-differentiated variants. For example, the rules on sterilization may involve prohibitions for the youngest and forced reflection periods for the second youngest. Moreover, the fine for driving a motorcycle without a crash helmet, as well as the price of cigarettes, could increase with age. While there may be reasons (e.g. of a practical nature) for not supporting such age differentiation, such policies find support from the factors addressed in this chapter. First, a ban for the youngest seems justified as it is plausible to assume that children are insufficiently competent relative to the decision to sterilize, drive without a helmet, or smoke. (However, it is debatable where exactly the age limit should be set.) Moreover, it may be considered worse to interfere with people’s self-regarding decisions the more life experience and greater insight into their own interests they have achieved. On the other hand, it may sometimes be considered particularly problematic to interfere with young people, who should have a wide range of options open for them to develop and pursue their own life plans. However, in situations involving a risk of severe and irreversible harm, the developmental value of learning from one’s mistakes often seems to be limited or to come at disproportionate cost. Second, young people generally risk losing more (good) years of living than older people do. Furthermore, irrespective of differences in additional life expectancy, preventing a person from losing additional years (of good quality) to pursue life plans seems to be more valuable the fewer life-years the person has

Ageing without Ageism?

51

already had. For these reasons, there seems to be a potentially greater good associated with preventing younger people from incurring risks. These points are consistent with my overall conclusion that the justification of paternalism generally weakens as the people interfered with advance in age. The competence-based concern and the good-promoted-based concern often converge. However, in specific cases, one or more of the justificatory elements will point towards a different conclusion. For example, some Alzheimer’s patients may justifiably be forced to live in care homes with appeal to their weakened decision-making capacities; and sometimes, young people must be allowed to learn from their mistakes. Therefore, it is important to be aware of the distinct ways in which the justification of paternalism is sensitive to age. The two factors discussed in this chapter vary in their relevance for soft and hard paternalism. Competence and voluntariness are not necessary to hard paternalists, who would sometimes be willing to interfere with competent people’s voluntary choices. By contrast, competence and voluntariness are essential to soft paternalists, who refuse to count it as a reason for interfering with a competent person’s voluntary choice that such interference will promote the person’s interests, good, or well-being. Still, as we have seen, the soft paternalistic position nevertheless incorporates a consideration for the good that the imprudent person puts at risk. When the size of this good is great, the degree of competence and voluntariness should be so too. Despite the differences between soft and hard paternalism, it applies to both positions that paternalism is generally harder to justify towards older people than towards younger people.

Acknowledgements For helpful comments on this chapter, I would like to thank Greg Bognar, Axel Gosseries, Refia Kaya, Louis Larue, Kasper Lippert-Rasmussen, Soren Flinch Midtgaard, Lauritz Munch, Lasse Nielsen, Jens Jørund Tyssedal, Manuel Valente, and an anonymous referee. This work was funded by the Danish National Research Foundation (DNRF144).

References Anderson, Elizabeth. 1999. ‘What is the Point of Equality?’ Ethics 109 (2): 287–337. Beckman, Ludvig. 2014. ‘The Accuracy of Electoral Regulations: The Case of the Right to Vote by People with Cognitive Impairment’. Social Policy and Society 13 (2): 221–233. Birnbacher, Dieter. 2015. ‘Paternalism in Education and the Future’. In The Nature of Children’s Well-Being. Children’s Well-Being: Indicators and Research 9, ed. Alexander Bagattini and Colin Macleod. Dordrecht: Springer, pp. 107–122. Bognar, Greg. 2015. ‘Fair Innings’. Bioethics 29 (4): 251–261.

52

In Defence of Age-Differentiated Paternalism

Buchanan, Allen, and Dan Brock. 1986. ‘Deciding for Others’. Milbank Quarterly 64: 17–94. Buchanan, Allen, and Dan Brock. 1989. Deciding for Others: The Ethics of Surrogate Decision Making. Cambridge: Cambridge University Press. Curtis, Kathryn, Anshu Mohllajee, and Herbert Peterson. 2006. ‘Regret Following Female Sterilization at a Young Age: A Systematic Review’. Contraception 73: 205–210. Dworkin, Gerald. 1972. ‘Paternalism’. The Monist 56 (1): 64–84. Dworkin, Gerald. 1988. The Theory and Practice of Autonomy. Cambridge: Cambridge University Press. Dworkin, Gerald. 2017. ‘Paternalism’. In The Stanford Encyclopedia of Philosophy (Winter edn), ed. Edward Zalta. https://plato.stanford.edu/entries/paternalism. Feinberg, Joel. 1986. Harm to Self. New York: Oxford University Press. Godwin, Samantha. 2011. ‘Children’s Oppression, Rights and Liberation’. Northwestern Interdisciplinary Law Review 4: 247–302. Goodin, Robert. 1989. ‘The Ethics of Smoking’. Ethics 99 (3): 574–624. Grill, Kalle. 2012. ‘Paternalism’. In Encyclopedia of Applied Ethics (2nd edn), ed. Ruth Chadwick. London: Elsevier Inc., pp. 359–369. Grill, Kalle. 2020. ‘Respecting Children’s Choices’. Moral Philosophy and Politics (published online ahead of print). doi: https://doi.org/10.1515/mopp-2019-0046. Hanna, Jason. 2018. In Our Best Interest. Oxford: Oxford University Press. Hughes, Paul. 1998. ‘Exploitation, Autonomy, and the Case for Organ Sales’. International Journal of Applied Philosophy 12 (1): 89–95. Kagan, Shelly. 2012. Death. New Haven, CT: Yale University Press. Kleinig, John. 1983. Paternalism. Totowa, NJ: Rowman & Allanheld. Mill, John Stuart. 1859 [2011]. On Liberty. Northampton: White Crane Publishing. Nielsen, Lasse. 2021. ‘Contractualist Age Rationing under Outbreak Circumstances’. Bioethics 35 (3): 229–236. Parfit, Derek. 1984. Reasons and Persons. Oxford: Oxford University Press. Pedersen, Viki. 2021. ‘Respectful Paternalism’. Law and Philosophy 40 (4): 419–442. Pedersen, Viki, and Soren Midtgaard. 2018. ‘Is Anti-Paternalism Enough?’ Political Studies 66 (3): 771–785. Quong, Jonathan. 2011. Liberalism without Perfection. New York: Oxford University Press. Shiffrin, Seana. 2000. ‘Paternalism, Unconscionability Doctrine, and Accommodation’. Philosophy and Public Affairs 29 (3): 205–250. Tentori, Katya, Daniel Osherson, Lynn Hasher, and Cynthia May. 2001. ‘Wisdom and Aging: Irrational Preferences in College Students but Not Older Adults’. Cognition 81 (3): 87–96. Thaler, Richard, and Cass Sunstein. 2008. Nudge. London: Penguin Books.

4 Age and the Social Value of Risk Reduction Three Perspectives Matthew D. Adler

1. Introduction How does an individual’s age figure into the social value of reducing their fatality risk? This is an important question for policymakers, in two ways. First, in a situation of scarcity, policymakers may need to allocate (ration) risk-reducing interventions among the population—and the question then arises whether an individual’s age should be one of the criteria that drives the allocation. For example, vaccines were quickly developed to protect against COVID-19, but the limited vaccine supply and logistics precluded immediately vaccinating the entire population even of the most affluent countries (let alone the whole world); and governments regularly used age as one basis for vaccination, with older individuals taking priority.1 Second, the social value of fatality risk reduction as a function of age comes into play when policymakers are balancing the costs of risk reduction against the benefits. To continue with the COVID-19 example: social-distancing policies that shut down or slowed down parts of the economy reduced individuals’ risks of dying from COVID-19 but at the cost of lost income to workers and business-owners (as well as other costs, depending on the specific intervention). If the value of risk reduction indeed depends on age, then a proper accounting of the benefits of a given social-distancing policy—for purposes of comparing those benefits to the policy’s costs—will need to consider how the risk reduction that the policy would deliver correlates with individuals’ ages. In this chapter, I compare how three policy-assessment frameworks value fatality risk reduction in light of age. The three frameworks are: monetized cost–benefit analysis (abbreviated as ‘CBA’), utilitarianism, and prioritarianism. The first two are well established. CBA is widely used by governments, and its theory and practice are supported by a large body of scholarship in economics.2 Utilitarianism is well established in the academic literature, if not governmental practice.3 Prioritarianism is a newer 1 See http://ourworldindata.org/covid-vaccinations. 2 See sources cited in Adler (2019: 39). 3 See sources cited in Adler (2019: 158). Matthew D. Adler, Age and the Social Value of Risk Reduction. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Matthew D. Adler (2023). DOI: 10.1093/oso/9780192894090.003.0005

54

Age and the Social Value of Risk Reduction: Three Perspectives

approach but has important advantages over CBA and utilitarianism—both in general (Adler 2012, 2019; Adler and Holtug 2019; Adler and Norheim 2022), and specifically as a metric of the social value of risk reduction, which will be my focus here. In what follows, I first describe the three frameworks, then discuss their implications for risk reduction, and finally illustrate the points elaborated in the second part with a COVID-19 example.

2. CBA, utilitarianism, prioritarianism 2.1 CBA CBA evaluates governmental policies by summing individuals’ monetary equivalents (Freeman et al. 2014; Boardman et al. 2018; Adler 2019). A policy is some intervention by government, be it enacting a regulation, building infrastructure, disseminating information, increasing taxes and/or spending, etc. Some baseline is chosen (typically, the status quo: what occurs if government leaves in place existing policies and does nothing new). CBA assumes that each individual’s preferences are such that they have a well-defined willingness-to-pay or willingness-to-accept for any given policy P. Monetary equivalents are then defined in terms of these quantities. (i) If the individual is better off with policy P than with the baseline scenario, then (CBA assumes) the individual has a well-defined willingness-to-pay for P, namely, the amount of money deducted from their income in P which just suffices to make them indifferent between that policy and the baseline. In this case, the individual’s monetary equivalent for P is positive—capturing the fact that they are better off with the policy—and equal to their willingness-to-pay amount. (ii) If the individual is worse off with policy P than with the baseline scenario, then (CBA assumes) the individual has a well-defined willingness-to-accept for P, namely, the amount of money added to their income in P which just suffices to make them indifferent between the policy and the baseline. In this case, the individual’s monetary equivalent for P is negative—capturing the fact that they are worse off with the policy—and equal to (−1) times their willingness-to-accept amount.⁴ (iii) If the individual is equally well off with P and the baseline, their monetary equivalent for the policy is zero. CBA then assigns each policy a score equalling the sum total of monetary equivalents and ranks policies in the order of these scores. Any policy with a positive sum-of-monetary-equivalents is better than baseline; any policy with a negative sumof-monetary-equivalents is worse; and one policy is better than a second if it has a larger sum. CBA came to the fore in economics at a time when there was widespread scepticism in the discipline about interpersonal comparisons of well-being (Adler and Posner 2006). This scepticism, in part, concerns the puzzle of comparing the ⁴ For example, if they are worse off with P but adding $700 to their income in P would make them indifferent between P and baseline, their monetary equivalent for P is −$700.

Ageing without Ageism?

55

well-being of individuals with different preferences. Consider first two individuals, Ariela and Batu, with the same preferences—a preference here meaning a ranking of bundles of individual attributes (income, health, leisure) and lotteries over such bundles. If Ariela has bundle a and Batu a∗ , then we can readily compare their well-being. If the common preference is such as to prefer a to a∗ , then Ariela is better off; if a∗ to a, then Batu is better off; and if indifferent, then the two are equally well off. But now imagine that there are two individuals, Caty and Dov, with different preferences. Caty has a, which she prefers to a∗ ; Dov has a∗ , which he prefers to a. Who is better off? CBA requires no interpersonal well-being comparisons. As long as we can model the population as divided into groups of similarly situated individuals, individuals in each group having the same preferences, we can calculate each group’s willingness-topay or willingness-to-accept for a given policy using the preferences of that group. We thereby translate the policy’s effect on individuals in each group into monetary equivalents, relative to baseline. We compare policies by comparing sum totals of monetary equivalents—not by any direct commensuration of the well-being of groups with different preferences. CBA uses money as the scale for measuring the impact of policies on individuals because those impacts can be measured on that scale—via the concept of monetary equivalent—without requiring interpersonal well-being comparisons. To be sure, CBA does make interpersonal comparisons of money. But economists who reject interpersonal well-being comparisons, while embracing CBA, are not committed to seeing CBA’s monetary measure of policy impact as a scale of wellbeing. Instead, such economists might argue that we need to measure policy impacts on a monetary scale rather than a scale of well-being just because money is an interpersonally comparable concept while interpersonal well-being comparisons are impossible (Adler 2019: 30–37).

2.2 Utilitarianism By contrast with CBA, utilitarianism does require interpersonal well-being comparisons. Let us use the term ‘outcome’ to mean a possible social outcome, that is, a description of what might happen to each person (or to each group of similarly situated individuals). Utilitarianism proceeds as follows. In a given outcome x, each person is assigned a well-being number depending on their attributes in x and their preferences. Outcome x is then assigned a social value equalling the sum total of well-being across individuals. The well-being numbers summed up by the utilitarian rule are interpersonally comparable. If Leila is assigned a larger well-being number in x than Mathias in y, this indicates that Leila in the first outcome is better off than Mathias in the second. A utilitarian rule that adds up the well-being numbers of different persons but denies that interpersonal well-being comparisons are possible is incoherent, for reasons I have elaborated elsewhere (Adler 2019: 41–47). If individuals are modelled as having the same preferences, assigning these interpersonally comparable well-being numbers is (conceptually) relatively straightforward. Economics shows how a given preference R (again, a ranking of attribute

56

Age and the Social Value of Risk Reduction: Three Perspectives

bundles and lotteries over such bundles) can be represented by a so-called utility measure uR (∙). The subscript ‘R’ indicates that this is the utility measure corresponding to preference R. If the preference is such as to rank a over a∗ , then uR (a) > uR (a∗ ); if a∗ over a, then uR (a∗ ) > uR (a); and if indifferent, then uR (a) = uR (a∗ ). The well-being number of a given bundle is just its utility according to a utility measure representing the common preference (Adler 2019: Chapter 2). If the simplifying assumption of common preferences is dropped, utilitarianism needs to address the Caty/Dov case above (section 2.1). It needs to provide some normatively defensible account of how to compare and measure the well-being of persons with different preferences. This challenge can be surmounted (or so I have argued elsewhere) (Adler 2016; Adler 2019: Chapter 2; Adler and Decancq 2022). I have written thus far about the utilitarian rule with respect to outcomes. For purposes of policy assessment, we need to rank policies. The outcome of a given policy may well be uncertain. Policy P might lead to outcome x, outcome y, etc. What is needed, here, is a procedure for applying utilitarianism under uncertainty. The academic literature concurs that utilitarianism should be applied under uncertainty as follows: choices (such as governmental policies) should be ranked according to the sum of individuals’ expected well-being (Adler 2019: Chapters 4–5).

2.3 Prioritarianism While utilitarianism originated in the eighteenth century with the writings of Jeremy Bentham (Eggleston and Miller 2014), prioritarianism is a much more recent entrant to ethics. The idea was introduced into the philosophical literature towards the end of the twentieth century, most visibly in work by the philosopher Derek Parfit (Parfit 2000).⁵ Note that utilitarianism is insensitive to the distribution of well-being. Assume that we have to choose between increasing the well-being of a given person by some increment ∆w (in well-being terms) and increasing the well-being of a better-off person by some increment ∆w∗ . If ∆w∗ is greater than ∆w, utilitarianism prefers to give ∆w∗ to the better-off person—even if they are much better off than the worse-off one and even if ∆w∗ is only slightly larger than ∆w. Prioritarianism rectifies this feature of utilitarianism by giving priority to wellbeing changes affecting individuals at lower levels of well-being. Formally, such priority is effectuated by plugging individuals’ well-being numbers into a concave transformation function (see Figure 4.1) and assigning each outcome a value equalling the sum total of transformed well-being. Imagine that Jasmine is worse off than Kat. We have to choose between giving a well-being increment ∆w to Jasmine (at well-being level wJ ) or the very same ⁵ See Holtug (2017) for an overview of the philosophical literature on prioritarianism. For a formal presentation of the approach, see Adler (2012, 2019); Adler and Norheim (2022). For a defence against competing philosophical views, see Adler and Holtug (2019).

Ageing without Ageism?

57

g(wK + Δw) g(wK)

Transformed wellbeing, g(w) g(wJ + Δw)

g(wJ)

wJ

wJ + Δw

wK

wK + Δw

Well-being, w

Figure 4.1 A prioritarian transformation function Source: author.

increment to Kat (at well-being level wK ). Utilitarianism is indifferent between these two allocations, but prioritarianism prefers to give the well-being improvement to Jasmine. As Figure 4.1 illustrates, the increase in Jasmine’s transformed well-being that occurs if we allocate ∆w to her is greater than the increase in Kat’s transformed well-being if we give ∆w to Kat. While utilitarianism corresponds to one specific rule for ranking outcomes, according to the sum total of individual well-being, prioritarianism is a family of such rules. Prioritarianism ranks outcomes according to the sum total of transformed well-being. There are many possible concave transformation functions. By selecting a specific function, we define a specific rule within the prioritarian family. The selection of this transformation function is a matter for ethical judgement. As it becomes more and more concave (i.e. as its slope decreases at a faster and faster rate), prioritarianism gives increasing priority to the worse off. Consider again the case of two individuals, one better off than the second, where we are choosing between giving a larger increment ∆w∗ to the better-off one and a smaller increment ∆w to the worse-off one. While utilitarianism chooses the first option, we can always find a sufficiently concave prioritarian transformation function to prefer the second—regardless of how much larger ∆w∗ is than ∆w. Prioritarianism, like utilitarianism, requires interpersonally comparable wellbeing numbers. Whatever method we employ for assigning such numbers in the utilitarian case, we can employ the same method in the prioritarian case.

58

Age and the Social Value of Risk Reduction: Three Perspectives

I noted above that the utilitarian rule under uncertainty—for ranking policies, the outcomes of which are uncertain—is to sum up individuals’ expected well-being numbers. The analogous rule, for prioritarianism, is to sum up individuals’ expected transformed well-being numbers (Adler 2019: Chapters 4–5).

2.4 Lifetimes and sub-lifetimes An important question has not yet been addressed. Indeed, this question is quite critical, if we are considering how age differences in the population should figure in policy assessment. Well-being can be ascribed on a lifetime basis or, instead, on a time-slice basis. Consider two possible outcomes x and y, and a particular individual, Jelani. We can ask whether Jelani is better off, over his entire lifetime, in x as compared to y. Alternatively, we can pick out some time-slice of Jelani’s life (a year, a week, a moment) and ask whether Jelani in x during that time-slice is better off than Jelani in y. Implicit in my presentation above of utilitarianism and prioritarianism was the assumption that these methodologies operate on lifetime well-being numbers. For a given outcome x, we assign each individual a single number, measuring their lifetime well-being in x; and with these lifetime numbers, we assign x a utilitarian value (the sum total of lifetime well-being) or a prioritarian value (the sum total of transformed lifetime well-being). Utilitarianism and prioritarianism could, alternatively, be specified on a time-slice basis. Whether to do so raises profound questions about age and ethics of distribution (see Paul Bou-Habib (Chapter 5) and Axel Gosseries (Chapter 6), this volume). I have argued elsewhere that lifetime well-being, not time-slice well-being, should be the currency for utilitarian and prioritarian assessment (Adler 2012: Chapter 6). This argument rests upon the continuity of personal identity over a normal human lifetime, upon the difference for distributive purposes between intra- and interpersonal compensation,⁶ and upon the arbitrariness of dividing a life into time-slices. Rather than recapitulating the argument here, I will simply work within the frameworks of lifetime utilitarianism and prioritarianism. What this chapter will now discuss is how lifetime utilitarianism and prioritarianism, as compared to CBA, take account of age in assigning social value to risk reduction.

3. Fatality risk reduction Section 2 described three different policy-evaluation methodologies: CBA (the sum of individuals’ monetary equivalents); utilitarianism (the sum of individuals’ ⁶ I can be compensated for a bad time-slice of my life by a better time-slice later on while you, a stranger to me, can’t be compensated for a bad time-slice of your life by a better time-slice for me.

Ageing without Ageism?

59

expected lifetime well-being); and prioritarianism (the sum of individuals’ expected transformed lifetime well-being). Let us now compare how these methodologies value risk reduction.⁷

3.1 A simulation model with survival curves I will illustrate the three methodologies by using a simulation model that is based upon the US income distribution and survival curve.⁸ An individual’s ‘survival curve’, as that term is used here, means a profile of annual survival probabilities beginning with the individual’s current age. A ‘survival probability’ means the probability of surviving to the end of a given year, conditional on being alive at the beginning. For an individual of a given age (say, 30), the survival curve tells us: the probability of surviving to age 31; the probability of surviving to age 32, conditional on surviving to 31; the probability of surviving to age 33, conditional on surviving to age 32; and so forth on to the maximum possible age. I model the population as divided into 35 groups: individuals falling into 7 different current ages (ages at the present: 20, 30, 40, 50, 60, 70, and 80), with each age divided into 5 different income quintiles. Combining 7 ages and 5 incomes, we end up with 35 age–income groups. The current-year incomes of the 35 groups are displayed in Table 4.1.⁹ Each group is assigned a survival curve, based upon the overall US survival curve, which is then adjusted to match data regarding life expectancy by income group. So as to facilitate the construction of the well-being measure required by utilitarianism and prioritarianism, and also to simplify the comparison of those approaches Table 4.1 Current-year incomes of the 35 groups Income ($) Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

8,331 22,098 28,426 28,681 24,930 19,719 14,757

11,425 30,306 38,984 39,334 34,189 27,043 20,238

15,686 41,607 53,522 54,003 46,939 37,128 27,784

21,827 57,896 74,476 75,145 65,316 51,664 38,662

51,152 135,680 174,536 176,103 153,069 121,075 90,605

⁷ On this topic, see generally Adler, Hammitt, and Treich (2014); Adler (2019: Chapter 5, 2020); Adler, Ferranna, Hammitt, and Treich (2021); Hammitt and Treich (2022). ⁸ See Adler (2020) for a detailed description of the model; tables here are taken from that publication. ⁹ The reader will note that income in each quintile increases up to age 50, then declines. The time path of income is taken, once more, from US data.

60

Age and the Social Value of Risk Reduction: Three Perspectives

to CBA, I assume that all individuals in the population (individuals in all 35 groups) have the same preferences—specifically, in this model, the same preferences over life histories, modelled as longevity–income bundles. A life history specifies what might happen to an individual over their entire lifetime. In this stripped-down model, this means specifying a lifespan and, for each year alive, an income amount. The utility of a given life history, as calculated by the utility measure u(∙) representing the common preferences, equals the sum total of the logarithm of income for each year alive.1⁰ This utility measure is both the basis for calculating the lifetime well-being numbers used by utilitarianism and prioritarianism and the basis for calculating monetary equivalents of the various groups (defined in terms of current income). For example, consider a life history with a longevity of 70 years where the individual lives with an income of $10,000 for the first 20 years; $30,000 for the next 20 years; and $40,000 for the next 30 years. The utility of this life history is 708.28:11 20 × log ($10, 000) + 20 × log ($30, 000) + 30 × log ($40, 000) . There are major simplifications here. In reality, individuals have heterogeneous preferences, and quality of life for each year alive depends upon much more than income. However, this simplified set-up suffices to illustrate the key differences between CBA, utilitarianism, and prioritarianism—differences that would carry over to a more realistic and complicated model. In the status quo, each group has an income profile (income for each year of life if alive) and survival curve. This combination of income profile and survival curve defines a baseline lottery over life histories for the group. A fatality-risk-reduction policy P changes the baseline survival curve for one or more groups. (In particular, P will increase survival probability, in the current year and/or future years, for one or more groups.) P will also likely change some groups’ baseline income profiles. By altering a given group’s survival curve and income profile, P produces a different lottery over life histories than the baseline lottery.

3.2 The social value of risk reduction: CBA, utilitarianism, prioritarianism As just explained, a policy in the simulation model here will have two types of effects: a change in the survival probability of one or more groups in the current year and/or future years and a change in the income profile of one or more groups. The overall value of a policy—as per CBA, utilitarianism, and prioritarianism—will depend upon both effects. 1⁰ The logarithm of income is a very standard utility measure in economic modelling. The model uses the natural logarithm, but the results would be the same were the logarithm to the base 10 or any other base used. 11 Strictly speaking, the model normalizes the logarithm of annual income by subtracting the logarithm of a subsistence level of income (taken to be $1,000). For further discussion, see Adler (2019: 292–294); Adler (2020).

Ageing without Ageism?

61

To isolate how the risk-reduction component of a policy affects social value as calculated by CBA, utilitarianism, and prioritarianism, let us consider a ‘pure’ risk reduction policy that increases a group’s survival probability in the current year. It will be most illuminating to begin with utilitarianism. Table 4.2 displays the utilitarian value of a small increase ∆p in a given group’s current-year survival probability. Table 4.2 should be read as follows. The utilitarian value of a change ∆p in a given group’s current-year survival probability is the resultant increase in that group’s expected lifetime well-being.12 The number in the cell for each group is the utilitarian value of a change ∆p in the current-year survival probability of someone in that group relative to the utilitarian value of that very same change in the current-year survival probability of someone in the 80-year-old, low-income group. (Note that the entry in the cell for the 80-year-old, low-income group is 1.) For example, the number in the cell for a 50-year-old, middle-income individual is 6.0. This means that a reduction ∆p in current-year mortality risk for someone in that group has 6.0 times the utilitarian value of the very same reduction for the 80-year-old, low-income individual. As Table 4.2 shows, the utilitarian values of risk reduction decrease with age. Within each income quintile (each column), the numbers become smaller as age increases. For example, in the second income quintile, the values decrease from 9.9 (age 20) to 8.6 (age 30), 6.9 (age 40), 5.2 (age 50), 3.6 (age 60), 2.3 (age 70), and 1.3 (age 80). Why is this? Consider the simplest case, in which an individual in one or another of these groups would die for certain during the current year, absent a policy intervention, and the intervention is a life-saving intervention ensuring that they survive the year. The utilitarian value of this life-saving intervention is just the difference between (i) the individual’s expected lifetime well-being if they survive the year

Table 4.2 Utilitarian value of increase ∆p in current-year survival probability (relative to utilitarian value of that increase for 80-year-old, low-income group) Income Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

8.6 7.4 5.9 4.4 3.0 1.9 1.0

9.9 8.6 6.9 5.2 3.6 2.3 1.3

11.2 9.7 7.9 6.0 4.3 2.7 1.6

12.5 10.8 8.8 6.7 4.8 3.1 1.8

15.6 13.5 11.0 8.6 6.2 4.2 2.5

12 Recall that utilitarianism scores policies by summing groups’ expected lifetime well-being numbers; and so the utilitarian value of ∆p is just the increase in this score that results from the risk change.

62

Age and the Social Value of Risk Reduction: Three Perspectives

and (ii) their realized lifetime well-being if they die now.13 Because life expectancy remaining decreases with age—a 20-year old who survives the year has a much longer life expectancy remaining than a 50-year-old who does so; in turn longer than an 80year-old who does so—the utilitarian value of life-saving decreases with age, holding constant quality of life (in our model, income).1⁴ Coming now to the more general case of a risk reduction: the utilitarian value of a ∆p increase in a given group’s survival probability is just ∆p times the utilitarian value of the life-saving intervention. Because the utilitarian value of the life-saving intervention decreases with age within each income quintile, the utilitarian value of risk reduction does too. The pattern of utilitarian value as a function of age in Table 4.2 is intuitively plausible and consistent with surveys of citizen preferences (Dolan et al. 2005: 202). Ceteris paribus, utilitarianism prefers to allocate a life-saving intervention or, more generally, a given unit of fatality risk reduction to a younger individual. What is less intuitively plausible is another aspect of Table 4.2—namely, that utilitarian values increase with income, holding constant age. Within each row (holding constant age), the utilitarian values become larger as we move to the right. For example, in the age 50 row, utilitarian values increase from 4.4 (low income) to 5.2 (moderate income), 6.0 (middle income), 6.7 (high income), and 8.6 (top income). Why is this? If Ariela and Bey are the same age and have the same survival curve, but Bey has a higher income, the utilitarian value of a life-saving intervention to save Bey is greater than the utilitarian value of a life-saving intervention to save Ariela. Because the well-being of each year alive increases with income (after all, more income is a good thing), the well-being gap between living an incremental year and not being alive that year is larger for those with larger incomes. Since the utilitarian value of lifesaving increases with income, so (more generally) does the utilitarian value of a ∆p reduction in fatality risk.1⁵

13 Recall that we are calculating lifetime well-being, in this model, as the sum of the logarithm of income for each year alive. Each individual, with a given age, is endowed with a survival curve (a list of survival probabilities for the current year and each succeeding year) and an income profile (specifying the income amount that they have earned in each prior year, and that they will earn in the current year and each succeeding year, if they are alive during that year). From this information, we can calculate both their realized lifetime well-being if they die now and their expected lifetime well-being if they survive the current year. 1⁴ Life expectancy remaining does not invariably decrease with age, but it does in the model here (see Adler 2020: table 9) and generally will do so as long as survival probability does not increase with age. See Adler, Ferranna, Hammitt, and Treich (2021). In the model here, income within each quintile is not constant with age but rather first rises and then falls. See Table 4.1. If income increased sufficiently sharply with age, utilitarian risk-reduction values might increase with age even though life expectancy remaining decreases. This does not occur in the model here; in Table 4.2, utilitarian risk-reduction values decrease uniformly with age, as would occur if income were constant in each quintile. 1⁵ Incomes in this model are after-tax-and-transfer incomes for the five quintiles of the US population (Adler 2020). The model assumes that the individual either (i) receives income if they survive to the end of a given year, and if so reaps well-being from the income (measured as the logarithm of income); or (ii) dies and does not receive the income. A third possibility (namely, that the individual dies but still receives the income, which is distributed to their heirs or the rest of the population) is not considered in the model. In the case of investment income, a more realistic model would consider this third possibility. The third possibility does not arise for wage income: if the individual is not alive during a time period, they cannot exchange their labour for wage income.

Ageing without Ageism?

63

Let us turn to Table 4.3, which illustrates how CBA values risk reduction for the different groups. Table 4.3 displays the monetary equivalents for a ∆p reduction in current-year fatality risk, normalized in a manner analogous to Table 4.2. ‘1’ indicates the monetary equivalent for a ∆p reduction in fatality risk for an 80-year old, low-income individual. The number in every other cell is the ratio between the monetary equivalent of an individual in that group for a ∆p reduction, and the monetary equivalent of an 80-year-old, low-income individual. For example, the number in Table 4.3 in the cell for a 50-year-old, middle-income individual is 19.0. This means that a 50-year-old, middle-income individual has a monetary equivalent for a given ∆p reduction in risk that is 19.0 times the monetary equivalent of an 80-year-old, low-income individual for the same reduction. CBA’s metric of value is just the monetary equivalent. What Table 4.3 shows, then, is the CBA value of risk reduction to various individuals as a function of their age and income. Comparing Table 4.3 to Table 4.2, two differences emerge. First, CBA value (the monetary equivalent) increases with income much more steeply than utilitarian value. Consider individuals aged 50. As noted above, utilitarian values for low, moderate, middle, high, and top income are, respectively, 4.4, 5.2, 6.0, 6.7, and 8.6. By contrast, CBA values for low, moderate, middle, high, and top income are, respectively, 7.4, 12.0, 19.0, 29.6, and 88.3. To appreciate more concretely what these numbers mean, consider, for example, that the utilitarian value for the top-income, 50-year-old individual is (8.6/4.4) = 1.95 times that of the low-income 50-year old. So, utilitarianism determines that reducing the low-income individual’s risk by ∆p is equally valuable as a smaller reduction of the top-income individual’s risk, namely (∆p/1.95). By contrast, the CBA value for the top-income, 50-year-old individual is (88.3/7.4) = 11.93 that of the low-income, 50-year-old individual. So, CBA determines that reducing the low-income individual’s risk by ∆p is equally valuable as a much smaller reduction of the top-income individual’s risk, namely (∆p/11.93).

Table 4.3 CBA value of increase ∆p in current-year survival probability (relative to CBA value of that increase for 80-year-old, low-income group) Income Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

4.2 9.6 9.9 7.4 4.5 2.3 1.0

6.6 15.2 15.7 12.0 7.4 3.9 1.7

10.3 23.6 24.6 19.0 11.9 6.3 2.9

15.9 36.4 38.1 29.6 18.8 10.0 4.6

46.5 106.7 112.4 88.3 56.9 31.0 14.6

64

Age and the Social Value of Risk Reduction: Three Perspectives

As explained earlier, an individual’s monetary equivalent when they are better off equals their willingness-to-pay. What Table 4.3 shows is that willingness-to-pay increases more steeply with income than utilitarian value. Why is this? Utilitarian value increases because the gain in expected lifetime well-being from saving a richer person from death is larger than the gain in expected lifetime well-being from saving a poorer person. CBA increases with income for this reason (the rich are willing to pay more for risk reduction because they reap a larger utility gain in forestalling premature death) and for an additional reason: an incremental dollar has a smaller utility impact for a richer as opposed to a poorer person (the so-called diminishing marginal utility of income). A richer individual’s willingness-to-pay for a given utility gain ∆u is larger than that of a poorer individual since the utility loss from expenditure is smaller for the richer individual. The second difference between Table 4.3 and Table 4.2 is that the utilitarian value of risk reduction decreases smoothly with age (within each column, the numbers decrease moving down), while this is not true of CBA value. Within each column, CBA value has a ‘hump’ shape, increasing with age until age 40 and then decreasing. The explanation, once more, is the diminishing marginal utility of income. Because income is not constant with age (see Table 4.1), different age groups are willing to pay different amounts for the very same utility gain ∆u. Summing up: utilitarian risk-reduction values decrease with age because older individuals have less life expectancy remaining but also increase with income. Many will find this latter feature of utilitarianism to be quite troubling.1⁶ CBA departs from utilitarianism in two ways: its valuations are even more sharply skewed towards the rich, and its valuation of risk reduction as a function of age reflects the time path of the marginal utility of income.1⁷ Finally, Table 4.4 illustrates the prioritarian value of risk reduction. To implement prioritarianism, we need to choose a transformation function encapsulating the degree of priority to the worse off. I do so using a function that (generally) suffices to neutralize the utilitarian skew towards the rich.1⁸ The prioritarian value of a change ∆p in a given group’s current-year survival probability is the resultant increase in that group’s expected transformed lifetime well-being. As in the CBA and utilitarian tables, the numbers in Table 4.4 are normalized so that 1 represents the value (here, prioritarian value) of a ∆p change in the current-year survival probability of someone in the 80-year-old, low-income group. To understand these values, consider the simplest case of a life-saving intervention. An individual would die for certain, absent an intervention which ensures that they survive the year. The prioritarian value of this intervention is the difference between 1⁶ Emanuel et al. (2020: 2051), discussing the public health literature, writes: ‘Consensus exists that an individual person’s wealth should not determine who lives or dies.’ 1⁷ The US government in practice departs from textbook CBA by using a population-average monetary equivalent. See Adler (2019: Chapter 5, 2020). This avoids the dramatic skew with income in Table 4.3 but also means that the valuation of risk reduction is independent of age. 1⁸ I use an ‘Atkinson’ transformation function with the priority parameter γ = 1.5. See Adler (2019: 154– 158, 2020). Note that, in Table 4.4, the prioritarian value of risk reduction decreases with income except at ages 70 and 80.

Ageing without Ageism?

65

Table 4.4 Prioritarian value of increase ∆p in current-year survival probability (relative to prioritarian value of that increase for 80-year-old, low-income group) Income Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

31.5 19.7 11.9 7.0 4.0 2.1 1.0

29.6 18.6 11.4 6.9 4.0 2.2 1.1

27.9 17.7 11.0 6.8 4.0 2.3 1.1

26.3 16.7 10.5 6.5 4.0 2.2 1.1

23.2 14.8 9.5 6.0 3.7 2.2 1.1

(i) the individual’s expected transformed lifetime well-being if they survive the year and (ii) their realized transformed lifetime well-being if they die now. More generally, the prioritarian value of a ∆p increase in a given group’s current-year survival probability is just ∆p times the prioritarian value of the life-saving intervention. Note that prioritarian values decrease with age more steeply than utilitarian values within each column. Utilitarian values decrease with age because life expectancy remaining decreases with age. Prioritarian values decrease with age for this reason and for the additional reason that prioritarianism (applied to lifetime well-being, as here) gives extra weight to gains to lifetime well-being that accrue to those who can expect to end up at a lower level of lifetime well-being. As between a younger person in a given column in Table 4.4 (with a given lifetime income profile) and an older person in the same column (with the same lifetime income profile), the younger person has a lower level of expected lifetime well-being. There is some risk that they will die before they reach the age of the older one. If Sara has a lower level of expected lifetime well-being than Timur, and we have to choose between a given increment to expected lifetime well-being for Sara and the very same increment for Timur, utilitarianism is indifferent—but prioritarianism prefers that the increment be allocated to Sara since she can expect to be worse off in lifetime terms.1⁹ Prioritarian values are generally flat or decreasing with income, by contrast with utilitarian values. For example, in Table 4.2, the utilitarian value for risk reduction in the 40-year-old row increases from 5.9 (low income) to 6.9 (moderate income), 7.9 (middle), 8.8 (high), and 11.0 (top). Prioritarian values decrease from 11.9 (low income) to 11.4 (moderate income), 11.0 (middle), 10.5 (high), and 9.5 (top). Richer individuals at a given age can expect to be better off in terms of lifetime well-being than poorer individuals of the same age. While a ∆p increment to the current-year survival probability of the richer individual yields a bigger increase in expected 1⁹ See Adler, Ferranna, Hammitt, and Treich (2021), providing a formal analysis of this ‘fair innings’ effect.

66

Age and the Social Value of Risk Reduction: Three Perspectives

lifetime well-being (see above), that increment is downweighted by prioritarianism because it accrues to someone who can expect to be better off in lifetime terms. In short, prioritarianism accentuates utilitarianism’s preference for the young in risk reduction and (with a sufficient degree of priority for the worse off) can neutralize its preference for the rich.

4. A COVID-19 example In this section, I illustrate how the differences between the methodologies in valuing risk reduction play out with respect to a key aspect of COVID-19 policy, namely, allocating vaccines.2⁰ A striking feature of COVID-19 is the stark difference by age in the ‘infection fatality risk’ (IFR): the risk of dying from the disease if infected. Table 4.5 provides one estimate of these IFRs.21 Continuing with the simulation model introduced above: consider the problem of distributing a scarce COVID-19 vaccine among the population of 35 age–income groups. Which groups should come first? Vaccinating an individual has a double benefit with respect to fatality risk reduction: it reduces the risk that they die from infection and the risk that they infect others who die as a result. Let us focus on the first type of benefit. Here, the social value of vaccinating a person is a function of three factors: (i) the reduction in infection risk that comes from vaccination, (ii) the reduction in fatality risk from a given reduction in infection risk (which depends on the person’s IFR), and (iii) the social value of fatality risk reduction. Ignoring the first factor, let us imagine that vaccination produces the very same reduction in infection risk for individuals in all 35 groups. Then, the social value of vaccination (as per CBA, Table 4.5 COVID-19 infection fatality risk (IFR) Age

Infection Fatality Risk (%)

20 30 40 50 60 70 80

0.03 0.08 0.15 0.6 2.2 5.1 9.3

2⁰ See Ferranna, Sevilla, and Bloom (2022) for a much more detailed analysis 21 These estimates are taken from an influential report released early in the COVID-19 pandemic, in March 2020. They were used in estimating baseline survival probabilities by age in the simulation model here, as described in Adler (2020) and so, for consistency, are used to estimate the social value of vaccination. Later estimates of IFRs by age are broadly consistent with Ferguson et al. (2020). See the meta-analysis in Levin et al. (2020).

Ageing without Ageism?

67

utilitarianism, or prioritarianism) for a given group is just the IFR for that group multiplied by the social value of risk reduction (as per that methodology). Concretely, we take the social-value-of-risk-reduction numbers in Tables 4.2, 4.3, or 4.4 and multiply by the IFRs for each age in Table 4.5 to determine the social value of vaccination. These values are displayed in Table 4.6 (utilitarianism), Table 4.7 (CBA), and Table 4.8 (prioritarianism). The values are normalized so that 1 indicates the social value of vaccinating an 80-year-old, low-income individual. Within a given table, a higher number for one group as compared to a second means that vaccinating the first group has higher social value and thus that this group should be higher up the priority list for allocation as compared to the second. Utilitarian social values of vaccination generally increase with age (the numbers in Table 4.6 get larger moving down in each column, except as between 70- and 80year-olds in the low-income column). Even though the utilitarian social value of a given unit ∆p of risk reduction decreases with age (Table 4.2), the sharp increase in IFR by age means that, on balance, utilitarianism prefers to vaccinate an older person as compared to a younger one. Indeed, even though the prioritarian social value of risk reduction decreases yet more quickly with age than utilitarianism (Table 4.4), the increase in IFR by age is sufficiently pronounced that, on balance, prioritarianism

Table 4.6 Utilitarian social value of vaccination Income Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

0.03 0.06 0.10 0.28 0.71 1.04 1.00

0.03 0.07 0.11 0.34 0.85 1.26 1.30

0.04 0.08 0.13 0.39 1.02 1.48 1.60

0.04 0.09 0.14 0.43 1.14 1.70 1.80

0.05 0.12 0.18 0.55 1.47 2.30 2.50

Table 4.7 CBA social value of vaccination Income Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

0.01 0.08 0.16 0.48 1.06 1.26 1.00

0.02 0.13 0.25 0.77 1.75 2.14 1.70

0.03 0.20 0.40 1.23 2.82 3.45 2.90

0.05 0.31 0.61 1.91 4.45 5.48 4.60

0.15 0.92 1.81 5.70 13.46 17.00 14.60

68

Age and the Social Value of Risk Reduction: Three Perspectives Table 4.8 Prioritarian social value of vaccination Income Age

Low

Moderate

Middle

High

Top

20 30 40 50 60 70 80

0.10 0.17 0.19 0.45 0.95 1.15 1.00

0.10 0.16 0.18 0.45 0.95 1.21 1.10

0.09 0.15 0.18 0.44 0.95 1.26 1.10

0.08 0.14 0.17 0.42 0.95 1.21 1.10

0.07 0.13 0.15 0.39 0.88 1.21 1.10

prefers to vaccinate an older person as compared to a younger one (except as between 70- and 80-year-olds). CBA, too, generally prefers to vaccinate the old. In short, the methodologies concur in preferring to vaccinate the old. They diverge on the income dimension. Utilitarianism and, even more so, CBA prefers to vaccinate richer individuals within each age tranche. Prioritarianism counteracts this troubling preference for the rich.

5. Conclusion This chapter has compared CBA, utilitarianism, and prioritarianism with respect to fatality risk reduction. Each approach uses some metric of social value and calculates the value of reducing an individual’s risk in terms of that metric. The CBA metric is the monetary equivalent; the utilitarian metric is expected lifetime well-being; the prioritarian metric is expected transformed lifetime well-being (using a concave transformation function, as in Figure 4.1). An individual’s age is one of the characteristics that might determine the social value of reducing their fatality risk, as per these approaches. In the simulation model described in section 3, individuals are differentiated by age as well as income. Implementing the three approaches in the context of this model, it emerges that none are neutral with respect to age. Rather, the utilitarian value of risk reduction decreases as individuals become older and prioritarian value even more so. CBA value first rises then falls with age.22 A key intuitive advantage of prioritarianism is how it treats income differences in valuing risk reduction. Utilitarianism is skewed towards the rich, and CBA is skewed even more so. Prioritarianism can neutralize this preference (with a sufficiently concave transformation function), while preserving the (intuitively plausible) preference for the young. 22 See Tables 4.2–4.4.

Ageing without Ageism?

69

References Adler, Matthew D. 2012. Well-Being and Fair Distribution: Beyond Cost–Benefit Analysis. New York: Oxford University Press. Adler, Matthew D. 2016. ‘Extended Preferences’. In The Oxford Handbook of Well-Being and Public Policy, ed. Matthew D. Adler and Marc Fleurbaey. New York: Oxford University Press, pp. 476–517. Adler, Matthew D. 2019. Measuring Social Welfare: An Introduction. New York: Oxford University Press. Adler, Matthew D. 2020. ‘What Should We Spend to Save Lives in a Pandemic? A Critique of the Value of Statistical Life’. Covid Economics 33: 1–45, 30 June. https://cepr.org/ content/covid-economics. Adler, Matthew D., and Koen Decancq. 2022. ‘Well-Being Measurement’. In Prioritarianism in Practice, ed. Matthew D. Adler and Ole F. Norheim. Cambridge: Cambridge University Press, pp. 128–171. Adler, Matthew D., and Nils Holtug. 2019. ‘Prioritarianism: A Response to Critics’. Politics, Philosophy & Economics 18: 101–144. Adler, Matthew D., and Ole Norheim, eds. 2022. Prioritarianism in Practice. Cambridge: Cambridge University Press. Adler, Matthew D., and Eric A. Posner. 2006. New Foundations of Cost–Benefit Analysis. Cambridge, MA: Harvard University Press. Adler, Matthew D., James K. Hammitt, and Nicolas Treich. 2014. ‘The Social Value of Fatality Risk Reduction: VSL versus the Social Welfare Function Approach’. Journal of Health Economics 35: 82–93. Adler, Matthew D., Maddalena Ferranna, James K. Hammitt, and Nicolas Treich. 2021. ‘Fair Innings? The Utilitarian and Prioritarian Value of Risk Reduction over a Whole Lifetime’. Journal of Health Economics 75: 102412. Boardman, Anthony E., David H. Greenberg, Aidan R. Vining, and David L. Weimer. 2018. Cost–Benefit Analysis: Concepts and Practice (5th edn). Cambridge: Cambridge University Press. Dolan, Paul, Rebecca Shaw, et al., 2005. ‘QALY Maximisation and People’s Preferences: A Methodological Review of the Literature’. Health Economics 14: 197–208. Eggleston, Ben, and Dale E. Miller, eds. 2014. The Cambridge Companion to Utilitarianism. Cambridge: Cambridge University Press. Emanuel, Ezekiel J., Govind Persad, et al. 2020. ‘Fair Allocation of Scarce Medical Resources in the Time of COVID-19’. New England Journal of Medicine 382: 2049– 2055. Ferguson, Neil M., et al. 2020. ‘Impact of Non-Pharmaceutical Interventions (NPIs) to Reduce COVID-19 Mortality and Healthcare Demand’. Imperial College, London, working paper. https://doi.org/10.25561/77482.

70

Age and the Social Value of Risk Reduction: Three Perspectives

Ferranna, Maddalena, J.P. Sevilla, and David E. Bloom. 2022. ‘The COVID-19 Pandemic’. In Prioritarianism in Practice, ed. Matthew D. Adler and Ole F. Norheim. Cambridge: Cambridge University Press, pp. 572–650. Freeman III, A. Myrick, Joseph A. Herriges, and Catherine L. Kling. 2014. The Measurement of Environmental and Resource Values: Theory and Methods (3rd edn). Abingdon: Taylor & Francis Group. Hammitt, James K., and Nicolas Treich. 2022. ‘Prioritarianism and Fatality Risk Regulation’. In Prioritarianism in Practice, ed.Matthew D. Adler and Ole F. Norheim. Cambridge: Cambridge University Press, pp. 317–359. Holtug, Nils. 2017. ‘Prioritarianism’. Oxford Research Encyclopedia of Politics, http://dx. doi.org/10.1093/acrefore/9780190228637.013.232. Levin, Andrew T., William P. Hanage, et al. 2020. ‘Assessing the Age Specificity of Infection Fatality Rates for COVID-19: Systematic Review, Meta-Analysis, and Public Policy Implications’. European Journal of Epidemiology 35: 1123–1138. Parfit, Derek. 2000. ‘Equality or Priority?’ In The Ideal of Equality, ed. Matthew Clayton and Andrew Williams. Houndmills: Palgrave, pp. 81–125. [Delivered as the Lindley Lecture at the University of Kansas in 1991].

5 Can Egalitarians Justify Spending More on the Elderly? Paul Bou-Habib

The modern welfare state devotes a significant proportion of its budget to the needs of the elderly. A graph depicting the proportions of welfare state expenditure across the periods of people’s lives would be U-shaped, with the greatest proportions devoted to the first and last periods. Or, as I will also refer to this phenomenon, the welfare state devotes a ‘disproportionate’ amount of expenditure on the elderly (as well as on the very young).1 Now this fact may not seem puzzling at first sight. Many amongst the elderly have significant needs: many elderly people cannot work and so lack income, or they need health care or assistance in carrying out day-to-day tasks. When the welfare state provides the elderly with pensions, health care, and daily nursing care, this makes sense, so one might think, because the welfare state should provide assistance for people with significant needs. However, one concern with disproportionate expenditure on the elderly that has been extensively debated amongst moral and political philosophers over the past 50 years is that it may amount to an inefficient allocation of resources over the lifespan. By this I mean that disproportionate expenditure on the elderly may fail to maximize the amount of well-being that many people can experience in their lives. Consider, for example, the health conditions of the elderly for which only ‘high-cost, low-value’ care is available, such as, arguably, the condition suffered by elderly people with advanced Alzheimer’s disease. The care available for this condition is expensive. However, because of the nature of the disease, it may not make a great difference to the well-being of the elderly person cared for and may not even be sufficient to preserve a decent standard of life. If some of the expenditure for this care were shifted to an earlier period of this same person’s life, it would arguably increase the amount of well-being in their life.2 Now, if one adopts a benefit-maximizing principle for determining the kinds of health care that should be publicly funded, one is likely to reject the funding of high-cost, low-value care. One example of a benefit-maximizing principle is the familiar utilitarian principle that requires that funding for care must maximize aggregate utility in society. Less obvious, however, is the fact that even egalitarians, who 1 According to Gál et. al. (2018), private expenditure on younger age groups outstrips public expenditure on the elderly, and the elderly may not therefore be the greatest recipients of social expenditure. 2 This example is drawn from McKerlie (2013: 47). I return to McKerlie’s discussion of it in section 1. Paul Bou-Habib, Can Egalitarians Justify Spending More on the Elderly?. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Paul Bou-Habib (2023). DOI: 10.1093/oso/9780192894090.003.0006

72

Can Egalitarians Justify Spending More on the Elderly?

do not endorse a benefit-maximizing principle for allocating resources, may struggle to justify the funding of high-cost, low-value care.3 While egalitarians reject reliance on benefit-maximizing principles when determining how resources should be distributed between people, they tend to allow reliance on a benefit-maximizing principle when determining how resources must be distributed between stages of people’s lives. (More generally, for ‘leximin’ egalitarians, efficiency gains are permitted whenever they are likely to benefit the least well off.) They can thus allow that conditions like Alzheimer’s disease not be treated if a greater amount of lifetime well-being can be achieved for people when the resources saved are used at other stages of their lives. Indeed, if egalitarianism requires that people enjoy equal lifetime well-being at the highest possible level of equality, they may even require this result.⁴ This chapter examines how egalitarians can defend a position in favour of disproportionate expenditure on the elderly, including expenditure on treatment for Alzheimer’s disease. To succeed, such a defence must justify constraints on how resources may be distributed within people’s lives, and, in particular, justify constraints which ensure that people avoid hardship at all stages of life, including old age. Section 1 raises objections against three ways egalitarians have sought to identify such constraints, which appeal to prudence, time-specific principles, and relational egalitarianism, respectively. Section 2 then diagnoses the challenge egalitarians face in identifying those constraints and a way forward. I suggest that egalitarians struggle to justify constraints on how resources must be distributed within lives because of the possibility of compensation within life—that is, the possibility that the hardship a person experiences during one period of life can be compensated for by benefits they enjoy during other periods. To justify constraints that prevent hardship during all stages of life, including the hardship of Alzheimer’s disease during old age, egalitarians must therefore identify reasons for preventing hardship that cannot be overridden by the fact that the people experiencing hardship have enjoyed compensating benefits during other periods of life. Section 2 sets out three such reasons: the anonymous badness of suffering, the value of dignity, and the symbolic value inherent in our treatment of others. These reasons, when added to relational egalitarianism, provide a more complete statement of constraints on distribution within lives and justify disproportionate expenditure on the elderly.

3 By ‘egalitarians’, I have in mind people who oppose unequal distributions between people either because they believe equal distribution is intrinsically valuable (intrinsic egalitarians) or because they believe it is morally more important to benefit people the worse off they are (prioritarians). The classic exposition of prioritarianism is Parfit (2000). While prioritarianism can be construed as a benefitmaximizing view, note that this is not the only the way to construe it. For a discussion of a ‘contractualist’ interpretation of prioritarianism, according to which it is morally more important to benefit people the worse off they are because they have stronger claims to benefits than other people, see Williams (2012). ⁴ Some egalitarians do not focus on well-being as the currency of just distribution. For example, Ronald Dworkin argues that a just distribution of health care should be determined on the basis of how people would insure against poor health under certain hypothetical conditions (2000: 307–319). But a version of the same problem arises for Dworkin. Given the high-cost, low-value nature of care for Alzheimer’s, many may decide not to insure for such care, which would imply, counterintuitively, that it is not required as a matter of justice. For this problem in Dworkin’s approach, see Bou-Habib (2011).

Ageing without Ageism?

73

1. Three stages of egalitarian theorizing about age-group justice My discussion begins with Norman Daniels’s pioneering work on age-group justice in the 1980s. The immediate context for this work was the concern that the welfare state might be unfairly privileging the interests of elderly people at the expense of younger people.⁵ What was needed in order to resolve this controversy was some way of figuring out what egalitarian justice demands by way of expenditure on the needs of different age groups. Daniels suggested an approach to this question that drew on the following important insight: ‘[j]ustice between age groups [. . .] is a problem best solved if we stop thinking of the old and the young as distinct groups. We age. The young become old’ (1988: 18). We should instead adopt what Daniels called a ‘prudential lifespan account’ for working out what people in different age groups are owed. This account involves a thought experiment in which we imagine how prudent people would allocate a fair individual share of resources across all the stages of a complete life, assuming they do not know their age or their specific conceptions of the good life but instead construe their well-being in terms of their access to certain allpurpose goods. (These restrictions ensure that the thought experiment is not biased in favour of any specific age group or conception of the good life.) Each stage of life within this prudent lifespan should then be regarded ‘as a proxy for an age group’ (1988: 18). Different age groups should receive the resources that would be allocated to the corresponding stage of life in the prudent lifespan. A key assumption of the prudential lifespan account is that justice is fundamentally concerned with how people’s complete lives compare to each other rather than with how periods of people’s lives compare to each other. It is this assumption that explains why we need not worry about inequalities between the young and old as such, given that these are, after all, only inequalities between periods of lives. What reason justifies this assumption that it is distribution between complete lives that should be the focus of moral attention for egalitarians? According to Nagel (1979: 120) and McKerlie (2001: 153–154), it is the fact that benefits and harms at different periods of the same person’s life can compensate for each other. As McKerlie explains, this ‘means that a larger benefit at one time outweighs a smaller harm at some other time, so that it is reasonable for a person to choose to have both the harm and the benefit rather than neither’ (2001: 154). If the harm that a person experiences during a specific period can be compensated for by benefits during other periods of their life, then our concern for them should not be focused only on how they fare during that specific period but should also take into account how they fare at other periods. Furthermore, assuming the periods over which benefits and harms can compensate for each other in a life stretch across a complete life, our concern for them should focus on their complete life rather than on specific periods within it.⁶ ⁵ Apart from his book-length exposition of the prudential lifespan approach (1988), Daniels also discusses the prudential lifespan account in a few earlier and later publications (1983, 1985: Chapter 5, 1996). ⁶ One might object that the possibility of compensation between different periods of the same life is not a justification of the complete lives view rather than a consequence of that view. But note that the claim

74

Can Egalitarians Justify Spending More on the Elderly?

While Daniel’s prudential lifespan account was a significant step forward in theorizing about age-group justice, it faced the serious problem that some health conditions in old age, such as Alzheimer’s disease, prevent significant improvements in well-being even with expensive care. The implication of this is that a prudent allocation of one’s fair lifetime share should lead one to refrain from devoting resources to such care, given that the resources can better be used during earlier stages of life. The prudential lifespan account could not then provide a basis for some instances of disproportionate, but seemingly common-sensical, expenditure on the elderly. It is partly this problem that leads to the second stage of egalitarian theorizing about age-group justice initiated by Dennis McKerlie.⁷ McKerlie proposes that we adopt a different constraint on the distribution of resources within lives than the one adopted by the prudential lifespan account. As we have just seen, the prudential lifespan account constrains the distribution of resources within lives in line with what prudence demands under impartial conditions. The constraint suggested by McKerlie is entirely different. He suggests that we endorse a ‘time-specific’ principle of distribution. A time-specific principle of distribution maintains that justice requires that certain distributions obtain between people-during-specific periods. There are many different ways in which one might characterize such a principle.⁸ McKerlie calls the time-specific principle he endorses ‘time-specific priority’. Imagine that we divide the lives of two people, A and B, into 10 same-length periods so that there are 20 periods in all. The time-specific priority principle tells us that there is more moral value in assisting A or B during any one of those 20 periods, the worse off A or B is during that period. McKerlie’s view is that we should constrain egalitarianism between complete lives with this time-specific priority principle.⁹ We should give moral importance to assisting people who will have the worst-off complete lives and to assisting people who have the worst-off periods of lives. When these criteria conflict in their implications, we should try, somehow, to strike a balance between them. (McKerlie does not provide a rule for how to strike that balance.) McKerlie’s time-specific principle may be able to justify disproportionate expenditure on the elderly despite the concern about inefficiency this causes across the lifespan. Elderly people with advanced Alzheimer’s disease are, after all, very badly off in absolute terms during the periods that they suffer from that disease, and timespecific priority can thus require that they receive assistance, even if this way of using their fair lifetime share of resources fails to maximize their lifetime well-being. (This assumes that we should give great weight to time-specific priority, so that the

about compensation is an axiological claim about what is valuable for people (as McKerlie puts it, it is a claim about what it is reasonable for a person to choose assuming they are prudent). The complete life view, by contrast, is a moral claim about how resources ought to be distributed between people. Arguably, a claim about prudential value is not a consequence of that moral claim but is rather a reason for it. ⁷ McKerlie developed his theory of age group across many publications. For a culminating statement, see (2013). ⁸ For a good discussion of the various possibilities, see McKerlie (1989). ⁹ The form of egalitarianism that McKerlie supports at the level of complete lives is also prioritarian. (As I suggest in n. 2 above, I regard prioritarianism as a form of egalitarianism.) He thus believes we should combine complete life prioritarianism with time-specific prioritarianism.

Ageing without Ageism?

75

moral importance of alleviating people during a very bad period outweighs the moral importance of maximizing their life-time well-being.) However, there are two main problems with McKerlie’s view. The first is that it lacks a rationale for time-specific priority other than that it fills an intuitive gap in our theorizing. This makes it appear ad hoc. Second, time-specific priority may, in any case, not have the form that an intuitively plausible time-specific principle should have. Consider the different time-specific principle defended by Axel Gosseries (2011), namely, time-specific sufficiency. Time-specific sufficiency tells us to give very great weight to ensuring that each person lies above a minimal threshold of well-being at each period of their life and no weight to ensuring that they are able to experience well-being beyond that threshold at each period of life. Intuitively, time-specific sufficiency better captures the kind of concern we should have for people during specific periods than does time-specific priority. Suppose a person is experiencing great hardship during a given period of their life but is not disadvantaged compared to others in terms of how they fare over their complete life. In that case, while it seems strongly intuitive that we should help raise them to the sufficiency threshold during their period of hardship, it is not intuitive that they should have claims (albeit of decreasing strength) to still further increases in well-being during that period, given that they are not disadvantaged compared to others over their complete life. The latter is what time-specific priority requires. That our concern for how people fare at specific periods is time-specific is also something we seem to express when we say that people must always live ‘in a decent condition’ or ‘with dignity’ (cf. Bou-Habib 2011). These seem to be threshold-focused convictions. We come, finally, to a third stage of egalitarian theorizing about age-group justice, inaugurated by Bidadanure (2016), who introduces a third type of constraint on how resources may be distributed within lives. Bidadanure argues that we should constrain egalitarianism over complete lives with a principle of relational equality. Very roughly, this is the principle that people should not be treated by others in ways that imply that they have less than equal status. Examples of such troubling treatment include insulting people, marginalizing them (i.e. excluding them from participating in important social events), or dominating them (i.e. leaving them with no option but to act as one prefers they should).1⁰ The principle of relational equality is relevant to the question of whether disproportionate expenditure on the elderly is justified because the elderly are vulnerable to the kinds of mistreatment the principle condemns and because disproportionate expenditure may be necessary in order to protect them against this vulnerability. To see this possibility, consider Bidadanure’s discussion of an example once proposed by McKerlie: [I]magine that the same city block contains a condominium complex and a retirement home. The residents of the complex are middle-aged, affluent, and happy. The retirement home is old and overcrowded. Its residents have adequate medical care but little dignity or happiness. Our first reaction is that this inequality raises an 1⁰ For a thorough examination of relational egalitarianism, see Lippert-Rasmussen (2018).

76

Can Egalitarians Justify Spending More on the Elderly?

issue of justice [. . .] But the young do grow old, and the very old were once young. So let us also suppose that the people in the condominiums will end up in such a place as the nursing home, and that the people in the home used to be as fortunate as their neighbors now are. (McKerlie 2001: 152–153)

Many readers will find this example intuitively troubling and would endorse policy reform that prevents the situation it depicts from arising, even if the elderly in the example have as good lives, on the whole, as the young. Whereas McKerlie sought to justify this intuition with the idea of time-specific priority, Bidadanure seeks to justify it with that principle of relational equality (i.e. the principle that people should not be treated by others in ways that imply they have less than equal status). As she explains, the elderly in McKerlie’s example are ‘spatially segregated from the affluent youth’ and ‘[s]patial segregation and differential levels of affluence easily become associated with unequal status and unequal levels of respect’ (Bidadanure 2016: 246). The problem with the appeal to relational egalitarianism is that it makes support for the elderly depend on the fact that their difficult circumstances may, or will, have certain knock-on effects on their marginalization and status, when intuitively, they should not. Bidadanure is right that spatial segregation and lower levels of affluence can lead the elderly to suffer from unequal status, and this is indeed a powerful reason to protect them from such circumstances. But if the elderly found themselves in difficult circumstances that did not have those knock-on effects, we would still, presumably, find this troubling. Suppose a group of elderly people suffer from a painful illness but are not segregated from others or poor. We would not, in that case, simply dismiss their claims to support. Their painful illness is still a reason for us to offer them generous support even if their complete lives are no worse than those of others in society. The appeal to a principle of relational equality is, in other words, incomplete. While relational equality surely matters, it does not vindicate all of the disproportionate expenditure on the elderly that common sense seems to require.

2. Suffering, dignity, and symbolic value I have been examining whether egalitarians can justify disproportionate expenditure on the elderly by considering three kinds of constraints egalitarians have proposed on how resources may be distributed. I have argued that constraining intra-life distribution with prudence under conditions of impartiality fails to adequately protect people against hardship, constraining it with a principle of time-specific priority is ad hoc, and constraining it with a principle of relational equality yields an intuitively incomplete set of protections against hardship in old age. I have also suggested that a different time-specific principle defended by Gosseries (2011)—namely, timespecific sufficiency—is more promising, although it, too, faces the problem that it is ad hoc. We need a theory or set of reasons that explains why a principle such as

Ageing without Ageism?

77

time-specific sufficiency, or something similar to it, is justified as a constraint on the distribution of resources within lives. We can see a way forward, I now propose, if we notice a seeming obstacle to our imposing constraints on how resources are divided within lives. This is the idea that benefits and harms at different periods of life can compensate for each other. The possibility of compensation presents a challenge to constraints on how resources should be distributed within lives because if a person in hardship during one period of life can be compensated for this with sufficient resources or well-being during other periods, the idea that we must constrain intra-life distribution to prevent that hardship seems unwarranted. To defend constraints on resource distributions within lives, we must find reasons for showing concern for people in hardship that remain valid even if the people in question have enjoyed, or will enjoy, a great deal of resources of well-being at other times of their lives. The appeal to relational egalitarianism is one example of this strategy: it tells us that we should worry about people in hardship not because they experience a severe deficiency in resources or well-being per se but because they are vulnerable to being marginalized or disrespected. Notice that these latter features of their hardship cannot be compensated for by resources or well-being they have enjoyed, or will enjoy, at other times of their lives. As we noted earlier, however, the appeal to relational egalitarianism is incomplete. To provide a more complete justification of disproportionate expenditure on the elderly, egalitarians must therefore try to find further reasons for time-specific concern that are immune to the possibility of compensation within life. Let me now briefly suggest three such reasons. (a) The anonymous badness of suffering. Consider a situation in which a person suffers terribly—for example, suppose I am in excruciating pain because I have a cracked tooth. Now, one fact that could justify alleviating my suffering is, of course, that doing this will improve my level of well-being. But some writers have suggested that there is a distinct fact that would justify alleviating my suffering. This is the fact that someone is in pain, quite apart from the fact that that someone happens to be me (Nagel 1989; Mayerfeld 1999).11 Thomas Nagel, who endorses the ‘anonymous’ badness of suffering, finds its badness self-evident, and, thus, for that reason, difficult to justify, other than by illustrating its force with examples (1989: 160). One example might be this: when we witness a person who is suffering excruciating physical pain, most of us do not seem to need to know who they are, let alone how many resources or how much well-being they may have happened to enjoy during other periods of their life, in order to believe that we are obliged to assist them. Furthermore, the belief that we are obliged to assist them is not one that we come to regard, on reflection, as some sort of misplaced empathy for them. Even after reflecting on the situation, we continue to believe that we were obliged to assist them. The anonymous badness of suffering justifies constraints on the distribution of resources within lives. Suppose, for example, that some people would prefer exposing 11 As Thomas Nagel puts it: ‘if I lacked or lost the conception of myself as distinct from other possible or actual persons, I could still apprehend the badness of pain, immediately’ (1989: 161).

78

Can Egalitarians Justify Spending More on the Elderly?

themselves to relatively short periods of suffering in old age by diverting the resources from their fair shares needed to alleviate that suffering to earlier stages of their lives. A constraint that prohibited them from doing this would be inefficient in the sense that it would prevent them from satisfying their preferences as much as possible. However, if suffering is anonymously bad, such a constraint is justified because, unless these people complied with this constraint, others would have to pay to protect them from their suffering in old age, which would be unfair. The anonymous badness of suffering does not, of course, single out old age for special consideration but constrains intralife distribution so that it protects against suffering whenever it might occur in life. But there is good reason to think that it will justify disproportionate expenditure on old age if people are more prone to experience suffering in old age. One concern one might have about appealing to the anonymous badness of suffering as a basis for disproportionate expenditure on the elderly is that it is difficult to specify the meaning of ‘suffering’ precisely. As is suggested by examples such as excruciating physical pain and mental anguish, suffering seems to be an aspect of our experience: to suffer is, necessarily, to experience something.12 Furthermore, to suffer is to experience something awful—the pain of a cracked tooth, rather than mild backache, or of intense anguish rather than boredom. But can we say something more specific about the kinds of awful experience that constitute suffering? I shall not try to say something conclusive here but will make one suggestion. When people suffer, they have awful experiences that tend to absorb them fully in the sense that they can think only of how to avoid those experiences. Suffering overwhelms people. This may help to explain why suffering can arouse concern independently of the fact that it happens to be the suffering of this or that particular person: the fact that the capacities of a person are absorbed by something awful is disturbing in its own right, apart from the fact that the person in question is experiencing a very low level of well-being. (Of course, the latter fact unavoidably coincides with the former, but the former fact may nevertheless be an independent reason for concern.)13 (b) The dignity of persons. A second reason for assisting people during specific periods of their lives that remains valid regardless of how many resources or how much well-being they have enjoyed during other periods of their life appeals to the dignity of persons. The term ‘dignity’ has been interpreted in many ways, indeed, so many that some have come to regard it as a meaningless term (Macklin 2003). I construe it here in line with one of the main traditions of interpreting it, namely, as closely tied to a person’s capacities for autonomy (see, e.g. Beyleveld and Brownsword 2001).

12 In construing suffering as necessarily tied to experience (so that a person cannot suffer from a condition that they do not experience), I follow several writers, including Mayerfeld (1999: 11); Brady (2018: 11); and Kauppinen (2019: 19–20). 13 One puzzle that is raised by the suggestion that suffering consists of being absorbed by awful experience concerns people who ‘suffer with dignity’, that is, who manage to conduct themselves calmly during their awful experiences. My suggestion may seem to imply that these people do not, in fact, suffer because they seem to manage not to be fully absorbed by their awful experiences and, worse, that their condition is not therefore anonymously bad in ways that justify assistance. One response to this puzzle is this. That they manage to comport themselves with calmness does not imply that they are not absorbed by awful experience but only that they are not reacting to their awful experience in certain ways.

Ageing without Ageism?

79

More specifically, the dignity of persons is a value that inheres in their capacities for autonomy. Respect for this value requires that they be able to adequately exercise those capacities at all times. As with suffering, the concern with dignity requires us to focus on the state of a person’s capacities rather than on their well-being, strictly speaking. While a person’s well-being is affected by their being able to adequately exercise their capacities for autonomy, the fact that they are able to adequately exercise those capacities can matter independently of their well-being. The preservation of dignity can thus generate a duty for others to assist people during specific periods of their lives regardless of the amount of well-being they may have enjoyed during other periods in their lives. If the predicament in which a person lacks opportunities with which to exercise their capacities for autonomy is more prevalent amongst the elderly, the consideration that we must protect the dignity of persons can thus justify disproportionate expenditure on the elderly.1⁴ (c) Symbolic value. A third reason is the symbolic value of acting in ways that express one’s appreciation for others or of avoiding action that expresses indifference for others.1⁵ This consideration overlaps with relational egalitarianism, but it is not co-extensive with it. What ultimately moves relational egalitarianism is a deeply important concern that policies and practices should avoid expressing the attitude that some are inferior to others in a ranking of status. But that our policies and practices should avoid implying that some are inferior to others is not the only feature of those policies that has symbolic significance. To see this, consider the kind of care that should be provided for an elderly person who suffers from advanced Alzheimer’s disease. Presumably, the expenses incurred when their hair is cut, when their clothes are washed, and when the windows and carpets in their bedroom are kept clean are justified. But it is not obvious that this is because withholding such care sends a message that they have lower status than others. If a society were to withhold such care from everyone who suffers from advanced Alzheimer’s disease, regardless of a person’s sex, race, or sexual orientation, it would not necessarily be treating anyone as if they had less status than others. The message that is sent by providing proper care for an Alzheimer’s patient is not exactly a message about how their worth compares to others—namely, that it is of equal worth to others—but is rather a message about their non-comparative worth. Leaving them looking dishevelled, in dirty clothes and in a dirty room, expresses an attitude of indifference towards them when they are still valuable and this is something that a decent society should not do even if it did this to all people who suffer from advanced Alzheimer’s disease regardless of their sex, race, or sexual orientation. Once again, if the elderly, more than other age groups, find themselves in circumstances in which our failure to assist would amount to an expression of indifference towards their value, then disproportionate expenditure on the elderly might be justified on grounds of symbolic value.

1⁴ For a more detailed discussion of the relevance of dignity for egalitarian theorizing about age group justice, see Bou-Habib (2011). 1⁵ For an excellent defence of the idea of symbolic value, see Adams (2002: Chapter 9).

80

Can Egalitarians Justify Spending More on the Elderly?

3. Conclusion To justify disproportionate expenditure on the elderly, egalitarians must justify constraints on how resources may be distributed within lives. After reviewing a number of proposed justifications, I have suggested that the best way to defend such constraints is to identify reasons for alleviating hardship during specific periods of life that remain valid even if the people in question enjoy compensating amounts of resources or well-being at other periods in their lives. Along with relational equality, we should regard the anonymous badness of suffering, the value of dignity, and the symbolism inherent in how we treat others as reasons of that kind. Let me conclude by describing a picture of age-group justice for egalitarians that is consistent with this conclusion. The welfare state should follow the recommendations of the prudential lifespan account when designing its various policies—that is, it should design those policies in the way a prudent person would design them when devoting a fair share of resources across the different stages of a complete life. However, the welfare state must constrain the results of the prudential lifespan account by ensuring that people in each period of life are respected as having equal status to others, prevented from suffering, protected in a minimal range of opportunities for exercising their capacities for autonomy, and symbolically affirmed in their individual worth. These constraints have priority over the prudential lifespan account in the sense that the welfare state may distribute resources over the life course in ways that are prudent only after it has ensured that these constraints are respected. My conjecture is that, in this picture of age-group justice, we will see a greater amount of welfare state expenditure devoted to the elderly than to most other age groups in society.

Acknowledgements I am grateful to Greg Bognar, Axel Gosseries, and Serena Olsaretti for extensive and helpful comments on previous drafts of this chapter.

References Adams, Robert. 2002. Finite and Infinite Goods. Oxford: Oxford University Press. Beyleveld, Deryck and Roger Brownsword. 2001. Human Dignity in Bioethics and Biolaw. Oxford: Oxford University Press. Bidadanure, Juliana. 2016. ‘Making Sense of Age-Group Justice: A Time for Relational Equality?’ Politics, Philosophy & Economics 15: 234. Bou-Habib, Paul. 2011. ‘Distributive Justice, Dignity, and the Lifetime View’. Social Theory and Practice 37: 285. Brady, Michael. 2018. Suffering and Virtue. Oxford: Oxford University Press.

Ageing without Ageism?

81

Daniels, Norman. 1983. ‘Justice between Age Groups: Am I My Parents’ Keeper?’ Milbank Memorial Fund Quarterly Health and Society 61: 489. Daniels, Norman. 1985. Just Health Care. Cambridge: Cambridge University Press. Daniels, Norman. 1988. Am I My Parents’ Keeper? Oxford: Oxford University Press. Daniels, Norman. 1996. Justice and Justification: Reflective Equilibrium in Theory and Practice. Cambridge: Cambridge University Press. Dworkin, Ronald. 2000. Sovereign Virtue. Cambridge, MA: Harvard University Press. Gál, Róbert, Pieter Vanhuysse and Lilli Vargha. et. al. 2018. ‘Pro-Elderly Welfare States within Child-Oriented Societies’. Journal of European Public Policy 25: 944–958. Gosseries, Axel. 2011. ‘Qu’est-ce que le suffisantisme?’ Philosophiques 38: 465. Hart, H. L. A. 1955. ‘Are There Any Natural Rights?’ Philosophical Review, 64: 175–191. Kauppinen, Antti. 2019. ‘The World According to Suffering’. In Philosophy of Suffering:Metaphysics, Value, and Normativity, ed. David Bain, Michael Brady, and Jennifer Corns. London: Routledge, 19–36. Lippert-Rasmussen, Kasper. 2018. Relational Egalitarianism. Cambridge: Cambridge University Press. Macklin, Ruth. 2003. ‘Dignity is a Useless Concept’. British Medical Journal 27: 1419–1420. Mayerfeld, Jamie. 1999. Suffering and Moral Responsibility. Oxford: Oxford University Press. McKerlie, Dennis. 1989. ‘Equality and Time’. Ethics 99: 475. McKerlie, Dennis. 2001. ‘Justice between the Young and the Old’. Philosophy & Public Affairs 30: 152. McKerlie, Dennis. 2013. Justice between the Young and the Old. Oxford, Oxford University press. Nagel, Thomas. 1979. Mortal Questions. Cambridge: Cambridge University Press. Nagel, Thomas. 1989. The View from Nowhere. Oxford: Oxford University Press. Parfit, Derek. 2000. ‘Equality or Priority’. In The Ideal of Equality, ed. Matthew Clayton and Andrew Williams. Palgrave: London, 81–125. Williams, Andrew. 2012. ‘The Priority View Bites the Dust?’ Utilitas 24: 315–331.

6 Age Limits and the Significance of Entire Lives Egalitarianism Axel Gosseries

In this chapter, I explore the relationship between the fairness of upper or lower age limits and the idea of equality over entire lives. We extensively rely on age criteria to treat people differently. We do so in health care, education, employment, immigration law, criminal policy, or to mark out political rights. This raises the question whether (and why) doing so is less problematic than differential treatment based on, for example, race, gender, or disability. In order to find out, we can try to identify what renders age ‘special’ compared with such other grounds of differential treatment and whether anything follows normatively. In searching for such ‘specialness-makers’, we often begin with phrases like ‘Our age changes’ or ‘We all age.’ Both phrases—through the use of ‘our’ or ‘we’—suggest some personal continuity across the ages. Yet, the former phrase—using ‘age’ as a noun—is better at conveying the idea of transient, successive, and multiple age-group memberships along a single life, as well as the fact that our age is outside of our control (we don’t say ‘we change our age’). In contrast, the latter phrase—using ‘age’ as a verb—is better at stressing the universality of this change (we don’t say ‘some of us age’) or the partly endogenous nature of individual ageing (we say neither that ‘we are being aged’ nor that ‘it ages’, as in ‘it rains’). In addition, ‘age’ as a verb helps us to see the link between our age and the passage of time. The latter is associated with a variety of processes such as undergoing brain development or losing muscular strength, gaining or losing confidence, acquiring affective or intellectual skills, or losing a sense of innocence. This link between time and processes is key in understanding and assessing our use of age limits. What I want to explore in more detail here is the interplay between the changing nature of our age, the use of age limits, and a certain understanding of the egalitarian ideal, namely, that which is concerned about equality over entire lives. I will first briefly present this understanding of the egalitarian ideal of justice. I will then explore, in three steps, the nature of the relationship between age criteria and this understanding of the egalitarian ideal. My goal is to assess the extent to which this egalitarian ideal can account for the normative specialness of age (if any).

Axel Gosseries, Age Limits and the Significance of Entire Lives Egalitarianism. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Axel Gosseries (2023). DOI: 10.1093/oso/9780192894090.003.0007

Ageing without Ageism?

83

1. Entire-life egalitarianism Developing an egalitarian account of justice involves several steps. One step is to decide on a principle or pattern. For instance, if one aims at an egalitarian principle, one needs to find out whether one aims at equality as such or whether, for instance, one ought to accept, or even promote, inequalities if they are necessary to maximize the situation of the least well off, for example, under leximin egalitarianism.1 Another step is to settle on a ‘currency’ or ‘distribuendum’, that is, define in what respect people should be equal as a matter of justice. This is the ‘equality of what?’ issue. Philosophers have proposed a variety of accounts of such currencies, ranging from ‘resources’ to ‘capabilities’, ‘primary goods’, ‘opportunities for welfare’, ‘Real freedom’, and many others.2 Each proposed currency of justice answers in its own way a set of meta-questions. Among them is the issue of whether the currency should be welfarist or not and, if so, what theory of welfare it should endorse. Another such meta-question is whether the inequalities that one should care about should be limited to those that do not result from people’s problematic choices, whether they should be restricted to those for which individuals should not be held morally responsible. Egalitarians answering affirmatively—which is signalled by words such as ‘luck’ in ‘luck egalitarianism’ or ‘opportunity’ in ‘opportunity for welfare’—may worry about disadvantages resulting from, for example, genetic diseases or a person’s place of birth, while denying compensation to disadvantages resulting from, for example, certain forms of imprudent behaviour (e.g. helmetless driving or mind-changing tattooing). They defend a ‘responsibility-adjusted’ currency.3 Time is also crucial in defining the demands of an egalitarian theory of justice. In some cases, time touches on dimensions that do not reduce to currency issues. This is so, for example, when asking whether egalitarian duties apply towards people with whom we will never coexist; be they born in 421 or in 2221, one is answering the ‘Equality among whom?’ rather than the ‘Equality of what?’ question.⁴ In contrast, in other cases, the time dimension directly touches on the ‘What?’ question. This is so when egalitarians wonder whether they should assess ‘snapshot’ inequalities somehow in isolation (e.g. those of a calendar year) or whether they should go beyond.⁵ By ‘beyond’, I do not merely mean that we should link current distributions with their genesis in earlier periods or their likely consequences in the future. I rather mean that we should aggregate, in one way or another, what individuals enjoy today with what they have benefited from or enjoyed in the past as well as with what they are

1 See, e.g. Van Parijs (2001). 2 See, e.g. Cohen (1989). 3 See, e.g. Arneson (2004). Here, I use ‘responsibility-adjusted’ rather than the more common ‘responsibility-sensitive’. The latter phrase may suggest a focus on causal responsibility that is too exclusive. What is at stake here is rather what people should be held morally responsible for. ⁴ See, e.g. Gosseries (2019). ⁵ Here, I leave aside corresponding segments. See, on the latter, McKerlie (2012). The ‘snapshot’ label is from Broome (2006).

84

Age Limits and the Significance of Entire Lives Egalitarianism

likely to benefit from or enjoy in the future. And this, irrespective of whether or not such burdens and benefits are causally related to current ones. Deeming snapshot inequalities relevant independently of what happens at other periods renders egalitarianism responsive to random fluctuations in, for example, stock market value for an investor or rainfall for a farmer. Focusing on snapshot inequalities also renders our view more sensitive to snapshot inequalities that result from the combination of two facts: the coexistence of different age groups at any given moment and the social practice of differentiating entitlements and obligations on the basis of age. One alternative to an exclusive focus on snapshot inequalities is to move all the way to focusing on inequalities over people’s entire lives. Such an entire-life egalitarianism (hereinafter ‘EL egalitarianism’) defends the negative, anti-isolationist claim that, in assessing the fairness of a distribution, we should not assess the current situation of two individuals of different ages in isolation from what happened and will happen at other periods of their life. This allows us, for instance, to compare periods of people’s life during which they do not coexist, which is typically not the case for snapshot egalitarians.⁶ Yet, EL egalitarians are not alone in endorsing anti-isolationism. Consider the responsibility-adjusted perspective mentioned above. It has to assess whether current inequalities are unjust by looking at the respective choices made by the individuals at stake in the past. If some of the choices are such that society should not compensate for their negative consequences, current inequalities could be regarded as just from a responsibility-adjusted perspective. Yet, such a conclusion can be reached without looking at burdens and opportunities in the future and over entire lives. Hence, endorsing a responsibility-adjusted view does not necessarily commit us to an entire-life one. The reverse is true as well. An entire-life perspective can endorse a responsibility-adjusted view but also reject it altogether or attenuate its weight through allowing for regular fresh starts, avoiding holding us responsible for the consequences of choices made far in the past.⁷ Hence, EL egalitarianism involves refusing to assess snapshot inequalities (i.e. inequalities at a given moment or period) without looking at the situation of the individuals outside the period in question. In addition, it requires that we look at all the periods of their lives before being able to say anything about how well someone’s life went and about whether the current distribution of a good is unjust. A present inequality in one dimension may be counterbalanced by a reverse inequality in this or another dimension between the same individuals at another moment. The present inequality may then not necessarily be considered unjust as a result. Compensation across periods and ages is key for EL egalitarians. A disadvantage at period x or age y can be compensated by an advantage at period x + z or x − z or at age y + z or x − z.

⁶ With the exception of corresponding segments. ⁷ See Fleurbaey (2005).

Ageing without Ageism?

85

This is not the case for egalitarians rejecting certain types of snapshot inequalities irrespective of what happens at surrounding periods.⁸ Hence, while both snapshot and EL egalitarians have something to say about whether a temporary inequality in a given dimension is unjust, they differ in the range of facts deemed relevant in assessing it. Both of them may agree on the existence and magnitude of a temporary wealth or happiness inequality at time t. And yet, they might disagree about whether this inequality is unjust. In a nutshell, snapshot and EL egalitarians differ, axiologically, in how they bridge the finding of a momentary wealth or happiness level to assessing how well a person’s life is currently going and how well it will end up having gone overall. And they differ, normatively, in how to bridge the finding of a momentary wealth or happiness inequality to its assessment in terms of fairness. To be totally clear, EL egalitarians endorse the view that even if snapshot comparisons were feasible and meaningful (i.e. even if we could simply aggregate successive snapshot inequalities without going to a straight holistic comparison), snapshot equality under the relevant currency would still not necessarily be required from a normative point of view. Here, it is worth stressing, by way of illustration, how EL egalitarians handle differential longevity, that is, the fact that some people live longer lives than others. Let us assume that being short-lived is significantly worse than being long-lived, all else being equal. Snapshot egalitarianism seems unable to capture this intuition. In contrast, the EL approach is at home with considering longevity as an important dimension of how well our lives went. This is so not merely because it would aggregate or link some successive snapshot inequalities but more simply because it looks at our existence in full. Of course, EL can take on board the intuition that a short and intense life could be better than a long and dull one. It could also factor in the view that a longer life is better, even if summing up of snapshot happiness levels in the lives at stake were to lead to the same figure. For one could hold the view that a longer life is a better one even if it does not translate into higher levels of cumulated happiness. In other words, without pre-empting how longevity is supposed to be valued exactly, taking longevity into account is something that EL egalitarians do especially well because it requires that we look at people’s lives in full. Before I move ahead, it is also important to stress that endorsing EL egalitarianism does not prevent one from being a pluralist when it comes to the ‘entire-life’ issue. Radical EL egalitarians will claim that snapshot inequalities only interest us to the extent that they translate into inequalities over entire life. This is not practically insignificant since accumulating disadvantages at every single stage of our lives will lead us to a disadvantaged position over our entire lives. And yet, EL egalitarians don’t need to be radical. They may say, for instance, that while entire-life inequalities should be our priority concern, some weight should also be given, in addition, to some snapshot inequalities. And it is the latter moderate view that I will be endorsing here. To put it differently, I will investigate what role the entire-life intuition can play in accounting for the normative specialness of the age criterion, including when the ⁸ See, e.g. McKerlie (2012) and Bidadanure (2016).

86

Age Limits and the Significance of Entire Lives Egalitarianism

entire-life intuition is used in combination with snapshot concerns. This potentially leaves more room for the entire-life intuition than if I were to limit myself to radical EL egalitarianism.

2. Types of age limits The purpose of section 1 was to give a brief characterization of EL egalitarianism. Let me now turn to age limits. Consider the following type of absolute age limit: Those under (or above) age x are (not) entitled (or obliged) to y.

To illustrate this, I will take a few stylized examples inspired from actual practices across various jurisdictions. Children under 12 are denied the right to work. Teenagers under 16 are obliged to attend school. Citizens under 18 are denied the right to vote. Patients under 18 are offered a full refund for dental care. Teenagers under 18 are not allowed to marry. Judges above 70 are denied the right to stay in office. Drivers above 70 have to undergo regular medical check-ups. Workers under 65 are not entitled to retire with full pension benefits. Patients above 75 are denied access to intensive care units (ICUs) in case of a pandemic surge. These examples all involve plain absolute ages. They differ from transition ages, differentiated ages, relative ages, or age gaps. Transition age rules involve the use of absolute ages combined with a given calendar date. This is so when we exempt from the application of a new rule those who have already reached a given age on the day of the rule’s entry into force. Differentiated age rules involve the use of absolute ages differentiated along other dimensions such as gender. This is so when women are entitled to pension benefits at a different age than men. Relative age rules give priority to those who are older or younger in a given pool of people without setting any predefined age. For instance, in allocating organs for transplant, one may give priority, among those on the waiting list, to those who were born last (the youngest patients) rather than to those who signed on the list first. As to age gap rules, they adjust duties and rights to the magnitude of the gap between those interacting. This is the case with a maximum age difference between child and prospective parents for adoption eligibility or for optimal age gap for marriage in social acceptability rules. These five types of age limits can combine in different ways. For instance, criminalizing sexual intercourse may rely both on absolute ages and on age gap, as when sexual relationships with persons within the 12–16 years old range (absolute ages) are prohibited unless the age gap between the sexual partners is no larger than 4 years. Also, absolute age limits may be a second-best alternative to relative age limits whenever the full group of potential right holders cannot be identified on the day of decision, which is the case for patients admitted in a continuous manner to hospital and seeking access to ICU beds. In what follows, I mostly focus on plain absolute age limits, that is, absolute ages that are neither transition nor differentiated ages. A comprehensive analysis

Ageing without Ageism?

87

would need to be run on the other four types of uses of age limits. In addition, I also leave aside whether the age limit is merely used as a tie-breaker (when other criteria don’t suffice to select) or as an exclusive criterion instead. Finally, I am not assuming here that each and every one of such norms is ultimately defensible, all things considered. I merely aim to analyse how an EL egalitarian should approach absolute age limits and whether the ‘entire-life’ idea is central in accounting for the normative specialness of age.

3. Defensive entire-life argument I begin with the following EL egalitarian argument for the acceptability of age limits. I call it ‘defensive’ because rather than establishing a positive case for the use of age, it rejects, instead, a claim against the permissibility of age limits. The defensive argument: P1. (Age) criterion x is permissible if it does not increase inequalities over entire lives. P2. (Age) criterion x does not increase inequalities over entire lives. C. Therefore, (age) criterion x is permissible.

By way of illustration, consider someone objecting to disenfranchising the young on grounds of the gender analogy. A defensive EL egalitarian could respond that the difference with disenfranchising women rests on the fact that while gender-based disenfranchisement would definitely lead to inequalities over entire lives, it would not necessarily be the case with age-based disenfranchisement. Note, as well, the broader significance of the argument, beyond age-based practices, as long as there is a temporal dimension involved. For instance, in the case of storable votes regimes (i.e. an electoral system through which you can keep your votes for the next elections and exercise several votes on a single election as a result), people with the same age could have different number of votes on a given election day.⁹ This could partly be defended on grounds that, while allowing for inequalities in a given election, voters may end up with the same longevity-adjusted voting opportunities over their entire life. Now, let me begin by looking more closely at the first premise (P1). I take it that P1, in combination with P2, properly characterizes a widely shared intuition about what renders age limits acceptable. I am leaving aside, here, the possibility of alternative formulations of P1, for instance, a leximin formulation (to mention but one of the options pointed out above). What I want to stress here is the negative nature of P1. It does not provide us with any positive reason for using age criteria in the first place. It leaves things open in this respect. Actually, it does not even presuppose that such a positive reason is needed. And that potentially casts doubt on the claim that the ‘entire-life’ idea can account for the normative specialness of age. ⁹ See Casella (2011), Valente (2022).

88

Age Limits and the Significance of Entire Lives Egalitarianism

This opens the door to at least three types of responses. The first one—discussed in section 4—overcomes the merely negative nature of P1 by offering an argument in favour of age criteria from within the EL egalitarian perspective. The second one— discussed in section 5—points to a variety of positive arguments in support of age criteria, none of which relates positively or negatively to the ‘entire-life’ idea. The third one—which I don’t discuss in the present chapter—challenges the ‘entire-life’ perspective by claiming that age criteria might remain objectionable for egalitarians despite not worsening entire-life inequalities.1⁰ At this stage, I want to look at the second premise of the defensive argument (P2). Is it true that age limits do not worsen inequalities over entire lives? My claim here is that age criteria often do increase inequalities over entire lives. First, even if age limits were permanent and if we all had the same life expectancy, fluctuations in our socio-economic environments would generate inequalities. Hence, the impact of a given age limit would differ significantly over time. For instance, a minimum or maximum age for compulsory military service impacts people very differently, depending on whether one reaches these ages in peacetime or in wartime. Second, age limits change with time, with new ones being introduced (e.g. compulsory driving exams above a certain age), others being dropped (e.g. upper age limits for refunding contraceptive pills), lowered (e.g. shifting minimum voting age from 18 to 16), or raised (e.g. shifting minimum tobacco sales age from 18 to 21). In all such cases, we end up with differentiated ages across birth cohorts that will, everything else being equal, lead to differential impact over entire lives. Third, individuals have different life expectancies. Hence, setting an eligibility age for pension benefits at 67 impacts the short-lived differently than the long-lived. These are three illustrations of the fact that P2 will often be false. And this significantly restricts the domain of the entire-life view as a core specialness maker. Hence, considering the defensive argument ‘for’ age limits, we end up facing a twofold challenge to the centrality of EL egalitarianism as an account of the normative specialness of age limits. The first challenge relates to P1. It results from the conjunction of stressing the merely negative nature of P1 and of showing that a lot of the positive reasons for age limits can stand without the EL egalitarian idea, a point to which we return in section 5. The second challenge relates to P2. It stresses that there is a large subset of cases in which P2 is false. Together, these two challenges strongly question whether the EL egalitarian idea can play a central role in accounting for the normative specialness of age limits.

4. Affirmative entire-life argument So far, while I questioned its significance in accounting for the specialness of age, I have not challenged the EL egalitarian idea on its substance. I discussed one argument 1⁰ Objections to snapshot inequalities that do not lead to entire-life inequalities include human dignityrelated concerns (Gosseries 2003: 481; Bou-Habib: Chapter 5, this volume), concerns for non-domination and exploitation (Bidadanure 2016), and finally, concerns for the objectionability of discrimination despite the absence of any distributive inequality resulting from it (Lippert-Rasmussen 2018).

Ageing without Ageism?

89

that defensively endorses age limits while relying on the entire-life idea. In this section, I present and discuss a type of reason that both involves a positive case for age limits and relies on the EL egalitarian idea. Consider the following argument: The affirmative argument: P1. (Age) criterion x is desirable whenever it reduces inequalities over entire lives. P2. (Age) criterion x reduces inequalities over entire lives. C. Therefore, (age) criterion x is desirable.

I begin with a first example. Some jobs, such as judgeships or professorships, are in limited supply. Age-based compulsory retirement for such jobs at, for example, 65 is used in some countries. This can be for a number of reasons, ranging from diversifying a group’s generational composition (to ensure viewpoint diversity) to limiting power concentration or keeping the wage–productivity gap under control along people’s career.11 This list of purposes can also include the idea of job sharing across entire lives. By job sharing, I mean any device that caps the time during which one is allowed to work in order to better distribute access to jobs. Hence, agebased compulsory retirement can be seen through the same prism as weekly hours regulation. This job-sharing idea matters because it points to the potential of agebased mandatory retirement to reduce inequalities in access to certain jobs over entire lives, including across overlapping generations. Admittedly, this does not provide us with an all-things-considered case for age-based mandatory retirement.12 Efficiency losses may be such that while compulsory retirement reduces inequalities in access to employment, it may actually also leave the worse off in society in an even worse position. Yet, this affirmative EL egalitarian argument may play a significant role in justifying this type of age-based practice. As a second illustration, I return to the issue of differential longevity. Again, I assume that short-lived people are disadvantaged in comparison to long-lived people, everything else being equal. Consider two contexts. First, since we can only tell ex post who the short-lived individuals will turn out to be, it has been argued that we should shift consumption earlier in life to ensure that not just the long-lived among us enjoy it.13 This could be done through explicit age limits, for example, in the allocation of certain social benefits. Second, in the health-care context, prioritizing the young in access to ICUs can admittedly serve the purpose of maximizing the medical efficiency of an intervention. This entails a concern neither for inequalities in longevity nor for equality over entire lives in general beyond the strict dimension of longevity. Yet, the same priority to the young in access to ICUs could alternatively (or complementarily) be advocated with the purpose of reducing inequalities in longevity. By giving priority to a younger patient, including in cases in which she would turn out to have a worse prognosis than an older patient, we may aim to distribute longevities among people. And we may want to do so through relying on an age criterion, privileging the young. 11 On the latter, see Gosseries (2004). 12 See Lippert-Rasmussen (2019); Halliday and Parr (2022) 13 See Fleurbaey et al. (2014).

90

Age Limits and the Significance of Entire Lives Egalitarianism

What these social benefits and health-care contexts illustrate is that age limits can serve to reduce entire-life inequalities, either through taking the longevity differential as given and compensating in other ways or through acting on the longevity differential itself with the aim of reducing it. And what the compulsory retirement discussion above points to is the fact that the affirmative case for age limits is not limited to issues related to differential longevity. Hence, this confirms that there is a non-negligible range of cases in which the entire-life perspective can potentially support a positive case for age limits.1⁴

5. Positive reasons for age limits unrelated with the entire-life view So far, I have looked at age limits supported by a defensive EL egalitarian argument (e.g. minimum voting age) or by an affirmative EL egalitarian argument (e.g. agebased compulsory retirement). Since my aim is to assess the significance of the entirelife intuition in justifying age limits, we also need to consider possible arguments supporting age limits that do not rely on the entire-life idea. The list I am offering here is not meant to be comprehensive and is organized in three packages, each of which includes two illustrations.1⁵ First, the use of age criteria can result from sequence-related rationales, namely, concerns about doing things in the right order. For instance, one reason supporting age-based compulsory education is that it is more efficient that, up to a certain point, some school-based learning precedes on-the-job-training.1⁶ And one possible reason for the age-based prohibition on child labour or the disenfranchisement of young voters is that time for innocent life should be preserved early in life as innocence tends to be irreversibly eroded by life experience.1⁷ Hence, in both examples, there is a view that human lives should follow a certain temporal sequence, for reasons of efficiency or for other reasons, x having to be done preferably before y. Age limits are there to serve such a sequencing purpose. And they do not need to rely on an EL egalitarian support. The second package is about age and additional life expectancy. One reason for giving priority to younger patients in ICUs is a concern for maximizing additional life expectancy, that is, the number of years gained with a given medical intervention. Also, in redesigning our electoral systems, one may consider the age-adjustment of voting weights. One possible reason supporting the latter is that voting weights should espouse the degree to which people are potentially affected by political decisions. If older citizens are likely to die earlier, they are likely to be less affected by 1⁴ Note that since the defensive argument merely requires the non-worsening by age criteria of entirelife inequalities as opposed to a neutrality in terms of entire-life inequalities, it seems possible to combine the defensive and the affirmative arguments in a single case. 1⁵ See Gosseries (2014); Berndt-Rasmussen (Chapter 1, this volume). 1⁶ See Cormier and Brighouse (Chapter 13, this volume). 1⁷ See Gheaus (Chapter 8, this volume).

Ageing without Ageism?

91

many of our current decisions in the future.1⁸ Hence, in both examples, age criteria are used to reflect concerns associated with additional life expectancy. And in the voting example, it is a concern about adjusting voting power to the degree to which one is affected by collective decisions rather than an efficiency concern that is at work. Again, none of these two policies necessarily rely on an EL egalitarian support. The third package covers practices that rely on age as an informational short-cut, as a predictor for some x. They imply that age is preferable to other informational strategies. The first example is about setting a legal age for sexual majority. Of course, this involves false positives and negatives because some may end up reaching actual sexual maturity after or before the legal age. Yet, relying on an age limit can be deemed preferable to an individualized assessment because the latter might be too intrusive into people’s private lives. A second example has to do with whether, in the COVID-19 vaccination context, we should combine age criteria with comorbidities or whether we should rather use age only. Some have argued that we should exclusively rely on age as a proxy for vulnerability to COVID-19. Adding comorbidities runs the risk of violations of medical secret since information about the existence of comorbidities would have to be shared with those in charge of coordinating the distribution of vaccines.1⁹ I am not saying that each and every one of the six reasons presented by way of illustration is conclusive in the sense of being defensible all things considered. I merely want to stress that while they cover a representative range of the reasons supporting age limits and while they exhibit a significant degree of plausibility, none of them needs an ‘entire-life’ focus.

6. Conclusion It is an important task to understand what, if anything, explains the view, which many of us may share, that age limits may not be as problematic as race-based or genderbased forms of differential treatment are. Among the possible ‘specialness-makers’ of age is the idea of EL egalitarianism. I characterized the latter before distinguishing its defensive use from its affirmative use. My conjecture is that the range of cases in which age limits are justifiable is narrower than the use of age limits in our legal systems implies. Yet, whenever age limits are actually justified, two claims can be made. First, once one scratches beneath the defensive argument, one realizes that there is a very significant range of cases that are best accounted for through positive reasons that are unrelated to the EL egalitarian idea—as demonstrated in section 5. Second, looking at the affirmative argument, while it clearly extends beyond cases in which differential longevity is a core concern, it is unclear how far it does so. Note, as well, that neither of these two claims engages in a conversation with alternative 1⁸ See Gosseries (2022). 1⁹ The Association Belge des Syndicats Médicaux, the main Belgian medical practitioners trade union, challenged the Belgian vaccination strategy in Court in February 2021 on such grounds.

92

Age Limits and the Significance of Entire Lives Egalitarianism

reasons that are egalitarian in spirit and would show concern for snapshot inequalities that are irreducible to entire-life concerns. The two claims just made suggest that whenever age limits are justified, the EL egalitarian idea may actually play a central role in only a limited number of such cases. Hence, the features that render age limits special and justifiable cannot be reduced to the EL egalitarian intuition, even if the latter still has some significance. Note that I am not assuming here that an account needs to cover the full range of significant instances of age limits to qualify as a specialness-maker. What I am assuming is merely that if an account is to play a central role in our normative approach to age limits, it needs to play an affirmative role in a significant number of such cases. And what the present chapter suggests is that while the debate on the entire-life perspective definitely raises profound philosophical questions about issues such as personal identity or how we should aggregate successive events in a lifetime, and while it is one of definitive relevance for some of the important age limit practices, it seems wise not to aim to account for the moral acceptability of age limits and/or for their normative specialness through focusing exclusively, or even perhaps centrally, on the EL egalitarian intuition. While age limits tend to provide one of the best illustrations of the practical relevance of the ‘entire-life’ debate, it does not follow that the latter offers us insights that are as significant as expected once we consider the use of age limits over their whole range.

Acknowledgements Many thanks to Kim Angell, Greg Bognar, Tyler John, and Tim Meijers for suggestions on earlier drafts of this chapter.

References Arneson, Richard. 2004. ‘Luck Egalitarianism Interpreted and Defended’. Philosophical Topics 32 (1/2): 1–20. Bidadanure, Juliana. 2016. ‘Making Sense of Age-Group Justice: A Time for Relational Equality?’ Politics, Philosophy and Economics 15(3): 234–260. Broome, John. 2006. ‘The Value of Living Longer’. In Public Health, Ethics and Equity, ed. S. Anand, F. Peter and A. Sen. Oxford: Oxford University Press, 243–260. Casella, Alexandra. 2011. Storable Votes. Protecting the Minority Voice. New York: Oxford University Press. Cohen, Gerald Allan. 1989. ‘On the Currency of Egalitarian Justice’. Ethics 99: 906–944. Fleurbaey, Marc. 2005. ‘Freedom with Forgiveness’. Politics, Philosophy and Economics 4(1): 29–67. Fleurbaey, Marc, Marie-Louise Leroux, and Gregory Ponthiere. 2014. ‘Compensating the Dead’. Journal of Mathematical Economics 51: 28–41.

Ageing without Ageism?

93

Gosseries, Axel. 2003. ‘Intergenerational Justice’. In The Oxford Handbook of Practical Ethics, ed. Hugh LaFollette. Oxford: Oxford University Press, pp. 459–484. Gosseries, Axel. 2004. ‘Are Seniority Privileges Unfair?’ Economics and Philosophy 20(2): 279–305. Gosseries, Axel. 2014. ‘What Makes Age Discrimination Special? A Philosophical Look at the ECJ Case Law’. Netherlands Journal of Legal Philosophy 43(1): 59–80. Gosseries, Axel. 2019. ‘Are Inequalities between Us and the Dead Intergenerationally Unjust?’ Critical Review of International Social and Political Philosophy 22(3): 284–300. Gosseries, Axel. 2022. ‘Should Old Age Votes be Granted Less Weight?’ in Studies on the Boundary Problem in Democratic Theory, ed. Paul Bowman, Stockholm: Institute for Future Studies, pp. 149–168. Halliday, Daniel, and Tom Parr. 2022. Ageing, Justice and Work: Alternatives to Mandatory Retirement, The Cambridge Handbook of the Ethics of Ageing, ed. Christopher Wareham, Cambridge: Cambridge University Press, pp. 228–242. Lippert-Rasmussen, Kasper. 2018. ‘Discrimination’. In The Oxford Handbook of Distributive Justice, ed. Serena Olsaretti. Oxford: Oxford University Press, pp. 479–497. Lippert-Rasmussen, Kasper. 2019. ‘The EU and age discrimination: Abolish mandatory retirement!’ In Twelve Stars: philosophers chart a course for Europe, eds. Meyer, Marco, and Joachim Fritz-Vannahme. Gütersloh: Bertelsmann Stiftung, pp. 127–133. McKerlie, Dennis. 2012. Justice between the Young and the Old. New York/Oxford: Oxford University Press. Valente, Manuel Sá. 2022. ‘Proportionality without Inequality: Defending Lifetime Political Equality through Storable Votes’. Res Publica 28(4): 715–732. Van Parijs, Philippe. 2001. ‘Difference Principles’. In The Cambridge Companion to John Rawls, ed. Samuel. Freeman. Cambridge: Cambridge University Press, pp. 200–240.

7 Age Universalism Will Benefit All (Ages) Simon Birnbaum and Kenneth Nelson

1. Introduction All welfare states provide social protection for age-related social risks. However, they differ considerably in the extent to which they address various needs associated with different stages of life.1 How should these age-related structures of social protection be organized? Should we provide higher levels of protection for vulnerabilities arising at specific stages of life (such as childhood, maturity, or old age) or is it better to offer equally ambitious forms of support throughout the life course? A sensible approach for tackling these issues for the practical purposes of policy guidance should not only be informed by ethical reasoning on the values at stake, such as justice between age groups (Daniels 1988; McKerlie 2013), but also entail empirical efforts to identify and understand the social outcomes of the welfare state and the organization of income support for different stages of life (Vanhuysse and Tremmel 2019). An important lesson from the comparative social policy literature of the past few decades is that the actual propensity of voters to support (or at least tolerate) that governments take a large responsibility for the social protection of its citizens depends importantly on the political dynamics associated with different types of social policy programmes. Even if we are guided by the aim of maximizing the life prospects of the least advantaged over complete lives, or to reduce as much as possible the social impact of ‘brute luck’ on the distribution of life prospects, this does not mean that social spending should target only the poor and less fortunate citizens. Social programmes that direct economic resources exclusively to the most disadvantaged exist in all countries. However, such targeted programmes tend to achieve redistributive objectives far more effectively when they are embedded in a wider, universalist system of benefits (‘targeting within universalism’), including adequate earnings-related social insurance programmes for broader sections of the electorate (e.g. Marx et al. 2016: 21–22). This universality of the welfare state tends to be essential for increasing the budget available for social spending and ensuring the long-term legitimacy and 1 This chapter draws on our book The Generational Welfare Contract: Justice, Institutions and Outcomes (2017), written together with Tommy Ferrarini and Joakim Palme. The chapter develops a discussion of the normative implications of our findings and an updated empirical analysis in the form of a brief synthesis, taking more recent data into account. We are grateful to Axel Gosseries and an anonymous reviewer for their helpful comments on earlier drafts.

Simon Birnbaum and Kenneth Nelson, Age Universalism Will Benefit All (Ages). In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Simon Birnbaum and Kenneth Nelson (2023). DOI: 10.1093/oso/9780192894090.003.0008

Ageing without Ageism?

95

resilience of an ambitious redistributive agenda (Korpi and Palme 1998; Olivier and Noël 2018). In this chapter, we explore the neglected issue of whether similar processes of universalism are at work in relations between age groups and, if so, what this implies for how welfare states should be organized. Recognizing the apparent explanatory significance of social insurance programmes that stabilize access to accustomed living standards widely in the population, we here identify the age balance of social insurance arrangements as a possible key factor for the size and performance of welfare states. Age-balanced systems of social insurance provide (roughly) equal rates of income replacement for the needs associated with different stages of life (e.g. during parenthood, unemployment, sickness, or retirement). For example, such systems offer parental leave benefits, unemployment insurance, and old age pensions that all compensate loss of earnings at equal or very similar rates. In age-unbalanced forms of social insurance, the degree of income replacement (and, thus, the protection of income stability and accustomed living standards) instead differs substantially between programmes that target different age-related risks, thereby prioritizing some life stages over others. We argue that age-balanced social insurance matters for the levels of welfare state support available for all age groups and, in turn, a wide range of social outcomes, such as low poverty; adequate low-end incomes; and high levels of trust, happiness, and life satisfaction. The needs and claims associated with different life stages can give rise to clashes of interests and political tensions between age groups, with negative consequences for social cohesion. However, age-balanced social insurance can play a key role in enabling stable cooperative gains and preventing divisive competition between age groups over the social budget. Our contribution to normative debates on age policies is threefold. In the search for adequate policy prescriptions, we will (i) demonstrate the relevance of linking philosophical debates on justice between age groups to the welfare state literature on social policy impacts; (ii) suggest measures and methods for analysing core agerelated structures of welfare states; and (iii) develop a tentative, instrumental case for age-balanced social insurance. We begin by introducing our theoretically grounded expectations about the advantages of age-balanced social insurance and its moral significance (section 1), methodological considerations (section 2), and empirical results (section 3). Finally, we discuss the normative implications of our findings and questions for future research, also taking recent developments into account (section 4).

2. Positive-sum solutions in generational politics? Theory and hypothesis Our interest in age-balanced social insurance is motivated by empirically grounded expectations on when and how welfare states succeed in attracting robust political support and in achieving desirable social outcomes. Our argument suggests

96

Age Universalism Will Benefit All (Ages)

pragmatic advantages of age-balanced welfare states. We approach our topic with a broadly egalitarian mindset (Birnbaum et al 2017: Chapter 2), but our argument in this chapter does not presuppose or advance one specific normative conception of justice (or other relevant values). While recognizing that a plurality of goals and standards should be considered in evaluating arrangements for addressing age-related social risks, the following analysis will (in the interest of focus) address one key objective: how can we design social policies for age-related risks in ways that serve efficiency—positive-sum solutions—by improving the general prospects of all major age groups? This efficiency goal is important in its own right but it is also highly relevant for identifying institutions and policies that may serve wider egalitarian objectives. When all age groups perceive that their prospects are firmly interconnected through mutually beneficial cooperation, bonds of solidarity and trust are also far more likely to permeate our societies. Norman Daniels’s influential ‘prudential lifespan approach’ to justice between age groups emphasizes that we must not approach redistribution between younger and older members of the community as a context for divisive competition for scarce resources. Instead, the welfare state should be considered a (potential) means to establish and maintain a stable welfare-promoting form of cooperation between successive generations as they are moving through the different stages of life: ‘if these institutions are prudently designed, we each benefit throughout our lives’ (Daniels, 1988: 155). If the promotion of stable mutual welfare gains is a central rationale for addressing age-related social risks through the welfare state, it becomes crucial to reveal systematic differences in the extent to which specific policy designs improve overall living conditions (or fail to do so) in all major age groups. Our analyses of the prospects for mutual benefits focus on the policy impacts on different age groups (or, more precisely, age-related risks and needs—more on this below) at a given point in time. We will thus stay out of more historically orientated analyses of how the welfare state has benefited some cohorts and disadvantaged others over complete lives, as discussed in claims about ‘sacrificed generations’ (see, e.g. Chauvel and Schröder 2014). We do not (of course) deny the moral significance of whether cooperation gains under different age polices are stable or distributed evenly between cohorts over time. However, if we want to find out how processes of alliance-building and perceived legitimacy affect the social outcomes of age polices, it is particularly relevant to study the impact of how social insurance treats different age-related risks and needs in the same political community at a given point in time. The reason is simply that the causal and moral responsibility of states (or governments) for disparities between groups are likely to be perceived as much more direct, visible, and of greater relevance to legitimacy assessments in this case compared to the long-term distribution of benefits between cohorts.2 Still, gainful cooperative relations between age groups are likely to be a major social asset as states seek to

2 Judgements on the latter relate to actions and circumstances in the distant past and speculations about future welfare developments or distributive patterns—many of which will seem hard to identify and beyond the control of (current) governments.

Ageing without Ageism?

97

confront (other) injustices in an even-handed manner. The age-balanced approach can therefore also contribute to further the goal of justice between cohorts. We conceptualize age structures of social policies as forging an implicit generational welfare contract between older and younger members of the political community (Birnbaum et al. 2017). Our hypothesis is that age-balanced (or, for short, balanced) social insurance, offering equal levels of income replacement for needs and risks associated with different life stages, is more likely than unbalanced schemes to generate outcomes that are favourable to citizens of all ages. Balance-of-income replacement in social insurance does not imply equal size of benefits in absolute amounts to citizens of all ages. Instead, it is supposed to capture the extent to which the welfare state equally supports the need for income protection associated with different stages of life (e.g. whether parental leave benefits and oldage pensions both compensate loss of earnings by, let us say, 60 per cent). We make no wider claims about public services, such as school infrastructure, active labour market policies and lifelong learning, or health-care provision. Social policy and the welfare state are, of course, more than social insurance. Nonetheless, insurances for age-related social risks constitute an essential and highly visible part of every mature welfare state. It is also a policy area where extant data allow us to empirically analyse long-term policy changes in a large number of countries. Our expectation that age-balanced social insurance is more likely to support favourable outcomes in all age groups is based on an analytical framework that directs our attention to two types of processes associated with the welfare state: first, how the political economy mechanisms induced or reinforced by social insurance— formation of group identities and strategic alliance building—affect the prospects for long-term political support for redistribution; second, how social insurance affects the perceived legitimacy of ambitious welfare state institutions via considerations of fairness and political trust. To understand the performance of welfare states, we must consider how social policy designs contribute both to stable majorities that see it in their self-interest to support redistribution (Esping-Andersen 1990) and to the perceived legitimacy of welfare state institutions (Rothstein 1998).

2.1 Alliance-building The political economy argument on alliance-building in the welfare state builds on an influential strand of thought in comparative social policy research according to which adequate protection for a broad majority of citizens will sustain wider political support for raising taxes and increase levels of provision also to the poor and less advantaged people (Esping-Andersen 1990; Nelson 2004; cf. Titmuss 1968). Marketorientated welfare states—with only modest levels of transfers and a stronger emphasis on means-testing and selectivity in social policy—tend to reinforce conflicting interests and competitive relations between poor (net receivers) and better-off (net contributors) citizens. In contrast, encompassing forms of universalism, including a robust form of earnings-related social insurance, are more likely to establish and

98

Age Universalism Will Benefit All (Ages)

sustain powerful alliances between disadvantaged groups and the middle classes in support of the welfare state, ensuring that the poor need not stand alone (Olivier and Noël 2018). Thus, high levels of redistribution should primarily be expected in countries where social policies cut across interest groups and include the needs of all, or nearly all, citizens in one common risk pool. Korpi and Palme (1998) famously expressed this relationship between universalism and redistribution as the ‘paradox of redistribution’, stating that ‘the more we target welfare state benefits at the poor only, the less likely we are to reduce poverty and inequality’ (1998: 681). In spelling out this thesis, they emphasize that earningsrelated social insurance plays a key role in increasing redistributive budgets over time by including the middle class more fully in the welfare state. Thus, social policies that rely heavily on means-testing or equal flat-rate benefits, instead of earnings-related social insurance, will be less successful in serving egalitarian objectives. Why expect that similar processes of social stratification and alliance-building are at work in the context of age-related social insurance? Welfare states both structure and are structured by generational relations. They not only provide resources to people in different stages of life but may also encourage (or discourage) broader popular support for social policy that cuts across age-specific interests. As illustrated by the fears about growing generational tensions related to ageing populations (Kotlikoff and Burns 2012) and the increasingly precarious living conditions of young people (Bessant et al. 2017), the age structure of social policies may be a source of fragmentation, division of interest, and political tension in society. However, our hypothesis on age-balanced social insurance suggests that if designed properly, the welfare state can prevent divisive and competitive processes associated with age-based stratification. With an age-balanced approach, welfare states may help to promote resilient forms of gainful cooperation (e.g. between parents, workers, and the elderly) that are supportive to the income needs of all age groups. If this expectation is correct, age-balanced social insurance should be associated with higher levels of income replacement and more favourable social outcomes for all major age-related risk groups.

2.2 Perceived legitimacy The extent to which the welfare state reinforces stable cooperative exchange across age groups is not only linked to processes of stratification and alliance-building in social insurance but also closely intertwined with the perceived fairness and legitimacy of state institutions (Rothstein 1998). No alliances for expansive, redistributive solutions are likely to come far, or to remain steadfast in consenting to ambitious social policies over time, unless key actors trust the state to implement public commitments in reasonably fair, effective, and non-corrupted ways. Following Margaret Levi (1998), we may expect that citizens are far more willing to accept an expansion of government (including social insurance)—and contribute to its funding—when they trust that state institutions will not fail to provide people with their fair share of support and are confident that burdens are shared equitably.

Ageing without Ageism?

99

The universality of welfare states often seems to play a key role in satisfying and reproducing these conditions for people’s ‘contingent consent’ to social policy, thereby facilitating more ambitious redistributive programmes. Unpacking this link between universalism and the perceived legitimacy of welfare states, earlier findings have emphasized the perceived fairness, transparency, and efficacy of universal programmes (Kumlin and Rothstein 2005), as well as the equality of their outcomes as relevant factors (Uslaner 2002). Providing all (or most) members of society protection against various types of social risks without heavy reliance on the discretionary power and bureaucratic involvement associated with means-tested benefits appears conducive to such positive attitudes (Rothstein 1998; Rothstein and Uslaner 2005). Universalism may help to sustain the everyday experience of fair and well-functioning government institutions (political trust), thereby also facilitating cooperative relations across groups (social trust) and a welcoming attitude to ambitious social programmes. However, the extent to which welfare states are age-balanced is an overlooked— and potentially very important—element in the debate about universalism and the perceived legitimacy of welfare state institutions. Countries with age-balanced social insurance systems can help to support virtuous circles in policymaking by signalling in people’s everyday lives that no age group is left behind and that the needs of different life stages are addressed without apparent disparities in the treatment of different age-related risk groups. This speaks directly to key conditions for people’s contingent consent to the welfare state. Young parents in need, or working-age citizens facing high unemployment risks, may be reluctant when asked to contribute substantially to the needs of the elderly if they lack equivalent levels of security to build a future for themselves and their dear ones. The elderly may reasonably feel reluctant to contribute via taxes to the needs of younger citizens unless they are confident of receiving adequate pensions after a long working life. For these reasons, we expect that a universalistic strategy of smoothing income replacement in social insurance across major age-related social risks offers a promising basis for supporting a strong and reciprocal commitment to the welfare state in all age groups.

3. Measuring and comparing generational welfare contracts The conceptualization and measurement of welfare states and social policy is extensively discussed in research (Clasen and Siegel 2007; van Oorschot 2013; Otto 2018). The most common approach is to base the empirical analysis on social expenditures from the national accounts. Although spending patterns are valuable for some research purposes, they seem less relevant for exposing fundamental generational structures of welfare states of key importance for stratification processes and public support for ambitious welfare state commitments. It is well known that social expenditures are highly sensitive to changes in social need and demographic shifts. Expenditures on out-of-work benefits tend to increase during economic downturns

100

Age Universalism Will Benefit All (Ages)

as unemployment rises. And as the numbers of pensioners increase in periods of population ageing and fertility decline, so does the share of resources countries spend on old-age benefits and so forth. However, this does not necessarily mean that out-ofwork benefits and old-age pensions have increased on the recipient side. Quite the opposite; with the return of mass unemployment in Western Europe, several countries introduced cutbacks to their out-of-work benefits (Nelson 2004). Similarly, in many countries, population ageing has coincided with new pension benefits being lower than they were for earlier retirees (Tepe and Vanhuysse 2009). What matters for alliance-building in the welfare state and the perceived fairness of redistribution across ages is arguably not so much how aggregate spending is divided between the young and the elderly (which may be explained by demographic shifts or changes in business cycles, as noted above). Rather, what matters is to what extent the particular risks people face in different vulnerable periods of life are adequately covered by the welfare state (i.e. whether social insurance provides an income that secures accustomed living standards). In this study, we therefore shift focus from expenditures to income replacement. Our data on income replacement in social insurance is from the Social Policy Indicators (SPIN) database (Nelson et al. 2020). SPIN data is explicitly collected to facilitate analyses between countries, over time and across policy programmes. SPIN income replacement data is calculated on the basis of national legislation using model family techniques (Bradshaw et al. 1993). For child-related risks, we use a benefit package including child allowances, child tax allowances and credits, and post-natal parental leave benefits and maternity grants paid in relation to childbirth. For working-age risks, we include unemployment insurance benefits and sickness insurance benefits, and for old-age risks we use retirement benefits. For each age-related social risk, entitlements are calculated net of taxes and expressed as percentages of an average production worker’s net wage. In cases where entitlements are established with reference to previous wages, benefits are calculated at the wage level of an average production worker. Birnbaum et al. (2017) provide more details about the measurement of income replacement in age-related social insurance.

4. Results The focus below is on differences between age-balanced and age-imbalanced structures of social insurance. For each country, we have combined data for the period 1990–2015 to improve the stability of our classifications. For each year, the dispersion of income replacement across age-related social risks was assessed by the relative standard deviation, and we used the same cut-off as Birnbaum et al. (2017) to determine whether social insurance for different age-related risks is balanced or not.3 Countries that, on average, had a relative standard deviation below 20 per cent were 3 The relative standard deviation is a simple measure of dispersion and defined as the ratio between the standard deviation and the arithmetic mean, multiplied by 100.

101

Ageing without Ageism?

categorized as balanced. For individual years, countries may occasionally deviate from these broad patterns (see more below). Figure 7.1 shows the three profiles of income replacement in social insurance observed in our data: age-balanced, pro-work, and pro-old. Only country group averages are shown.⁴ Income replacement is balanced and very similar across the agerelated risk categories in Austria, Belgium, Denmark, France, Finland, Norway, and Sweden. Thus, no age-related risk category is left behind with substantially weaker protection, as in countries with more unbalanced profiles. Countries with (unbalanced) pro-work profiles provide higher levels of income replacement for working-age risks (unemployment and sickness) compared to the risks associated with childhood and old age. Included in this category are Germany, Japan, the Netherlands, and Switzerland. Another group of countries have (unbalanced) pro-old profiles, in which the degree of income replacement in social insurance is tilted in favour of the elderly. Included in this group are Australia, Canada, Ireland, Italy, New Zealand, the United Kingdom, and the United States.⁵ It is this pro-old profile state that tends to be at the centre of most discussions about generational conflict in social policymaking related to competing claims of different age groups (Laslett and Fishkin 1992; Tepe and Vanhuysse 2009; Berry 2012; Kotlikoff and Burns 2012). Notably, risk groups that are favoured in countries with major age imbalances in social insurance do not receive higher levels of protection compared to countries with age-balanced structures. Quite the opposite; income replacement in social insurance tends to be higher in the age-balanced cluster, thus supporting our hypothesis of Pro-work

e O

ld

ag

e W

or

ki

ng

ag

ho od

Percentage

100 80 60 40 20 0

Ch ild

e ld O

ng ki or W

ag

ag e

Percentage

ho od

e

Pro-old

100 80 60 40 20 0

Ch ild

O

ld

ag

e ag ng ki or W

Ch ild

ho od

Percentage

Balanced 100 80 60 40 20 0

Figure 7.1 Age-related profiles of social insurance income replacement (country group averages 1990–2015) Note: The balanced profile includes Austria, Belgium, Denmark, Finland, France, Sweden and Norway. The pro-work profile includes Germany, Japan, the Netherlands, and Switzerland. The pro-old profile includes Australia, Canada, Ireland, Italy, New Zealand, the United Kingdom, and the United States. Source: The Social Policy Indicators (SPIN) Database. ⁴ The degree of income replacement in social insurance for each country is shown in appendix Tables 7.A1–7.A3. Yearly data is available upon request by the authors. ⁵ Income replacement in Canada and Italy describes a flatter generational profile than other countries with pro-old profiles, although replacement rates are substantially lower than in countries with clearly age-balanced social insurance schemes.

102

Age Universalism Will Benefit All (Ages)

positive-sum solutions in generational politics. The exception is for working-age risks in countries with pro-work profiles, where the degree of income replacement is on a par with (but not higher than) that of the age-balanced countries.⁶ There have been some interesting changes in the generational profiles of income replacement since 1990. Figure 7.2 shows the balance of income replacement in age-related social insurance by type of generational profile in 18 Organisation for Economic Co-operation and Development (OECD) countries for the period 1990– 2015. Up to the turn of the new millennium, there was a clear trend towards more age-balanced income replacement in social insurance, irrespective of type of profile. More recently, this uniform pattern has changed. Whereas countries with pro-old profiles continued to develop in a more balanced direction, the opposite is true for income replacement in the balanced and pro-work countries. It is not possible, in our data, to observe a country that, for an extended period of time, can be categorized as pro-child.⁷ However, a few countries seem to be moving 70 60

Percentage

50 40 30 20 10 0

1990

1995 Balanced

2000

2005 Pro-work

2010

2015

Pro-old

Figure 7.2 Age-balance of income replacement in social insurance by type of generational welfare contract in 18 OECD countries, 1990–2015 Note: The vertical axis shows the relative standard deviation of social insurance income replacement across age-related social risks. Source: The Social Policy Indicators (SPIN) Database.

⁶ The relationship between balance and generosity is not due to our classification of countries into different groups. The correlation between the degree of income replacement in social insurance and its dispersion (i.e. relative standard deviation) is clearly negative for each age-related social risk (−0.84 for childhood-related risks, −0.74 for working-age risks, and −0.49 for old-age risks). A perfect negative/positive correlation has a value of −1/+1. Values close to zero indicate lack of relationship. Birnbaum et al (2017) apply more sophisticated multivariate statistical techniques but observe similar negative associations between balance and generosity in age-related social insurance. ⁷ Taking private transfers between generations into account may change the picture and, in some countries, suggest more pro-child distributions, as argued by Gal et al. (2018) and Hammer et al. (2018). Our focus is on social policy and income replacement in social insurance, not intergenerational transfers (whether public or private) per se. While intra-familial transfers are an important form of redistribution,

103

Ageing without Ageism? Norway

Sweden 100

80

80

80

20

60 40 20

10

00

20

90

20

80

19

70

19

10

00

20

90

20

80

19

70

19

19

60

0

19

10

00

20

90

20

80

19

19

19

19

70

0

60

0

40

60

20

60

19

40

19

60

Percentage

100 Percentage

Percentage

Japan 100

Childhood Working-Age Old-age

Figure 7.3 Income replacement in social insurance for three age-related social risks in Japan, Norway, and Sweden, 1960–2015 Source: The Social Policy Indicators (SPIN) Database.

in this direction, including Japan, Sweden, and Norway (Figure 7.3). In all three countries, protection against child-related risks is nowadays higher than for working-age or old-age risks. Japan is becoming increasingly unbalanced due to an exceptional rise of income replacement for this category since 1990. In Norway and Sweden, the development towards less age-balanced systems primarily reflects cutbacks to old-age pensions. Table 7.1 shows six social outcomes (poverty, income level of the tenth decile, social trust, political trust, happiness, and life satisfaction) by type of generational profile. Once again, only country averages are shown.⁸ Common to previous income distribution research on OECD countries, we conceptualize poverty in relative terms, using a benchmark of 50 per cent of national median disposable household income in the total population. The income level of the tenth decile is expressed in purchasing adjusted US dollars.⁹ To adjust for economies of scale within households of different size and composition, we use the so-called old OECD equivalence scale. The first household member is assigned a value of 1, each additional adult a value of 0.7, and each child a value of 0.5. Income data is from the Luxembourg Income Study (LIS), where we have combined various waves from 1990 onwards. For social and political trust, happiness, and life satisfaction, we use two data sources: the European Social Survey (ESS) for the European countries and the World Value Survey (WVS) for the non-European countries. For both surveys, we merged different waves of data: nine waves 2002–2018 for the ESS and three waves 1999–2014 for the WVS. In the ESS, we used the following question for social trust: ‘Generally they belong to the private sphere and are difficult for governments to regulate through policy. Far from everyone receives substantial intra-familial transfers on a regular basis. It is therefore questionable whether private transfers between generations are an effective means of fostering cooperation (including sharing of risk and resources) between age groups for the benefit of citizens in all of society, regardless of social background. ⁸ Data for each country are shown in appendix Tables 7.A1–7.A3. ⁹ The tenth decile shows the income level below which 10 per cent of the poorest individuals fall.

104

Age Universalism Will Benefit All (Ages) Table 7.1 Type of generational profile in age-related social insurance and selected outcomes in 18 OECD countries Childhood

Working age

Old age

Poverty (%)

Balanced Pro-work Pro-old

7.9 10.0 14.8

6.1 6.3 10.3

3.0 6.5 5.4

Income 1st decile ($US PPPS)

Balanced Pro-work Pro-old

12,149 11,644 8,812

11,162 12,383 9,840

10,983 10,737 9,435

Social trust (%)

Balanced Pro-work Pro-old

60.0 53.9 51.5

57.9 50.9 47.9

53.9 47.6 49.1

Political trust (%)

Balanced Pro-work Pro-old

50.0 44.4 32.5

49.2 43.8 33.8

44.9 43.9 32.5

Happiness (%)

Balanced Pro-work Pro-old

93.3 90.6 88.2

91.1 88.0 87.4

89.9 89.4 89.1

Life satisfaction (%)

Balanced Pro-work Pro-old

88.3 84.3 81.5

87.2 82.8 80.9

86.9 87.0 84.9

Note: The balanced profile includes Austria, Belgium, Denmark, Finland, France, Sweden, and Norway. The pro-work profile includes Germany, Japan, the Netherlands, and Switzerland. The pro-old profile includes Australia, Canada, Ireland, Italy, New Zealand, the United Kingdom, and the United States. Source: authors’ own calculations based on various waves of the Luxembourg Income Study (LIS), the European Social Survey (ESS), and the World Value Survey (WVS).

speaking, would you say that most people can be trusted, or that you can’t be too careful in dealing with people?’ Respondents were asked to choose a score on the ranking of 0 (cannot be too careful) up to 10 (most people can be trusted). In the WVS, the question is: ‘Generally speaking, would you say that most people can be trusted or that you need to be very careful in dealing with people?’ Respondents could either answer ‘Most people can be trusted’ or ‘Need to be very careful’. For political trust, we used the ESS question, ‘How much do you personally trust each of the following institutions: parliament, political parties and the legal system?’ A response scale between 0 and 10 was used for each item (ranging from no trust at all to complete trust). In the WVS, we relied on the following question: ‘For each of the following organizations, could you tell me how much confidence you have in them: parliament, political parties and the courts?’ The WVS applies a four-point scale ranging from a great deal of confidence to none at all.

Ageing without Ageism?

105

For happiness and life satisfaction, we used the following questions in the ESS: ‘Taking all things together, how happy would you say you are?’ and ‘All things considered, how satisfied are you with your life as a whole nowadays?’ For each question, respondents were asked to pick a score between 0 (extremely dissatisfied/unhappy) and 10 (extremely satisfied/happy). In the WVS, the questions were: ‘Taking all things together, how happy would you say you are?’ and ‘All things considered, how satisfied are you with your life as a whole these days?’ For each question, respondents were asked to judge their score on a 10-point scale varying between 1 ‘completely dissatisfied’ and 10 ‘completely satisfied’. In order to harmonize the two data sets, we dichotomized all response categories into a value of 1 (trustful/satisfied/happy) and a value of 0 (distrustful/unsatisfied/unhappy). Our results show that countries with age-balanced social insurance not only tend to have low levels of (relative) poverty but also low-end incomes seem to be higher than in countries with major age imbalances in social insurance. Furthermore, people in countries with age-balanced profiles also tend to be more trustful, happier, and more satisfied with life. These cross-national patterns appear for each age-related risk category. Thus, it is not the case that elderly people are financially better off or happier and more inclined to trust others in pro-old systems, nor are people of working age better protected against poverty and more satisfied with life in countries with prowork profiles. Together, our data therefore suggest that countries with age-balanced income replacement in social insurance are, indeed, more likely to produce outcomes that are beneficial to the living conditions and well-being of people in all age-related risk groups.

5. Conclusion: The generational paradox of redistribution The paradox of redistribution conveys a warning against overreliance on meanstested or meagre flat-rate benefits that fail to adequately respond to the needs of income security among middle-income groups, thus excluding large portions of the better off from the welfare state (Korpi and Palme 1998). Our findings in this chapter suggest that a second element should be added to the paradox: the more we target income replacement to the needs associated with a specific life stage, the less likely we are to improve the levels of support and living conditions of this particular age group. Formulated in more positive terms: age-balanced welfare states tend to provide higher levels of income replacement for age-related risks throughout the life cycle and perform better in terms of key social outcomes in all age groups, including poverty, trust, and subjective well-being. This conclusion should be interpreted as a stable regularity in affluent, democratic welfare states rather than a universal iron law. Some recent findings raise new questions on the key role of encompassing forms of universalism for poverty prevention and the resilience of welfare states (cf. Brady and Bostic 2015). Our analyses

106

Age Universalism Will Benefit All (Ages)

in this chapter indicate that some countries with age-balanced welfare states—we mentioned Norway and Sweden above—have lately become more unbalanced and experienced cutbacks to social protection, albeit from very high levels. It is much too early to tell whether these recent developments offer an (hitherto unobserved) example of changing dynamics in the political economy of redistribution. But the question of whether this is just a brief, temporary deviation from a stable, longterm pattern or the beginning of a more fundamental structural shift calls for further analysis. Although the analyses on income replacement in social insurance above clearly reveal interesting connections, our methods cannot rule out the possibility that there is some other, underlying unobserved factor affecting our results on age-balanced welfare states and the outcomes they seem to produce. However, the more comprehensive triangulation of descriptive data analysis and statistical regressions in Birnbaum et al. (2017) tests and dismisses the most plausible candidates for such unobserved factors, including both economic and demographic variables. The consistency of our findings across a broad set of countries, years, and outcomes is striking. This conclusion calls for more research on the causes and consequences of age-balanced welfare states and the wider configurations of policies that may effectively support positive-sum solutions in generational politics. An extension of research beyond income replacement in social insurance, including also in-kind benefits and services, seems especially warranted, once the necessary institutional data becomes available.

Table 7.A1 Variables included in the empirical analyses: Childhood risk category by country, averages for the period 1990–2015 Country

Income replacement (%)

Poverty (%)

Income 1st decile ($US PPPS)

Social trust (%)

Political trust (%)

Happiness (%)

Life satisfaction (%)

Australia Austria Belgium Canada Denmark Finland France Germany Ireland Italy Japan Netherlands New Zealand Norway Sweden Switzerland United Kingdom United States

17.3 61.7 48.5 51.1 56.1 77.8 50.9 55.3 25.9 54.0 61.2 34.0 14.8 96.3 85.3 34.8 35.0 8.3

14.3 10.5 9.0 15.9 6.3 5.8 11.9 9.5 11.4 21.4 11.7 8.9 m.d. 5.7 6.4 9.7 16.5 23.9

10,017 11,697 10,654 10,475 13,292 12,266 9,861 11,234 8,961 5,674 10,263 11,758 m.d. 15,055 12,218 13,322 8,473 9,270

60.0 44.9 44.3 54.2 79.9 75.7 29.3 40.8 49.9 36.7 54.4 64.9 47.4 77.2 68.6 55.2 44.2 68.2

32.8 38.5 37.1 43.8 68.8 61.5 27.1 37.5 28.7 22.8 29.5 54.6 32.9 60.8 56.6 55.9 31.8 35.0

90.9 86.8 92.5 92.4 96.6 96.5 93.2 86.1 85.8 81.6 86.7 95.7 92.4 93.7 93.7 94.0 84.3 89.7

82.6 85.2 87.2 85.8 95.8 94.3 73.5 79.6 79.6 78.5 73.3 92.7 82.9 91.0 91.0 91.6 79.0 82.1

Note: m.d.= missing data. Source: authors’ own calculations based on various waves of the Social Policy Indicators Database (SPIN), the Luxembourg Income Study (LIS), the European Social Survey (ESS), and the World Value Survey (WVS).

Table 7.A2 Variables included in the empirical analyses: Working age risk category by country, averages for the period 1990–2015 Country

Income replacement (%)

Poverty (%)

Income 1st decile ($US PPPS)

Social trust (%)

Political trust (%)

Happiness (%)

Life satisfaction (%)

Australia Austria Belgium Canada Denmark Finland France Germany Ireland Italy Japan Netherlands New Zealand Norway Sweden Switzerland United Kingdom United States

23.9 72.6 70.8 51.5 57.5 73.9 57.8 83.1 31.0 56.0 72.5 69.3 24.5 81.8 74.3 71.7 17.6 25.6

10.5 8.9 6.6 12.0 4.7 4.3 7.5 7.1 10.6 13.6 6.9 4.9 m.d. 5.6 5.1 6.4 10.7 15.0

10,979 12,086 11,376 9,938 11,692 10,014 10,237 10,981 9,337 8,400 10,391 11,639 m.d. 11,830 10,897 16,521 8,940 11,444

51.8 47.1 44.5 46.7 74.3 72.6 30.1 40.7 50.9 37.2 46.3 63.1 57.9 72.1 64.5 53.5 46.6 44.2

35.9 38.5 40.1 47.7 64.7 61.0 27.5 38.2 31.1 22.0 24.9 56.8 29.4 58.6 54.0 55.3 34.0 36.8

89.1 85.3 92.0 91.0 94.4 93.0 90.9 83.4 83.9 82.6 80.9 94.3 91.2 91.7 90.6 93.4 85.9 87.8

78.3 82.9 87.2 86.4 93.6 91.4 76.9 78.4 78.3 78.8 70.1 91.9 83.1 89.3 89.0 90.8 80.0 81.4

Note: m.d.= missing data. Source: authors’ own calculations based on various waves of the Social Policy Indicators Database (SPIN), the Luxembourg Income Study (LIS), the European Social Survey (ESS), and the World Value Survey (WVS).

Table 7.A3 Variables included in the empirical analyses: Old age risk category by country, averages for the period 1990–2015 Country

Income replacement (%)

Poverty (%)

Income 1st decile ($US PPPS)

Social trust (%)

Political trust (%)

Happiness (%)

Life satisfaction (%)

Australia Austria Belgium Canada Denmark Finland France Germany Ireland Italy Japan Netherlands New Zealand Norway Sweden Switzerland United Kingdom United States

46.8 74.5 65.8 62.6 60.0 77.0 61.6 47.7 54.5 74.6 57.3 59.5 51.3 73.0 69.8 42.2 57.5 68.2

6.3 4.6 7.2 2.8 1.5 1.1 4.1 3.6 4.7 6.4 13.9 2.2 m.d. 0.9 1.2 6.1 6.1 11.7

9,112 11,500 9,909 12,617 11,292 9,938 10,461 10,534 9,116 7,617 7,858 12,574 m.d. 13,079 10,705 11,983 8,336 9,813

52.4 36.6 39.6 43.5 72.0 70.6 27.5 36.7 52.0 32.9 41.9 61.7 59.9 72.3 58.9 50.2 51.2 51.6

31.1 39.5 31.6 45.5 60.2 56.3 24.3 33.9 34.0 23.0 39.7 52.6 29.8 54.0 48.7 49.5 34.3 29.7

93.3 84.2 88.6 94.7 94.5 92.4 87.6 82.2 87.6 73.7 86.9 94.5 94.9 91.6 90.5 93.8 88.9 90.4

86.0 84.6 85.1 89.3 95.2 91.5 72.8 81.7 84.8 73.0 80.3 93.1 89.7 89.3 89.8 93.1 84.3 87.4

Note: m.d.= missing data. Source: authors’ own calculations based on various waves of the Social Policy Indicators Database (SPIN), the Luxembourg Income Study (LIS), the European Social Survey (ESS), and the World Value Survey (WVS).

110

Age Universalism Will Benefit All (Ages)

References Berry, Craig. 2012. The Rise of Gerontocracy? Addressing the Intergenerational Democratic Deficit. London: Intergenerational Foundation. Bessant, Judith, Rys Farthing, and Rob Watts. 2017. The Precarious Generation: A Political Economy of Young People. London: Routledge. Birnbaum, Simon, Tommy Ferrarini, Kenneth Nelson, and Joakim Palme. 2017. The Generational Welfare Contract. Cheltenham: Edward Elgar. Bradshaw, Jonathon, John Ditch, Hilary Holmes and Peter Vilhiteford. (1993). A Comparative Study of Child Support in Fifteen Countries. Journal of European Social Policy, 3(4), 255-271. Brady, David, and Amie Bostic. 2015. ‘Paradoxes of Social Policy Welfare Transfers, Relative Poverty, and Redistribution Preferences’. American Sociological Review 80 (2): 268–298. Chauvel, Louis, and Martin Schröder. 2014. ‘Generational Inequalities and Welfare Regimes’. Social Forces 92 (4): 1259–1283. Clasen, Jochen, and Nico A. Siegel, eds. 2007. Investigating Welfare State Change: The 'Dependent Variable Problem' in Comparative Analysis. Cheltenham: Edward Elgar. Daniels, Norman. 1988. Am I My Parents’ Keeper? An Essay on Justice between the Young and the Old. Oxford: Oxford University Press. Esping-Andersen, Gøsta. 1990. Three Worlds of Welfare Capitalism. Cambridge: Polity Press. Gal, Robert, Pieter Vanhuysse, and Lili Vargha. 2018. ‘Pro-Elderly Welfare States within Child-Oriented Societies’. Journal of European Public Policy 25 (6): 944–958. Hammer, Bernhard, Tanja Istenič, and Lili Vargha. 2018, ‘The Broken Generational Contract in Europe: Generous Transfers to the Elderly Population, Low Investments in Children’. Intergenerational Justice Review 12 (1): 21–31. Korpi, Walter, and Joakim Palme. 1998. ‘The Paradox of Redistribution and Strategies of Equality: Welfare State Institutions, Inequality, and Poverty in the Western Countries’. American Sociological Review 63 (5): 661–687. Kotlikoff, Laurence. J., and Scott Burns. 2012. The Clash of Generations. Cambridge, MA: MIT Press. Kumlin, Staffan, and Bo Rothstein. 2005. ‘Making and Breaking Social Capital: The Impact of Welfare-State Institutions’. Comparative Political Studies 38 (4): 339–365. Laslett, Peter, and James S. Fishkin, eds. 1992. Justice between Age Groups and Generations. New Haven, CT: Yale University Press. Levi, Margaret. 1998. Consent, Dissent, and Patriotism. New York: Cambridge University Press. Marx, Ive, Lina Salanauskaite, and Gerlinde Verbist. 2016. ‘For the Poor, but Not Only the Poor: On Optimal Pro-Poorness in Redistributive Policies’. Social Forces 95 (1): 1–24. McKerlie, Dennis. 2013. Justice between the Young and the Old. Oxford: Oxford University Press.

Ageing without Ageism?

111

Nelson, Kenneth. 2004. ‘Mechanisms of Poverty Alleviation in the Welfare State. A Comparative Study of Anti-Poverty Effects of Non Means-Tested and Means-Tested Benefits in Five Countries in the 1990s’. Journal of European Social Policy 14 (4): 371–390. Nelson, Kenneth., Daniel Fredriksson, Tomas Korpi, Walter Korpi, Joakim Palme, and Ola Sjöberg. 2020. ‘The Social Policy Indicators (SPIN) Database’. International Journal of Social Welfare 29 (3): 285–289 . Olivier, Jacques., and Alain Noël. 2018. ‘The Case for Welfare State Universalism, or the Lasting Relevance of the Paradox of Redistribution’. Journal of European Social Policy 28 (1): 70–85. Otto, Adeline. 2018. ‘A Benefit Recipiency Approach to Analysing Differences and Similarities in European Welfare Provision’. Social Indicators Research 137 (2): 765–788. Rothstein, Bo. 1998. Just Institutions Matter: The Moral and Political Logic of the Universal Welfare State. New York: Cambridge University Press. Rothstein, Bo, and Eric M. Uslaner. 2005. ‘All for All: Equality, Corruption, and Social Trust’. World Politics 58 (1): 41–72. Rowntree, B. Seebohm. 1901. Poverty: A Study of Town Life. London: Macmillan. Tepe, Markus, and Pieter Vanhuysse. 2009. ‘Are Aging OECD Welfare States on the Path to Gerontocracy?’ Journal of Public Policy 29 (1): 1–28. Titmuss, Richard. M. 1968. Commitment to Welfare. London: Allen and Unwin. Uslaner, Eric. M. 2002. The Moral Foundations of Trust. Cambridge: Cambridge University Press. van Oorschot, Wim. 2013. ‘Comparative Welfare State Analysis with Survey-Based Benefit Recipiency Data: The “Dependent Variable Problem” Revisited’. European Journal of Social Security 15 (3): 224–249. Vanhuysse, Pieter., and Jörg Tremmel. 2019. ‘Measuring Intergenerational Justice for Public Policy’. In Routledge Handbook of Ethics and Public Policy, ed. Andrei Poama and Annabelle Lever. London: Routledge, pp. 472–486.

PA RT II

POLICY PROPOSALS

8 ‘Let Them Be Children?’ Age Limits in Voting and Conceptions of Childhood Anca Gheaus

1. Introduction My aim in this chapter is to explore how recent developments in the philosophy of childhood can be brought to bear on debates about lowering the voting age. I don’t argue for any particular conclusion. Rather, I offer a map of how arguments for and against enfranchising children, or at least adolescents, can be enriched by reflecting on what, according to several contemporary philosophers, is uniquely good or bad about being a child. The vast majority of the literature about children’s disenfranchisement is dominated by a traditional, mere deficiency, account of childhood, according to which its distinguishing normative feature is the incomplete development of agential powers.1 I assume it is true that children’s agential powers are underdeveloped but that the mere deficiency account is incomplete: a conception of childhood needs to also include an account of value in childhood. On one axiological view, the same features that impair children’s agency are also prudentially bad for children. More recently, however, philosophers have advanced the claim that, during childhood, we also have unique, or at least privileged, access to some very important sources of value and that missing out on such value makes individuals’ lives worse overall. In section 2, I briefly contrast these two views. Which of the two axiological views of childhood is assumed as correct has implications for the arguments advanced—or not yet advanced!—in the debate about children’s disenfranchisement. The third, main section of this chapter explores these implications. Specifically, it does three things. First, it explains how the mere deficiency view of children’s agency seems to have been assumed by most arguments brought both in favour of and against lowering age limits in voting. Second, it points to further implications of the first axiological view on this issue. And, third, it elaborates on how the alternative axiological view of childhood supports the case against lowering the voting age. This case appeals to both the prudential interests of individual children and to a collective interest that we all have in allowing children to enjoy a 1 For an elaborate presentation of the views philosophers have traditionally held about children, see Matthews and Mullin (2018). Anca Gheaus, ‘Let Them Be Children?’. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Anca Gheaus (2023). DOI: 10.1093/oso/9780192894090.003.0009

116

‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

longer period of protection from the responsibilities, political knowledge, and other burdens that are characteristic of adulthood. Not only are the arguments explored in the chapter merely pro tanto but also their importance in the balance of reasons is likely to depend on the background justice of particular societies. The fourth section explains why (older) children’s interest in voting is more weighty if they live in a society in which many adults fail to comply with their duties of justice, and in particular with what they—the adults—owe to future generations. The fifth section concludes.

2. Two conceptions of childhood By ‘childhood’, in this chapter, I refer to roughly the first two decades of people’s lives, thus also counting teenagers as children (for simplicity). I assume that this is the time during which we mature biologically, emotionally, intellectually, and socially. Children, by this definition, are not fully mature in these ways. Children’s different kinds of immaturity have always been considered normatively relevant: this is the mere deficiency view. Their emotional, intellectual, and social immaturity bears on the content of their rights, assuming, as I do, that they are rights bearers. It does so, first, by making them subject to being ruled over by adults whose authority over children is justified by appeal to the children’s own well-being (Gheaus 2021; Møller Lyngby-Pedersen: Chapter 3 in this volume). This is because children lack the agential powers required for self-government, including the liberty right to significantly set back their own interests. Second, due to their immaturity, children are also widely believed to be incompetent in important respects that are relevant for the acquisition of rights the exercise of which imposes risks on others: the right to vote is such an example. So much is widely assumed by both axiological views of childhood. The difference between them lies in their identifications of unique sources of value and disvalue during childhood. The first view affirms that (i) there are special bads of childhood and rejects (ii) that there are special goods of childhood. The second view affirms (ii) (and, in some versions, rejects (i), but I will leave this complication aside here).

2.1 First view: ‘Childhood is worse’ In addition to bearing on children’s rights, children’s immaturity is also sometimes seen as a distinctive source of disvalue for children—a feature that makes being a child worse, other things being equal, than being an adult. Sarah Hannan (2018) notes how a child’s desires and plans are subject to legitimate frustration by children’s guardians and deems this to be, in one way, bad for the child. Patrick Tomlin (2018) also thinks that having silly or inconsistent desires is a source of ill-being for the bearers. If some

Ageing without Ageism?

117

kind of desire satisfaction is the correct view of well-being, then childhood seems worse than adulthood. Both Hannan and Tomlin also note that children are subject to domination by adults. While such domination is—again—justified by children’s underdeveloped agential powers, it is, in one way, bad for the child. Further, according to Hannan (2018), it is in itself bad for children to be asymmetrically physically and emotionally vulnerable to adults. Perhaps due to children’s limited capacities and agency, childhood was seen, until recently, as a stage of life whose value consisted mostly in preparing individuals for adulthood. Some philosophers explicitly deny that achievements during childhood contribute to the success of one’s life (Slote 1983) and Kantians tend to see childhood as a predicament that is best overcome (Schapiro 1999). In short: being childish is prudentially bad, and since children are unavoidably childish, childhood is, other things being equal, worse than adulthood.

2.2 Second view: ‘Life without childhood is worse’ Recently, however, several philosophers have challenged the claim that childhood is merely good as preparation for adulthood. They note that there are some goods which display the following combination of characteristics: (i) children have unique, or privileged, access to them; (ii) they make a very significant contribution to a person’s lifetime well-being; and (iii) this contribution is unique, in the sense that no other goods can make the contribution in question. Call these the ‘special goods of childhood’. Examples include a fully trusty kind of intimacy between children and their caregivers which is not to be encountered in relationships between adults, and sexual innocence (Brighouse and Swift 2014); a sense of one’s future as limitless (Brennan 2014); playfulness and carefreeness (Gheaus 2015b; Feraciolli 2020); unusually high levels of openness to experimentation (Franklin-Hall 2013) and a vivid imagination, allowing children to lead rich fantasy lives (Macleod 2015); and heightened levels of artistic creativity and ability to ask philosophical questions (Gheaus 2015b; Matthews and Mullin 2018). Not everybody identifies all the above examples as special goods of childhood, but there is general consensus that one particular good, unstructured free time, is itself such a good. This is in part because of its developmental value for children. But, above and beyond what they need for optimal development, children also need enough free time if they are to derive the full value of their capacity for being carefree and to make proper use of their unusual fantasy, creativity, and ability to explore the world and experiment. The distinctive feature of the second conception of childhood is that these goods are sufficiently valuable to outweigh any unavoidable ‘bads’ of childhood that stem from children’s immaturity. In short, being childlike is prudentially good, and since we can only be fully childlike during childhood, because the psychological and biological immaturity are important determinants of our ability to be childlike, a life

118

‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

without childhood with sufficient access to the special goods would be worse.2 Which (particular version) of these views of childhood is correct matters for the specification of children’s claims of justice. If missing out on these goods during childhood cannot be (fully) compensated during adulthood, which is what those defending the second conception of childhood seem to assume, then children are owed the enjoyment of the special goods of childhood. There are some reasons that support the belief that the goods I have listed above can only be enjoyed, or fully enjoyed, at an early age. First, at least some of these goods are made possible by children’s biological immaturity: the incomplete development of prefrontal cortices explains their creativity, curiosity, and imagination (Gopnick 2009). Second, access to some of the above goods seems to depend on being new to the world (Gheaus 2015b): a sense of time as limitless, carefreeness, and openness to experimentation. Finally, the full enjoyment of most of these goods appears incompatible with leading a productive, orderly, life; if so, it is possible that, even if adults were, in principle, capable of enjoying them to some extent, their enjoyment could be limited by adults’ duties (Gheaus 2015a) or else interfere with specific sources of value during adulthood, namely, the kinds of accomplishments that require long-term, self-disciplined engagement in goal-orientated activities. Perhaps there are good replies to all these arguments, at least with respect to some of the above goods. If so, to the extent to which it turned out to be both possible and desirable to make the goods available to adults, this would detract from the case that we owe them specifically to children. Assume that children are owed the special goods of childhood. This is directly relevant to arguments about age limits in policymaking. Section 3 illustrates this claim by looking, specifically, at voting.

3. Age limits in voting Children in general are denied the right to vote. Yet, it is not trivial to show that a blanket disenfranchisement of children is permissible. It is even harder to show that it is morally required. Alternatives to the status quo include lowering the voting age—by a few years, to 16 or, indeed, by many more years—and determining which children have the vote on a case-by-case basis.

3.1 Resources from the ‘childhood as mere deficiency’ view The argument most frequently invoked in favour of children’s disenfranchisement is their incompetence, stemming from their incompletely developed rational agency 2 If childhood is better than adulthood, it doesn’t necessarily follow that the longer our childhood, the better because it may be better for individuals, other things being equal, to enjoy a variety of sources of value (Gheaus 2015b).

Ageing without Ageism?

119

(Chan and Clayton 2006; Beckman 2018). According to Robert Dahl, for instance, children should not be allowed to vote because ‘they are not yet fully competent’ (Dahl 1991: 127). Their lack of competence (for instance, their incapacity to form moral judgements) means that excluding them from the vote does not wrong them by failing to treat them as moral equals since they are not, in this particular respect, adults’ moral equals (Christiano 2008). And, since voting is a form of exercising power over others, allowing incompetent individuals to vote creates negative externalities. Children’s lack of competence due to immaturity can feature in both views of childhood, but the discussion of their right to vote has, by and large, assumed that when it comes to voting rights, the case for children’s disenfranchisement is on a par with that of intellectually disabled individuals (Lopez-Guerra 2012). This fact in itself may not be enough to attribute to those writing about age limits in voting the view that ‘childhood is worse’ (at least, some authors may also reject a ‘disability is worse’ view). But nor do these authors move beyond the mere deficiency view of childhood: for instance, with one exception, they don’t discuss the significance of the opportunity costs that children would have to pay if they were enfranchised. It is particularly telling that, when considering the consequences that giving children the vote would have for children themselves, the focus is almost always exclusively on educational results: teaching children to take responsibility for political participation (Lopez-Guerra 2012: 120). Thus, a frequently invoked reason in favour of enfranchising children is how voting would impact on them qua future adults by cultivating their interest in politics and instructing them about electoral processes. Indeed, on the mere deficiency account of childhood, some of the most relevant considerations in favour of lowering the voting concern the ways in which early voting would prepare them for adulthood. The problem with the argument from children’s incompetence is that competence levels vary significantly, both amongst adults and amongst children. If we deny children the vote on grounds of incompetence, should we then also withhold the franchise form incompetent adults, and, if so, is this conclusion compatible with political equality between adults (Fowler 2014)? Moreover, a minority of children do possess the relevant capacities to the same extent as average adults and to a higher extent that many of the adults who have the vote; this fact seems particularly relevant to children who are close to maturity—for instance, those above age 16. Excluding these children from the vote appears unjustified (Schapiro 1999; Mraz 2020). On one influential account, democracy has non-instrumental value: the one-person–one-vote rule is constitutive of people standing in equal social relations, which is valuable in itself (Kolodny 2014). In this case, denying the vote to those who are minimally competent to exercise it amounts to a violation of a right they have on grounds of their moral status, and it involves a serious expressive harm. Several philosophers tried to provide an explanation of why we may disenfranchise children while enfranchising adults whose cognitive capacities are similar to those of typical children (Fowler 2014; Mraz 2020). But even the most sophisticated arguments struggle to justify the disenfranchisement of the minority of children who have outstanding rational and cognitive capacities.

120

‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

To explain why it is permissible (or maybe even required) to exclude even unusually competent children from the vote, one may appeal to the ‘not their voices’ argument: perhaps children’s views are not attributable to them in a way that would make their exclusion from the franchise disrespectful. Tamar Schapiro has described children’s lack of mature agential powers as a ‘predicament’ (Schapiro 1999). On her influential account, a child3 is a person who has not yet acquired a stable practical identity—that is, a stable set of first- and second-order preferences. Children’s limited ability to reflect on their desires and goals, she believes, makes even those who are outliers in terms of competence (i.e. the unusually competent children) normatively disabled because their will is not fully theirs—or, as Norvin Richards (2010) put it, they don’t yet speak in their own voice. Some authors who argue against lowering the voting age, such as Tak Wing Chan and Matthew Clayton (2006), do, in fact, rely in their justification on the claim that even the judgements of 16-year-olds are insufficiently stable, without explicitly endorsing the ‘not their voices’ view. If so, even children who are on a par with adults with respect to their cognitive competency, and level of interest in, and knowledge about, politics may be legitimately denied the vote. If their values and preferences—which the vote would express—are not properly theirs, denying them the opportunity to let their views bear on the political process does not amount to an expressive harm. This argument, however, may run into the same difficulty as the previous one: some people under 18 have stable practical identities—they mature more quickly in this respect, too. Why is it justified to deny them the vote? Chan and Clayton (2006) insist that age is a fairly good proxy and point to the high costs of individual testing for relevant competencies and for stability of judgement; but appeal to costs alone is unlikely to convince those who defend democracy on non-instrumental grounds. Defenders of the ‘childhood is worse’ view, at least in principle, have the resources to elaborate on these arguments; they can note that it would be desirable to speed up the maturation of children’s agential capacities and the stabilization of their values and preferences, if that was possible without imposing on them developmental costs (for instance, without impairing some aspects of their emotional or social development). If childhood is a predicament, as Schapiro (1999) puts it, and if being its subject is bad for children, as Hannan (2018) argues, then adults have decisive reason to help children to grow up quicker. If so, perhaps we owe it to all children to help them reach maturity earlier, which would then also justify lowering the voting age, and it is possible that granting them the right to vote could itself contribute to speeding up children’s maturation. Those who endorse this axiological conception of childhood and who are also ‘complete lives’ egalitarians (McKerlie 2012) should particularly welcome the ‘acceleration’ of children’s intellectual, emotional, and social maturation. Justice, they believe, is concerned with how well off individuals are relative to each other over the course of their entire lives; if so, people who die early tend

3 Unlike this chapter, Schapiro uses ‘child’ as a status, not as a biological concept: anybody who has not yet developed a stable practical identity. Her choice of the word ‘child’, however, suggests that age is an accurate proxy for identifying the extension of the concept.

Ageing without Ageism?

121

to be at a disadvantage. If, in addition, childhood has little non-preparatory value, it seems that the sooner it ends, the better for realizing equality between individuals’ whole lives. It is this claim that defenders of the alternative conception of childhood will squarely oppose.

3.2 The ‘Let them be children!’ reasons against lowering the voting age I don’t know whether it is at all possible to accelerate children’s maturation, let alone without negative side effects on their future emotional and social well-being.⁴ But even if this was possible, it would be wrong, according to the ‘life without childhood is worse’ view, because it would deprive individuals of their best, or only, chance, to enjoy the special goods of childhood. This view also provides distinctive reasons for holding on to age as a criterion for having the right to vote, in spite of some children’s high cognitive and rational capacities—and, thus, can offer a solution to the difficulties of justifying the disenfranchisement of these children. If, as children, we have unique access to important goods, and missing out on them makes our lives go worse overall, we also have an interest in the conditions necessary to enjoy these goods. This means that one must look beyond the consequences that children’s enfranchisement would have for their future, adult, self, at the likely noninstrumental impact on one’s childhood of being enfranchised. The only author who, as far as I know, has done this is Ludvig Beckman (2009), who notes that an important consideration against lowering the voting age comes from the burdens of responsibility incurred by having the franchise. His opposition to children’s enfranchisement is rooted in the worry that a right to vote comes with specific responsibilities and that children’s responsible exercise of the right to vote would limit their play time, thus setting back important interests of theirs. To this, a defender of the second conception of childhood may add that, alongside play, other important goods might be crowded out if children were to invest enough time in learning and thinking about politics in order to exercise their right to vote responsibly. Exploration of the world and of oneself, and engaging in creative endeavours, for instance, also take time. This line of reasoning can be strengthened by the further assumption that the value of engaging in certain activities, including political activities and therefore also voting, depend on the agent performing the activity in question well enough. Perhaps exercising a right to vote, if done well enough, crowds out the time necessary for enjoying the special goods of childhood—and, if so, this may be a good reason to deny the franchise to children. It is not clear that this argument settles the debate: perhaps, as critics noted, we could encourage children to limit their political participation in ways that is not ⁴ There is reason to be sceptical. The very early enrolment of children (at age four) in academically centred education in the United Kingdom has been found to backfire against their overall development (Alexander et al. 2010).

122

‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

contrary to their interest in free time (Lopez-Guerra 2012: 30). Still, readers who are receptive to Beckman’s point may worry that it is not possible to exercise the vote responsibly without getting interested in politics, and hence learning about the world to a significant extent, which, in turn, involves becoming burdened with concerns that will unavoidably spoil children’s general carefree attitudes towards the world. One interpretation is that children who become politically active and informed will rightly feel responsible not only for the manner in which they exercise their political right but also for the outcomes of this exercise and, as Beckman puts it, that ‘playfulness is suppressed by the burdens of responsibility’ (2009: 115). Another, even more plausible interpretation, is that knowledge about the world sufficient to afford children responsibility to exercise their right to vote will simply tamper with their carefree attitudes. Not only time but also a carefree attitude may be a precondition to children’s full enjoyment of the special goods of childhood and, some have argued (Feraciolli 2020), something to which children have a right. It is then possible that an interest in children’s well-being qua children, supported by the ‘life without childhood is worse’ view, could bear significantly on the debate about children’s disenfranchisement in a way that would bring out the difference from the case of mentally disabled individuals. If, as I noted above, it is difficult to establish a permission, let alone a moral requirement, to disenfranchise outstanding children merely by pointing to children’s competence or unstable practical identity, a concern with their access to the special goods of childhood could make the difference. Even if life without childhood would be worse, a full account of lifetime well-being would need a separate argument to determine the ideal length of childhood. It is possible that optimal enjoyment of the special goods of childhood require either less or more time than 18 years. This conception of childhood therefore doesn’t, in itself, speak against some degree of childhood acceleration—or, of course, against extending childhood. And it is compatible with lowering the voting age if we need less than 18 years to enjoy the special goods of childhood. Moreover, the claim that children should have sufficient years of protected time during which they can enjoy the special goods of childhood can appeal to two different arguments, and each of them may indicate as optimal a different number of childhood years. The first is the argument on which I have relied so far in this chapter, appealing to what is prudentially good for individuals. The second argument appeals to the common good: according to some developmental psychologists, human beings’ relatively long childhoods have important evolutionary advantages. Slow maturation is worth the high cost that we pay for it as adults (when we need to sustain high levels of parental investment) because it ensures behavioural flexibility during adulthood— flexibility which is, itself, necessary for our species’ successful adaptation to new circumstances (Bjorklund 2007; Gopnik 2016). Perhaps unsurprisingly, some of the same features of childhood that are seen as distinctive sources of value by the second conception of childhood are also believed to generate collective goods during adulthood: for instance, play in childhood has been said to promote social competence in adulthood (Bjorklund 2007). Assuming that some views of this form are correct, they indicate that we all have a powerful interest not to rush children through childhood.

Ageing without Ageism?

123

More generally, they are reasons to grant children sufficient unstructured time; to the extent to which responsible exercise of the right to vote—or, more generally, responsible political participation—is incompatible with having sufficient unstructured time, this is a reason for not involving children too heavily in the world of collective self-governing.

4. Children’s political participation in non-ideal circumstances Considerations of non-ideal circumstances generate an additional layer of complication in this debate. Even if, in relatively just societies, there was enough reason against the enfranchisement of (even outstanding) children, on grounds of their interest in having enough unstructured time and being carefree, such reasons may be overridden, in sufficiently unjust circumstances, by considerations about the prudential and moral value of enfranchising children. Even if political activity leaves children less well off in one respect—namely, with respect to their enjoyment of the special goods of childhood—it can make them better off in other respects. It is possible that, when many adults fail to comply with the demands of justice, children’s political participation, including a right to vote, could be instrumental to protecting their own stakes in inheriting a more just society; in particular, it would force the political class to pay more attention to their interests. This consideration might be sufficient to show that children have a right to some forms of political participation and, when they reach a modicum threshold of competence, a right to vote. This seems to apply especially when children themselves have high prudential stakes in addressing existing injustices—if, for instance, adults fail in their most stringent duties of intergenerational justice, and children’s participation is likely to improve policies that will affect children throughout the course of their lives. Tim Fowler (n.d.), for instance, notes the existence of political movements that rely significantly on children’s participation, such as ‘Fridays for Future’, which aims to raise support for just climate policies. He is advancing an argument similar to the one I explore here but aimed at defending children’s right to take part in political protest (rather than children’s enfranchisement): doing this would come at a cost, which he identifies as ‘innocence’, but the cost is, all things considered, worth paying since so much is at stake for current children when climate policies are concerned. In non-ideal circumstances, too, one’s conception of childhood can make a difference for the identification of the reasons relevant to children’s voting: if there are intrinsic goods of childhood as defined here, deliberation about lowering the age limit in voting must register the trade-off involved in the failure to protect children’s access to free time and carefreeness. If, on the other hand, there are no such goods, any trade-offs would be easier to make: perhaps, in unjust societies, there is more reason than in just ones to support the costs of testing children for competency and granting the right to vote in an age-independent manner.

124

‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

If giving children the vote earlier would result in more just intergenerational policies, this is certainly a weighty argument for enfranchising them. There is a second benefit that one may think children’s involvement in politics under significantly unjust circumstances may carry for them—at least on a moralized conception of what makes life go better for individuals. To have the vote, but also to be able to take part in political movements, is to have an opportunity to stand up to injustice; and, perhaps, standing up to grave injustices is, in itself, good for individuals, especially when injustice is pervasive.⁵ Children who are capable of appreciating the wrongs of injustice might, on this ground, be owed some kind of access to the world of politics, access which may or may not amount to having the franchise. Unlike many other goods (including the ones discussed here, such as free time or creative pursuits), this good—standing up to injustice—is not satiable. If failing to stand up to an injustice is bad for you, then it is good for you to stand up to injustice on every occasion. But this means that, in sufficiently unjust circumstances, if the balance of reason inclines towards granting older children rights to political participation on grounds that this is good for them, the same balance of reasons may point towards encouraging them to devote a great deal of their time to politics. The opportunity costs for children of being politically involved would then be higher in non-ideal than in just circumstances. It is particularly in such a context that it will matter whether we think that the goods foregone by children’s political involvement have special value. If the opportunity costs include forgoing goods impossible to compensate later in life, this is yet another illustration of how difficult it is to react well to injustice without thereby becoming oneself one of its victims.

5. Conclusion By examining the philosophical discussion about enfranchising children, this chapter illustrates how thinking about specific sources of value and disvalue in childhood matters for evaluating age limits in policymaking. The conclusions of this chapter generalize, at some level of abstraction, in two ways. First, different conceptions of childhood are likely to entail different judgements about how much time one would, ideally, spend as a child. It is not unthinkable that the length of childhood—defined as the time needed for our maturation—is itself, at least in part, influenced by social institutions. So policies may themselves determine, if only to a very limited extent, the normative relevance of age: the significance of being, say, 15, depends on whether or not 15-year-olds tend to be fully mature. Second, the method I adopt here generalizes because the specific goods and bads of childhood, or at least those that bear significantly on individuals’ well-being, matter to the specification of duties of justice towards children. Policies in which age limits seem, at least prima facie, normatively ⁵ An alternative basis for the line of reasoning I explore in this paragraph is appeal to the idea of a life well lived as understood by Ronald Dworkin (2011).

Ageing without Ageism?

125

relevant—the allocation of health care, labour, and marriage regulations, as well as the content of school curricula—are likely to make a difference to the discharging of these duties. I end with an observation that applies across several such policies: if children are entitled to a number of goods on grounds that their provision during childhood is most conductive to overall well-being, this provides a strong pro tanto reason to deny them certain freedoms when the exercise of these freedoms is incompatible with the enjoyment of the goods in question.⁶ Examples include not only the right to vote but also to marry or to be gainfully employed. Indeed, I hope that future debates about children’s civil and economic rights will make more explicit the ways in which particular conceptions of childhood inform the reasons at stake.

Acknowledgements I am grateful to the editors for feedback on a previous draft of this chapter.

References Alexander, Robin. 2010. Children, Their World, Their Education. Final Report and Recommendations of the Cambridge Primary Review. London: Routledge. Beckman, Ludvig. 2009. The Frontiers of Democracy. New York: Palgrave. Beckman, Ludvig. 2018. ‘Children and the Right to Vote’. In The Routledge Handbook of the Philosophy of Childhood and Children, ed. Anca Gheaus, Gideon Calder, and Jurgen De Wispelaere. London/New York: Routledge, pp. 384–394. Bjorklund, David. 2007. Why Youth Is Not Wasted on the Young: Immaturity in Human Development. Malden: Blackwell Publishing. Brennan, Samantha. 2014. ‘The Goods of Childhood and Children’s Rights’. In FamilyMaking: Contemporary Ethical Challenges, ed. Francoise Baylis and Carolyn McLeod. Oxford: Oxford University Press, pp. 29–48. Brighouse, Harry, and Adam Swift. 2014. Family Values: The Ethics of Parent–Child Relationships. Princeton, NJ: Princeton University Press. Chan, Tak Wing, and Matthew Clayton. 2006. ‘Should the Voting Age Be Lowered to Sixteen? Normative and Empirical Considerations’. Political Studies 54 (3): 533–558. Christiano, Thomas. 2008. The Constitution of Equality: Democratic Authority and Its Limits. Oxford: Oxford University Press. Dahl, Robert. 1991. Democracy and Its Critics. New Haven, CT: Yale University Press. ⁶ This is similar to contexts involving adults, when a prohibition to do x may be justified by claiming that it represents the best way to protect the right not to x. For an attempt to justify a prohibition to work right after one has given birth in this way, see Gosseries (2014).

126

‘Let Them Be Children?’: Age Limits in Voting and Conceptions of Childhood

Dworkin, Ronald. 2011. Justice for Hedgehogs. Cambridge, MA: Harvard University Press. Feraciolli, Luara. 2020. ‘Carefreeness and Children’s Wellbeing’. Journal of Applied Philosophy 37 (1): 103–117. Fowler, Tim. 2014. ‘The Status of Child Citizens’. Politics, Philosophy and Economics 13 (1): 93–113. Fowler, Tim. (n.d.). ‘The Minor’s Strike—Reflections on the Legitimacy and Limits of Children’s Political Activism’. Franklin-Hall, Andrew. 2013. ‘On Becoming an Adult: Autonomy and the Moral Relevance of Life’s Stages’. Philosophical Quarterly 63 (251): 223–247. Gheaus, Anca. 2015a. ‘The “Intrinsic Goods of Childhood” and the Just Society’. In The Nature of Children’s Well-Being: Theory and Practice, ed. Alexander Bagattini and Colin Macleod. Dordrecht: Springer, pp. 35–52. Gheaus, Anca. 2015b. ‘Unfinished Adults and Defective Children’. Journal of Ethics and Social Philosophy 9 (1): 1–22. Gheaus, Anca. 2021. ‘The Best Available Parent’. Ethics 131 (3): 431–459. Gopnik, Alison. 2009. The Philosophical Baby: What Children’s Minds Tell Us about Truth, Love, and the Meaning of Life. New York: Farrar, Straus and Giroux. Gopnik, Alison. 2016. The Gardener and the Carpenter: What the New Science of Child Development Tells Us about the Relationship between Parents and Children. New York: Farrar, Straus and Giroux. Gosseries, Axel. 2014. ‘Repos postnatal et vote obligatoire’. La Libre Belgique—Supplement Libre entreprise. https://www.lalibre.be/archives-journal/2014/05/03/repos-postnatalet-vote-obligatoire-NA7BJ3QEZZA3PI3KVXBBYACRSQ/ Hannan, Sarah. 2018. ‘Why Childhood Is Bad for Children’. Journal of Applied Philosophy 35 (S1): 11–28. Kolodny, Niko. 2014. ‘Rule over None II: Social Equality and the Justification of Democracy’. Philosophy and Public Affairs 42: 287–336. Lopez-Guerra, Claudio. 2012. ‘Enfranchising Minors and the Mentally Impaired’. Social Theory and Practice 38 (1): 115–138. Macleod, Colin. 2015. ‘Agency, Authority, and the Vulnerability of Children’. In The Nature of Children’s Well-Being, ed. Alexander Bagattini and Colin Macleod. New York: Springer, pp. 53–64. Matthews, Gareth, and Amy Mullin. 2018. ‘The Philosophy of Childhood’. Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/childhood. McKerlie, Dennis. 2012. Justice between the Young and the Old. Oxford: Oxford University Press. Mraz, Attila. 2020. ‘Disenfranchisement and the Capacity/Equality Puzzle: Why Disenfranchise Children but Not Adults Living with Cognitive Disabilities’. Moral Philosophy and Politics 7 (2): 255–279. Richards, Norvin. 2010. The Ethics of Parenthood. Oxford: Oxford University Press.

Ageing without Ageism?

127

Schapiro, Tamar. 1999. ‘What Is a Child?’ Ethics 109: 715–738. Slote, Michael. 1983. Goods and Virtues. Oxford: Clarendon Press. Tomlin, Patrick. 2018. ‘The Value of Childhood’. In The Routledge Handbook of the Philosophy of Childhood and Children, ed. Anca Gheaus, Gideon Calder, and Jurgen De Wispelaere. London/New York: Routledge, pp. 79–89.

9 Age and the Voting–Driving Analogy Alexandru Volacu

1. Introduction All democratic systems impose some age threshold below which citizens of the polity do not have the right to vote. This practice has been the subject of some academic debate in recent years.1 In contrast, the question of whether we should also impose a ceiling above which voting rights would have to be relinquished has been sparsely addressed by academics.2 In the broader public sphere, though, the idea that senior citizens should be disenfranchised has been more welcomed. For instance, a 2019 UK poll showed that 32 per cent of the respondents broadly agreed with the proposal that citizens above the age of 70 should not be allowed to vote, at least in some cases, such as the Brexit referendum.3 Some of these responses are most likely driven by the fact that the Brexit issue is bound to have serious long-term consequences, which will affect senior citizens less than younger ones.⁴ But some people also believe that the voting rights of senior citizens should be curtailed by appealing to another argument, one that is not based on interests but on an analogy between the right to vote and other rights where old age might warrant restrictions. Here is an example involving the right to work: in developed countries it’s common for a mandatory retirement for judges at 70 or 75—this is because early dementia and cognitive decline are very difficult to detect on your own and the gravity of a judge’s decisions is unarguable. Yet the elderly can vote—and in numbers that potentially have enormous influence on election and referendum outcomes. Jacobs (2019) 1 See, for instance, Chan and Clayton (2006); Goodin and Lau (2011); Lau (2012); Lopez-Guerra (2012); Umbers (2020); Weinstock (2021); O’Neill (2022). 2 One exception is Van Parijs (1998), who is mainly concerned with the effects of ageing democratic societies on intergenerational justice. I have criticized this argument elsewhere for relying on problematic empirical premises and for having unattractive implications for enfranchisement when it is consistently applied (see Volacu 2021). 3 See https://www.drg.global/wp-content/uploads/Wings-tables-for-publication-291019.pdf. ⁴ Poama and Volacu (2021) and Gosseries (2022) explore variants of such an argument in more depth, arguing that even under the most favourable normative and empirical assumptions (which are, at best, highly controversial on their own), all that this can entail is a form of weighted voting—and one which may have significant operational difficulties, implausible implications when features other than age are considered, and is likely to conflict with other important values. Alexandru Volacu, Age and the Voting–Driving Analogy. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Alexandru Volacu (2023). DOI: 10.1093/oso/9780192894090.003.0010

Ageing without Ageism?

129

And here is another one, appealing to the right to drive: ‘When older people are deemed unfit to drive a car they have their licences taken away. Removing their right to vote is merely an extension of that safety-first community thinking.’⁵ While these latter views call for a blanket denial of some rights for senior citizens, it is possible to design less radical right differentiation policies in the case of voting. I will call any policy which meets this latter condition, an age-adjusted electoral policy⁶ and use the expression age-adjusted voting rights to denote—in particular—the different voting rights package granted to senior citizens when age-adjusted electoral policies are in place.⁷ In this chapter, I aim to articulate a plausible version of an analogical folk argument in favour of age-adjusted voting rights and examine whether it can be defended. Section 2 presents the analogical argument for age-adjusted voting rights. Section 3 outlines several relevant distinctions for assessing the analogy, primarily between forms of age-adjusting policies, on the one hand, and the aspects relevant for the justifiability of such policies, on the other. Section 4 tentatively defends the idea that age-adjusted driving rights are morally permissible due to the fact that senior citizens are more likely to exercise driving rights in a manner which produces harmful outcomes. Section 5 offers several objections to the analogical argument, maintaining that there are at least four relevant dissimilarities between driving and voting, which make age adjustments in voting rights untenable.

2. The voting–driving analogy An argument from analogy derives a conclusion from an analogical relation rather than an inductive or deductive one.⁸ A normative analogical argument can be cast in the following form:⁹ Base Premise: A situation is described in C1 . Derived Premise: A is plausibly drawn as an acceptable normative conclusion in case C1 . Similarity Premise: Case C1 is similar to case C2 in all normatively relevant respects. ∴ A is plausibly drawn as an acceptable normative conclusion in case C2 .

Here is an example of how an analogical argument works, based on Freiman’s (2014) defence of voting markets. According to Freiman, vote-buying is not essentially different from other practices that we consider acceptable in democratic polities. ⁵ See Chesterton (2016). I thank Andrei Poama for suggesting both examples. ⁶ In section 2, I discuss several forms which an age-adjusted electoral policy might take. ⁷ Thus, although children or some other age group could also have age-adjusted voting rights, in this chapter I solely focus on the group of senior citizens. ⁸ See Ribeiro (2014) for a comprehensive view on various theoretical and applied aspects of these types of arguments. ⁹ I draw on Walton’s (2014: 24) structure and notations.

130

Age and the Voting–Driving Analogy

Consider, for illustration, how the argument unfolds with regard to earmarking, that is, the practice whereby public funds are allocated to some specific project with the intent that the person or party who has influenced that allocation wins the electoral support of voters benefiting from the project. The Base Premise is the existence of this practice. The Derived Premise is that such a practice is not illegal in contemporary democracies. The Similarity Premise is that the essential features of earmarking (namely, offering ‘material incentives to persuade voters to support a candidate or policy on the basis of their economic self-interest’; Freiman 2014: 765) are shared by the practice of vote-buying. The Conclusion is that vote-buying should be legal. Notice that in order for this type of argument to work, its defender must undertake the difficult task of showing that the cases examined are, indeed, similar in all normatively relevant respects. This raises the question of why we should pursue analogical reasoning to begin with since defending the premises of deductive arguments might be less controversial. Still, even though deductive arguments might seem more powerful, analogical arguments can also play useful roles in moral reasoning.1⁰ First, analogical arguments are common in ordinary thinking. If we are interested in doing ‘engaged’ political philosophy (Wolff 2018) rather than simply applying normative principles to particular issues, it is important to engage with the actual beliefs, attitudes, and moral intuitions commonly held by people. The analogical argument set out below is rooted in a folk argument on voting rights allocation, which reasonable citizens can, at least at first sight, endorse. Second, analogical reasoning can function as a heuristic device which could be useful in uncovering the deeper normative features of a particular issue. For example, Lomasky and Brennan (2000) argue that voter abstention is not a case of free-riding since the normatively relevant features of abstaining are akin to those of refraining from joining a particular occupation (e.g. farming), and the latter is not an instance of free-riding. For the purpose of this chapter, it could be the case that thinking about the allocation of other rights (e.g. driving rights) will provide illuminating insights into the reasons for allocating voting rights as well.11 Finally, analogical arguments can ensure that public policies are designed in a consistent manner, treating like cases alike. This consistency worry drives, for instance, Lau’s (2012) analogical defence of child enfranchisement, which is based on the fact that senior citizens and children share the same cognitive features, and they should therefore be treated on a par from an electoral standpoint.12 Returning to the aim of this chapter, if voting rights and other types of rights (e.g. driving rights) are, indeed, relevantly analogous, they should also be treated on a par from the perspective of public policies.

1⁰ For a more general discussion on the functions of normative analogical arguments see, for instance, Spielthenner (2014) or Aikin and Talisse (2019: 124–132). 11 While I will argue that this is, in fact, not the case, something like this idea seems to underlie the folk argument in favour of age-adjusted voting rights. 12 See Volacu (2021) for a critical discussion of Lau’s argument.

Ageing without Ageism?

131

Let us now move to what I will call the argument from analogy for age-adjusted voting rights. The argument can be articulated as follows: Base Premise: Senior citizens are more likely to produce harmful outcomes through their exercise of driving rights. Derived Premise: Driving rights should be age-adjusted, with the aim of mitigating the likelihood that harmful outcomes will be produced. Similarity Premise: Senior citizens are also more likely to produce harmful outcomes through their exercise of voting rights, for the same reasons. ∴ Voting rights should be age-adjusted, with the aim of mitigating the likelihood that harmful outcomes will be produced.

3. Age-adjusted rights An age-adjusted policy is, in the present sense, one which grants more limited rights to a certain age group (here, senior citizens). Age adjustments can take multiple forms. When it comes to adjusting electoral rights, a first relevant question is whether voting beyond a certain age should be either prohibited or subjected to extra requirements. The latter would not deny the right beyond a certain age but only demand more of senior citizens in order to gain access or continue to have access to that right. For example, this could amount to imposing an obligation of passing a political capacity test in order to cast a vote, exclusively for senior citizens. In the case of driving, this could amount to imposing a shorter time frame for licence renewal in the case of senior citizens, coupled with the requirement to pass a medical exam for each renewal.13 Call these types of limitations age adjustments in access. Second, we could ask whether senior citizens should be disenfranchised only on specific issues, for instance, on ballot proposals which will have long-term consequences, such as the Brexit referendum. Analogously, for the case of driving, senior citizens could be banned from driving but only during the night. More generally, this form of age adjustment would prohibit only a particular range of actions which could be performed by virtue of having the right, while leaving other actions equally possible between all groups. Call these types of limitations age adjustments in scope. Third, we could also ask whether the electoral weight of senior citizens should be reduced by introducing some form of plural voting.1⁴ More generally, this would amount to a case where an exercise of the right by senior citizens would be legally construed as less impactful than an identical exercise of the right by other adults.1⁵ Call these types of limitations age adjustments in weight. 13 I discuss these ideas in more detail in sections 4 and 5. 1⁴ See Van Parijs (1998); Poama and Volacu (2021); and Gosseries (2022) for examples of such a scheme. 1⁵ I can think of no instance where the right to drive could be adjusted in this way since exercising this right has an immediate impact, as opposed to exercising the right to vote, where the impact is arrived at only following a socially designed aggregation process.

132

Age and the Voting–Driving Analogy

Surely, there can be a wide range of justifications for rights adjustments of the types outlined above. In the present chapter, I am only interested with one account, namely, adjusting rights due to harm-based considerations.1⁶ A plausible view is that adjustments of rights should track the likelihood that a particular group will exercise the respective rights in a manner which will lead to more harmful outcomes for others. Evaluating this will require us to assess group features and the empirical evidence available, in order to see whether there is a significant difference between that group and others when it comes to producing harmful outcomes through the exercise of the right. If there is a significant difference, we have stronger reasons to adjust the right accordingly. In addition, we will need to assess the features of the harm at stake, that is, whether it can be assessed in the first place and whether it can take forms that are severe enough. In some cases, the exercise of a right can lead to consequences which are unambiguously harmful, for instance, an avoidable physical injury. In other cases, however, the consequences are not as easy to assess. In such cases, we have weaker reasons to adjust the respective right, just as we have weaker reasons to adjust it when the harms are not particularly severe. Furthermore, the features of the right in which we are interested must also be examined in order to see whether the harms in question are neatly traceable to the actions performed by an agent by virtue of having the respective right. In many cases, the unilateral action of an agent is sufficient for the production of harmful outcomes. But sometimes, harmful outcomes can only be produced when other people also perform the respective action and/or other people abstain from performing countervailing ones. In these latter cases, we have weaker reasons to adjust the respective right since the harm in question is more readily avoidable. Finally, even if we conclude that the exercise of a right by a particular group will indeed produce more harmful outcomes, it does not automatically follow that the right should be adjusted. Since having rights is one mechanism through which our interests are protected, the justification of rights adjustments also hinges on how weighty the underlying protected interest is for the groups in question and how much of a negative impact the adjustment itself will have on these interests. Thus, adjusting a right for a particular group may come with significant moral costs in terms of unfairly setting back those interests and, to the extent to which this is the case, we have weaker reasons to adjust the respective right. Therefore, the justification of a policy of age-adjusted rights, as well as the form which the adjustment would take, is a function of four considerations: (i) the magnitude of inter-group differences, (ii) harm severity and assessibility, (iii) the right exercise-harm nexus, and (iv) the moral costs of adjustment. I will now show that there are substantive differences between driving and voting along those lines and explore the implication of such differences. 1⁶ Leaving others, such as those based on interests or on comprehensive theories of justice, aside.

Ageing without Ageism?

133

4. The case for age-adjusted driving rights The second and third aspects (i.e. harm severity and assessibility and the right exercise-harm nexus) are fairly straightforward for the case of driving. The harmful outcomes which exercising driving rights lead to are clearly assessible, consisting primarily in physical injuries (including fatal ones) due to accidents and secondarily in the economic costs associated with damaged property. These harms can also be potentially severe. Moreover, singular improper exercises of driving rights are sufficient for these harmful consequences to be produced. Of course, not all accidents occur due to improper uses of driving rights, but many can be traced to the cognitive state of the driver. For instance, driving while under the influence of alcohol or under conditions of sleep deprivation increases the likelihood that the exercise of the driving right will have harmful consequences. This is the reason why numerous countries have stringent legal prohibitions against driving when the driver’s cognitive state is affected in this manner. The more controversial question, then, is whether the cognitive state of senior citizens is also affected in a manner which makes them more likely to produce harmful consequences when driving—which links to the first aspect identified above. The empirical evidence strongly supports the fact that as we age we do experience some forms of cognitive decline. The distinction between crystallized and fluid intelligence is relevant here. General knowledge, our vocabulary, and the type of skills that we acquire through experience, are part of crystallized intelligence. These skills are unaffected or still improving by the sixth and seventh decade of our lifetime. However, other skills, such as processing speed, attention, memory, visuospatial abilities, and executive functioning, which are part of fluid intelligence, are declining with age. And it is this particular type of cognitive decline that does affect the driving abilities of senior citizens: in many cases this is due to mild cognitive impairment (MCI) or dementia, other neurologic or musculoskeletal disorders, medical illness, vision problems, or medications. Unfortunately, even older adults who manage to avoid all of these challenges may still become unsafe drivers because of normal cognitive aging, which can cause small decrements in the multiple cognitive domains needed for driving. (Harada et al. 2013: 745)¹⁷

Drawing on these physiological circumstances, it is reasonable to expect that senior citizens would be, on average, more dangerous drivers. The empirical literature on

1⁷ See also Anstey et al. (2005), who provide meta-analytical evidence that largely points in the same direction, with processing speed, visual attention, executive functioning, short-term memory, visual acuity, contrast sensitivity, and physical function playing essential role in driving abilities.

134

Age and the Voting–Driving Analogy

the safety of elderly driving provides some support for this assumption, with a rough consensus existing over the following matters (Braver 2004): (a) Per licensed driver, senior citizens have a higher incidence of fatal crashes, which continuously increases above the age of 70. For this latter category, the rate of fatal crashes is similar to that of the 16–29 age group. (b) Per unit of travel, senior citizens have a higher incidence of overall crashes than the 30–69 age group but a rate which is similar to that of the 16–29 age group.1⁸ (c) Per unit of travel, senior citizens have a higher incidence of fatal crashes than other adults, except for young drivers, and above the age of 80, senior citizens have a much higher rate of fatal crashes than any other age group. (d) Elderly drivers, particularly above the age of 75, have a higher death rate when being involved in car crashes than all other age groups, but accidents involving them have a lower death rate for passengers of other cars and for pedestrians than for all other age groups. Hence, senior drivers are, indeed, more likely to cause harmful consequences for both themselves and for others. While this only partially holds when comparing senior citizens with very young adults, this does not undermine the argument as long as appropriate adjustments would also be prescribed against the group of very young adults. Finally, with regard to the moral costs of an age-adjusted driving policy, our fourth dimension above, it is true that the right to drive protects a weighty interest of individuals since the enhanced mobility which driving offers has both intrinsic and instrumental value. As Tripodis remarks: driver’s license holds special meaning for older persons. It serves as a symbol of freedom, independence and self-sufficiency. It also reinforces one’s identity as a functioning and socially capable adult [. . .] Older people rely on their cars to go to the grocery store, doctors’ appointments, places of worship, or family and friends’ houses. Taking away their licenses may lead to depression and a decline in their overall health and well-being. (1997: 1052)

This would count against at least some forms of right adjustments, including a blanket prohibition on driving for seniors. However, more moderate forms of age adjustment in access could reasonably be enacted. For instance, in Illinois, drivers above the age of 81 have to renew their driving licences every other year and those above the age of 87 on a yearly basis, while other adult drivers are required to renew their license once every 4 years. 1⁸ Some authors (e.g. Langford et al. 2008) claim that comparing the relative rate of car crashes between age groups by using units of travel provides tainted results due to a ‘low-mileage bias’. Specifically, drivers who engage in this practice more frequently typically have a lower rate of car crashes per mile than those who engage in it less frequently. Since senior citizens drive, on average, with considerably less frequency than other age groups, the authors claim that we should largely attribute the increased rate of car crashes per unit of travel to this fact rather than to a decline in driving skill affecting the elderly.

Ageing without Ageism?

135

Furthermore, there is a ceiling of 75 years for renewal of a driving license by mail, and citizens above the age of 75 must also take a written and road test as part of the renewal process, alongside the visual test, which is required for every renewal, regardless of age.1⁹ More generally, in both the United States2⁰ and the European Union, about two-thirds of all states have some age-based differences in licensing policies between senior citizens and other adults. These and other types of more stringent requirements for senior citizens’ continued access to driving rights have been met with mixed evidence when it comes to their consequences for road safety.21 However, to the extent that such requirements indeed turn out to be effective, a solid harm-based argument can be provided in their defence.

5. Driving and voting rights: Four relevant dissimilarities Is the right to vote sufficiently similar in its essential features to the right to drive? If the answer is affirmative, it might be plausible to extend harm-based adjustments to the case of voting. Perhaps, it should not be blanket disenfranchisement. But some weaker form of age adjustment in access could turn out to be defensible. One example might be a test of political capacity for senior citizens above a certain age, using an instrument such as the Competence Assessment Tool for Voting (CAT-V). The test, designed by Appelbaum et al. (2005), seeks to evaluate political capacity on the basis of six variables: (i) understanding the nature of voting, (ii) understanding the effects of voting, (iii) choice, (iv) comparative reasoning, (v) generating consequences, and (vi) appreciation. I will draw on the four dimensions identified in section 3 to show that there are fundamental differences between driving and voting which make age adjustments (even in thinner forms) unattractive in the case of voting.

5.1 Cognitive type Activities such as driving rely heavily on fluid intelligence, which declines with age. Yet, the core aspects of voting are more closely related to the knowledge and experience accumulated throughout life (i.e. crystallized intelligence), which does not seem to decline with age unless the individual suffers from some specific cognitive conditions, such as dementia. An empirical study relevant here has been conducted by Irastorza et al. (2011). The authors use the CAT-V instrument, mentioned earlier, in the case of 2 groups of people over 70 years of age, one with no mental illness and the other suffering from various stages of dementia. The results obtained are 1⁹ See https://www.nolo.com/legal-encyclopedia/illinois-driving-laws-seniors-older-drivers.html. 2⁰ See https://www.iii.org/state-drivers-license-renewal-laws-including-requirements-for-olderdrivers. 21 See Siren and Haustein (2015: 69–73) for a meta-analysis.

136

Age and the Voting–Driving Analogy

suggestive: none of the subjects in the former group scored 0 (on a three-point scale consisting of 0, 1, and 2) on any of the 6 dimensions relevant for political capacity, while a majority of respondents in the latter group scored 0 for 4 of the 6 dimensions, and a plurality scored 0 for another one. This indicates that while a condition such as dementia does take a heavy toll on political capacity, old age in itself does not. Moreover, Lau and Redlawsk (2008) show that political knowledge actually peaks at advanced ages (around 70) on average. Though it registers a slow decline after this, even citizens at very advanced ages still outperform young adults when it comes to political knowledge.22 This gives credence to the idea that the type of cognitive decline associated with age does not negatively affect voting-relevant cognitive features in any profound way.23

5.2 Harm assessment We have uncovered the first dis-analogy. The reasons associated with cognitive and physical decline and invoked in the driving case do not apply to the voting case. Senior citizens retain high levels of political knowledge and sufficient decisionmaking abilities. But let us assume, for the sake of argument, that they would be characterized by significantly diminished political capacity. Would voting and driving be on a par then? This leads us to examine the harm dimension. Harmful exercises of voting are intuitively more difficult to pinpoint than for driving since they will necessarily rely on more controversial value judgements. One way to proceed is to start our inquiry from the idea of bad voting. Jason Brennan argues that people ‘vote badly when they vote without sufficient reason for harmful policies or candidates that are likely to enact harmful policies’ (Brennan 2009: 538). Bad voting, in his view, is the result of failing to meet basic moral and epistemic standards, such as voting premised on immoral beliefs, ignorance, or epistemic irrationality. We can equate bad voting with harmful exercises of the vote since, by voting badly, we are marginally increasing the likelihood that harmful policy outcomes will be produced. However, some people believe that we should reject the idea of bad or harmful voting altogether: in a democratic society no process-independent moral criteria can be referred to in order to settle what counts as a harmful, unjust or morally unjustified exercise of the right to vote, for voting is a device that is only called for precisely when citizens disagree on what counts as harmful, unjust and morally unjustified. (Gonzalez-Ricoy 2012: 50) 22 Especially, as Bramlett (2013) argues, when senior citizens have ample opportunities for social interaction with their peers—for example, when they live in aged communities. 23 Perhaps it could be argued, then, that a case based on cognitive capacity could actually be made for increasing the voting weight of senior citizens. But this is doubtful as well. In the same article, Lau and Redlawsk (2008) show that, beginning, on average, from the age of 65, individuals experience a gradual decline in their ability to determine which candidate best represents their political preferences. So political knowledge is only one facet of what we may term political capacity, while the ability to properly align electoral choices with political preferences is another, and an increase in one can be counterbalanced by a decrease in the other.

Ageing without Ageism?

137

Both Brennan and Gonzalez-Ricoy’s positions have merits and drawbacks. The former brushes aside many of the reasonable disagreements which characterize diverse societies and the fact that such disagreements are supposed to be solved through democratic means.2⁴ The latter fails to account for cases where voting may influence the enactment of policies that are harmful beyond reasonable moral disagreement. For instance, assume that we face only two options: voting for party A or B. The candidates of party A have historically supported policies that systematically suppress the vote of socio-economic and ethnic groups who usually vote for party B. Moreover, they campaign on a platform which seeks to uphold such policies, disguising their real motives behind a veil of empirically false claims about voter fraud. I take it that unjustified voter suppression is wrong in a morally uncontroversial way: it directly harms the citizens who cannot vote; it indirectly harms other citizens who would have benefited from the respective votes; it harms democracy; and it does so while unjustly blocking some citizens from having their say on ‘what counts as harmful, unjust and morally unjustified’. Furthermore, there is no countervailing value which would be achieved if voter suppression policies were to be maintained. Other things being equal, voting for party A in these circumstances, then, represents a harmful exercise of the vote. Even if the idea of voting harmfully can be made sense of, elections for political offices are bound to involve a complex range of issues and policies as well as genuine value conflicts, which makes assessing whether a vote is harmful deeply controversial. For instance, Van Parijs’s (1998) main worry concerning voting in rapidly ageing societies relies on the fact that senior citizens produce harmful consequences through their votes because they mainly choose on the basis of immediate self-interest and disregard issues of intergenerational justice. Whether this indeed qualifies as harmful voting, however, is not as straightforward as it might first seem since it also requires a convincing normative case in favour of the idea that the conflict between shorttermism and intergenerational justice in institutional design (Gonzalez-Ricoy and Gosseries 2016: 4–6) should always be settled in favour of the latter and should fall outside the purview of democratic practice. Because identifying instances of harmful voting is usually bound to be controversial, harm-based adjustments of voting rights would only be warranted in a handful of cases and building a policy on the basis of such cases appears to be rather impractical.

5.3 Contributory efficacy The third aspect relevant for age-adjusted rights policies (i.e. rights features) largely leads us to the same conclusions. As discussed in section 4, in the case of driving, it is sufficient for one person to exercise the right in a bad way (e.g. through reckless 2⁴ For instance, one of the examples to which Brennan (2011) sometimes turns in order to illustrate what bad voting might mean is that of voting for protectionist policies or candidates espousing such policies. However, the idea that protectionist policies would necessarily be harmful relies on a moral judgement which prioritizes one value (presumably, that of economic efficiency) over others, when, in fact, it is plausible to think that this underlying value conflict should be handled in a democratic manner.

138

Age and the Voting–Driving Analogy

driving) so that harmful consequences obtain either for that person or for others (or for both). But, in the case of voting, an individual exercise of the right is not sufficient to produce harmful outcomes since it needs to be accompanied by many other similar exercises. Thus, while harmful outcomes may very well be a consequence of the electoral process, they will come about through a collective process, rather than through isolated individual actions. This is important because, even if it could be shown that a higher percentage of senior citizens tend to vote more harmfully, this behaviour can be rendered harmless by the actions of other voters. As Goodin and Lau note: in voting, just as in contracting, it is not strictly necessary that each and every participant meets high standards of competence, so long as sufficiently many of them do. After all, no one voter acting alone enacts legislation. A great many other voters have, in effect, to ‘co-sign’ before a proposition becomes law. (2011: 159)

The joint authorship of the law allows us to enfranchise some less competent citizens (e.g. children or cognitively impaired individuals) without fearing that the electoral outcomes will be distorted by voter incompetence. But the same argument can be applied to the case of harmful voting, which can be cast both due to political incompetence or for other reasons (e.g. self-interest, immoral beliefs, etc.). Even if senior citizens would, on average, vote more harmfully than other adults, the latter can still counterbalance any negative effect of an unrestricted senior citizen vote. Of course, the larger the group of senior citizens, the less powerful this dissimilarity in contributory efficacy between the act of driving and the act of voting becomes. But it is worth noting that even in the most aged contemporary democratic polities, senior citizens (above the age of 65) still represent less than one-third of the eligible electorate. Therefore, at least under current circumstances, this dissimilarity remains quite relevant.

5.4 Moral cost of age adjustments Finally, assume that we were to grant that the analogy is plausible in spite of the normative and empirical difficulties mentioned so far. Even so, it is possible for there to be distinct reasons for rejecting age adjustment of voting rights that would not apply in the case of driving rights. First, let us recall that there are both symbolic and instrumental grounds which need to be factored in when making an all-things-considered judgement on whether to impose age-adjusted driving policies. Blanket prohibitions of driving rights above a certain age would certainly damage some of the vital interests of senior citizens, raising significant moral costs. But weaker forms of age adjustment, such as more frequent testing of the driving skills of senior citizens, will probably not be particularly detrimental to them. Symbolically, what the state is signalling through such

Ageing without Ageism?

139

policies is simply that, because of old age, the actions which supposedly occur during the exercise of these rights are more likely to be physically dangerous for senior citizens than for others, and the state should impose stricter regulations in order to better protect them. Instrumentally, the requirements of passing more frequent tests will probably not be overburdening, and if it is the case that the senior citizen who takes the test fails, denying them a renewed licence might end up to be instrumentally better for them as well. Extreme forms of age-adjusted voting rights, such as a blanket disenfranchisement, would also be profoundly damaging on symbolic and instrumental grounds. Symbolically, by denying the franchise to senior citizens, the state is signalling that they are no longer entitled to equal political standing within the community and that they have the status of—to put it somewhat figuratively—second-class citizens.2⁵ Instrumentally, by denying the franchise to senior citizens, the state deprives them of the most important political instrument for promoting their own welfare and that of their peers. First, because the preferences of senior citizens would no longer carry any political weight, democratically elected officials would have no clear rational incentive to pursue policies that would be to the benefit of this group. And second, because of the unique nature of the right to political participation as the ‘right of rights’ (Waldron 1998: 307), a failure to take into account the input of senior citizens in defining the scope and content of rights in a political community might deprive them of important legal protections. The key point is that, unlike in the case of driving, all of these considerations still apply even for weaker age-based adjustments of voting. Senior citizens would still be denied equal social status with other adult citizens and would end up having less political power, less political influence, and fewer means to protect themselves than they would have under conditions of full electoral equality. Thus, even if we could draw on the voting–driving analogy in order to ground a pro tanto case in favour of age-adjusted voting rights, there are distinctive symbolic and instrumental reasons which count against such a policy, even in its most moderate forms, but not against age-adjusted driving rights.

6. Conclusion In this chapter, I sought to contribute to the literature on the justification of policies which seek to restrict voting rights on age-based grounds. To this end, I charitably articulated and then examined a folk argument in favour of such policies, which builds on an analogy between the right to vote and other rights, such as the right to drive. My conclusion is that, even if we accept that driving rights could be ageadjusted for senior citizens, the move to impose age adjustment on voting rights on the same grounds fails. 2⁵ See Gosseries (2022: 164–165) for a similar point, but grounded in the value of self-respect.

140

Age and the Voting–Driving Analogy

Acknowledgments I thank Greg Bognar and Axel Gosseries for valuable feedback on several drafts of this chapter, which greatly improved the quality of the final version. This work was supported by the Grant Agency of the Czech Academy of Sciences through the project ‘Taking Age Discrimination Seriously’ (grant ID: 17–26629S) awarded to the Institute of State and Law of the Academy of Sciences of the Czech Republic, Centre for Law and Public Affairs (CeLAPA), created under subsidies for a long-term conceptual development (RVO: 68378122).

References Aikin, Scott, and Robert Talisse. 2019. Why We Argue (And How We Should). A Guide to Political Disagreement in an Age of Unreason (2nd edn). New York: Routledge. Anstey, Kaarin, Joanne Wood, Stephen Lord, and Janine Walker. 2005. ‘Cognitive, Sensory and Physical Factors Enabling Driving Safety in Older Adults’. Clinical Psychological Review 25 (1): 45–65. Appelbaum, Paul, Richard Bonnie, and Jason Karlawish. 2005. ‘The Capacity to Vote of Persons with Alzheimer’s Disease’. American Journal of Psychiatry 162 (11): 2094–2100. Bramlett, Brittany. 2013. ‘Aged Communities and Political Knowledge’. American Politics Research 41 (4): 674–698. Braver, Elisa. 2004. ‘Older Drivers and Motor Vehicle Crashes’. Hospital Medicine 65 (10): 599–604. Brennan, Jason. 2009. ‘Polluting the Polls: When Citizens Should Not Vote’. Australasian Journal of Philosophy 87 (4): 535–549. Brennan, Jason. 2011. The Ethics of Voting. Princeton, NJ: Princeton University Press. Chan, Tak Wing, and Matthew Clayton. 2006. ‘Should the Voting Age Be Lowered to Sixteen? Normative and Empirical Considerations’. Political Studies 54 (3): 533–558. Chesterton, George. 2016. ‘We Should Ban Old People from Voting’. GQ, 14 June. https:// www.gq-magazine.co.uk/article/eu-referendum-old-people-should-not-vote. Freiman, Christopher. 2014. ‘Vote Markets’. Australasian Journal of Philosophy 92 (4): 759–774. Gonzalez-Ricoy, Iñigo. 2012. ‘Depoliticising the Polls: Voting Abstention and Moral Disagreement’. Politics 32 (1): 45–51. Gonzalez-Ricoy, Iñigo, and Axel Gosseries. 2016. ‘Designing Institutions for Future Generations: An Introduction’. In Institutions for Future Generations, ed. Iñigo GonzalezRicoy and Axel Gosseries. Oxford: Oxford University Press, pp. 3–23. Goodin, Robert, and Joanne Lau. 2011. ‘Enfranchising Incompetents: Suretyship and the Joint Authorship of Laws’. Ratio (New Series) 24 (2): 154–166. Gosseries, Axel. 2022. ‘Should Old Age Votes Be Granted Less Weight?’ In Studies on the Boundary Problem in Democratic Theory, ed. Paul Bowman. Stockholm: Institute for Future Studies, pp. 149–168.

Ageing without Ageism?

141

Harada, Caroline, Marissa Natelson Love, and Kristen Triebel. 2013. ‘Normal Cognitive Aging’. Clinics in Geriatric Medicine 29 (4): 737–752. Irastorza, Luis Javier, Pablo Corujo, and Pilar Banuelos. 2011. ‘Capacity to Vote in Persons with Dementia and the Elderly’. International Journal of Alzheimer’s Disease, Article ID 941041. DOI: 10.4061/2011/941041. Jacobs, Joanne. 2019. ‘Old People Shouldn’t Vote’. Disruptorsco, 25 March. https://www. disruptorsco.com/tech-positive/2019/03/25/old-people-shouldn’t-vote. Langford, Jim, Sjaanie Koppel, Dennis McCarthy, and Sivaramakrishnan Srinivasan. 2008. ‘In Defence of the “Low-Mileage Bias”’. Accident Analysis and Prevention 40 (6): 1996–1999. Lau, Joanne. 2012. ‘Two Arguments for Child Enfranchisement’. Political Studies 60 (4): 860–876. Lau, Richard, and David Redlawsk. 2008. ‘Older but Wiser? Effects of Age on Political Cognition’. Journal of Politics 70 (1): 168–185. Lomasky, Loren, and Geoffrey Brennan. 2000. ‘Is There a Duty to Vote?’ Social Philosophy and Policy 17 (1): 62–86. Lopez-Guerra, Claudio. 2012. ‘Enfranchising Minors and the Mentally Impaired’. Social Theory and Practice 38 (1): 115–138. O’Neill, Martin. 2022. ‘Radical Democratic Inclusion: Why We Should Lower the Voting Age to 12’. Royal Institute of Philosophy Supplement 91: 185–212. Poama, Andrei, and Alexandru Volacu. 2021. ‘Too Old to Vote? A Democratic Analysis of Age-Weighted Voting’. European Journal of Political Theory OnlineFirst, DOI: 10.1177/14748851211062604. Ribeiro, Henrique Jales, ed. 2014. Systematic Approaches to Argument by Analogy. Cham: Springer. Siren, Anu, and Sonja Haustein. 2015. ‘Driving Licences and Medical Screening in Old Age: Review of Literature and European Licensing Policies’. Journal of Transport and Health 2 (1): 68–78. Spielthenner, Georg. 2014. ‘Analogical Reasoning in Ethics’. Ethical Theory and Moral Practice 17: 861–874. Tripodis, Vasiliki. 1997. ‘Licensing Policies for Older Drivers: Balancing Public Safety with Individual Mobility’. Boston College Law Review 38 (5): 1051–1088. Umbers, Lachlan Montgomery 2020. ‘Enfranchising the Youth’. Critical Review of International Social and Political Philosophy 23 (6): 732–755. Van Parijs, Philippe. 1998. ‘The Disfranchisement of the Elderly, and Other Attempts to Secure Intergenerational Justice’. Philosophy and Public Affairs 27 (4): 292–333. Volacu, Alexandru. 2021. ‘Justice, Symmetry, and Voting Rights Ceilings’. Theoria 87 (3): 643–658. Waldron, Jeremy. 1998. ‘Participation: The Right of Rights’. Proceedings of the Aristotelian Society (New Series) 98: 307–337. Walton, Douglas. 2014. ‘Argumentation Schemes for Argument from Analogy’. In Systematic Approaches to Argument by Analogy, ed. Henrique Jales Ribeiro. Cham: Springer, pp. 23–40.

142

Age and the Voting–Driving Analogy

Weinstock, Daniel. 2021. ‘What’s So Funny about Voting Rights for Children?’ Georgetown Journal of Law and Public Policy 18 (2): 751–771. Wolff, Jonathan. 2018. ‘Method in Philosophy and Public Policy. Applied Philosophy versus Engaged Philosophy’. In Routledge Handbook of Ethics and Public Policy, ed. Annabelle Lever and Andrei Poama. Abingdon: Routledge, pp. 13–24.

10 Empowering Future People by Empowering the Young? Tyler M. John

A number of recent writers have argued that the obligations of modern states to people who will exist in the future may far outstrip their obligations to their present citizens, given the vast number of people who will exist in the future and whose livelihoods depend on our actions (Beckstead 2013; Greaves and MacAskill 2021; Tarsney 2019; John 2022). And yet modern states do precious little on behalf of future generations, choosing to allow and incentivize destructive practices such as the widespread burning of fossil fuels while failing to take preventative measures that could deter global pandemics and other catastrophes. The state is plagued with problems of political short-termism: the excessive priority given to near-term benefits at the cost of future ones (González-Ricoy and Gosseries 2016b). By the accounts of many political scientists and economists, political leaders rarely look beyond the next two to five years and into the problems of the next decade. There are many reasons for this, from time preference (Frederick et al. 2002; Jacobs and Matthews 2012) to cognitive bias (Weber 2006; Johnson and Levin 2009; Caney 2016) to re-election incentives (Mayhew 1974; Tufte 1978; Arnold 1990; Binder 2006)1 but all involve foregoing costly action in the short term (e.g. increasing taxes, cutting benefits, imposing regulatory burdens) that would have larger moderate- to long-run benefits.2 Such behaviour fails not only the generations of people who are to come but also the large number of existing citizens who still have much of their lives left to lead. One type of mechanism for ameliorating political short-termism that receives much attention these days involves apportioning greater relative political influence to the young. As the story goes, younger citizens generally have greater additional life expectancy than older citizens, and it therefore looks reasonable to expect that they have preferences that are extended further into the future. If we apportion greater relative political influence to the young, it therefore seems that our political system as a whole will show greater concern for the future. In light of this story, a number of particular mechanisms have been proposed for apportioning greater relative political influence to the young, including lowering 1 For a contrary view, see Beck (1982). 2 For a general overview of the causes of short-termism and some mechanisms for ameliorating it, see recent work from Caney (2016); González-Ricoy and Gosseries (2016b); John and MacAskill (2021). Tyler M. John, Empowering Future People by Empowering the Young?. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Tyler M. John (2023). DOI: 10.1093/oso/9780192894090.003.0011

144

Empowering Future People by Empowering the Young?

the voting age (Piper 2020), weighting votes inversely with age (Van Parijs 1998; MacAskill 2019), disenfranchising the elderly (Van Parijs 1998), and instituting youth quotas in legislatures (Bidadanure 2016; MacKenzie 2016).3 In what follows, I argue that merely apportioning greater political power to the young is unlikely to make states significantly less short-termist, but underexplored age-based mechanisms may be more successful. In particular, states might mitigate short-termism by employing age-based surrogacy and liability incentives mechanisms within a deliberative body of young people charged with representing the young. In section I, I state precisely the argument for apportioning greater political power to the young on grounds of combating political short-termism. In section II, I argue that the extant empirical literature on the relationship between ageing and shorttermist policy support suggests that there is little relationship between the two, and so the argument for apportioning greater political power to the young fails. In section III, I identify age-based strategies which are better supported by existing political science research and advocate combining these strategies in a novel youth assembly.

1. Youth empowerment, efficiency, and justice According to a common view, a political system that is influenced by the elderly will tend to have a more short-term focus than a political system that evades or counteracts such influence (Van Parijs 1998). This is because the young can generally expect to live for a much longer period of time than the elderly and, so the argument goes, they will therefore generally have preferences that extend much further into the future. Younger people tend to have more remaining years of well-being, younger friends and families, and more personal goals in the years ahead of them. Thus, we may reasonably expect that younger people will be more concerned that the future goes well over longer time horizons. On the assumption that voters, policymakers, and other political actors are at least somewhat rational in acting on their preferences, we could infer that younger political actors will tend to support policies with a more long-term focus than older political actors. Thus, we should expect to find that political systems in which older age groups have more political influence are more short-termist. This common view provides the starting point for a powerful argument for increasing the influence of the young on politics. Short-termism has extremely deleterious effects on political decision-making, and so even any modest amelioration of political short-termism is a morally urgent priority. While it is somewhat difficult to measure precisely the harms of excessively prioritizing the near term, they appear to be substantial in aggregate. Hundreds of billions of US dollars are spent annually on global disaster relief despite studies finding regularly that investment in disaster preparation 3 Bidadanure (2016) does not accept this story and justifies youth quotas on other grounds.

Ageing without Ageism?

145

provides between 6 and 15 times as much benefit as the same size of investment in disaster relief (Healy and Malhotra 2009; Multihazard Mitigation Council 2017). The UN Office of Disaster Risk Reduction (2015) reports further that ‘an investment of $6 billion annually in disaster risk management would result in avoided losses of $360 billion over the next 15 years’. The US health-care system wastes between $88.6 billion and $111.1 billion each year by failing to adopt sufficient preventative care measures and instead adopting excessively reactive medical practices (Shrank et al. 2019). Net mitigation costs of global climate change, estimated at several hundred billion US dollars per year, increase, on average, by approximately 40 per cent for each decade of delay (US Council of Economic Advisers 2014). The failure of governments to adequately prepare for the COVID-19 pandemic—including by failing to accurately forecast its reproduction rate in 2020 and by failing to make sufficient investments in pandemic preparedness in years past—is estimated to cost the United States over $1 trillion per month along with millions of lives globally (Makridis and Hartley 2020). Certainly, failure to prevent future, more serious global catastrophes such as biological terrorism and nuclear war could cost us much more (Ord 2020). Short-termism has numerous causes and is exacerbated by global coordination problems and more. So no simple institutional fix will resolve all of these problems. And attending to short-term costs is clearly important. Global gross domestic product (GDP) trends indicate that our descendants will be wealthier than us, which may justify our borrowing some resources from the future to address near-term problems. Yet, the available indicators tell us that short-termism costs the global economy many billions and perhaps trillions of dollars annually and leads directly to millions of deaths from disasters and suboptimal spending. The exchange rate at which we are borrowing from our descendants and from our future selves clearly cannot be sustained. If we can identify institutional repairs to some of short-termism’s sources, we should doubtlessly pursue them.

1.1 Challenge of legitimacy The common view implies that it is both possible and desirable to ameliorate shorttermism by distributing greater political power to the young. This idea is based on utilitarian efficiency: all else being equal, we should not choose smaller welfare benefits for humankind over larger benefits. But this is arguably not the only moral consideration relevant to assessing such proposals. Many strategies for increasing the formal political power of the young entail the provision of unequal formal political power to members of the demos and unequal opportunities for influence. There is a sense in which giving the young disproportionate formal political power amounts to giving them power over the elderly, raising questions of political legitimacy: there are weighty moral reasons against giving greater formal political power to some citizens than to others.

146

Empowering Future People by Empowering the Young?

There are, however, at least three reasons why such legitimacy concerns are not sufficient grounds for dismissing proposals such as age-weighted voting, which give greater formal political power to the young.⁴ First, in a society where the elderly control a great deal of informal political power (such as social and economic influence), providing greater formal political power (such as votes) to the young may, in fact, lead to an increase in the equality of the political system overall, making it more egalitarian than a system in which all members of the demos have equal formal political power. Second, apportioning formal political power based on age is consistent with treating people equally across their whole lives. If younger people are apportioned greater voting power, this is a privilege allocated to everyone who will ever live in that political system, during the period in which they are young (MacAskill 2019). Total formal political power across people’s lives remains equal among the demos. Third, inequalities in power must be rectified not only within a generation but also between generations. The current generation wields immense power over future generations. Future citizens are subjected to our laws and the causal fallout of all of our decisions. A political system that allows unequal power relationships between some of its present members in order to give a greater say (by proxy) to future generations over the laws by which they are governed may therefore be a more legitimate system than one that ensures equal power relationships between its present members and leaves future generations disempowered (González-Ricoy and Gosseries 2016b; Gosseries 2016; John 2022).⁵ So we have what appears to be a promising argument for apportioning greater political power to the young, such as by weighting votes with age or instituting legislative youth quotas. Intuitively, people are likely to have more near-term-orientated preferences as they age and will therefore generally support more short-termist policy. Redistributing political power to younger citizens is therefore likely to make political systems less short-termist. Given that short-termism causes immense harms, both fiscal and material, this would have extremely good results, and considerations of legitimacy do not clearly prohibit us from redistributing political power in this way.

2. Why the common case for youth empowerment fails on current evidence When we look at the empirical literature on short-termism and ageing, we do not find confirmation for the intuitively plausible hypothesis that older people support more ⁴ For a contrary view, see Karnein and Roser (2015). ⁵ A fourth consideration is this: political legitimacy is fungible with other values. It is widely accepted that we may sometimes use undemocratic procedures to avoid costly errors (Halstead 2017). For example, most constitutions rightly protect people’s fundamental rights from overrule by the majority. It is sometimes morally laudable to accept costs to the democratic equality of a decision procedure to achieve large welfare and other benefits, if this is the best way to achieve those benefits.

Ageing without Ageism?

147

short-termist policy. It turns out that the common view is flawed in its core assumption: the empirical literature does not show any systematic correlations between ageing and shorter time-horizons.

2.1 Subjective discount rate The standard method of measuring intertemporal trade-offs or ‘time preference’ is the subjective discount rate (SDR). The SDR measures the extent to which people discount the value of goods the longer they must wait to receive those goods. In studies of SDR, participants are typically offered a series of choices between receiving some amount of good (such as money) or burden (such as required effort) at a time in the near future or a larger amount of that good or burden in the more distant future. After several choices, researchers can derive a discount function which maps participant willingness to discount goods across varying periods of time. The argument for increasing the political power of the young assumes that younger people will tend to have a lower SDR. It is based on the assumption that older people, given their smaller remaining lifespan, are less interested in receiving goods in the future (given the risk of dying beforehand) and more interested in pushing burdens into the future than younger people. Some work in economic theory predicts that the SDR will be a little more complicated than this. The SDR might be a U-shaped function: the youngest people have a high discount rate because they lack self-control, their SDR decreases throughout life as they gain self-control and have children whose futures they must care about, and people’s SDR then increases towards the end of life when they have fewer remaining life-years (Read and Read 2004; Chu et al. 2010; Chao et al. 2009). If this picture were accurate, we might instead empower the middleaged rather than the young to secure intergenerational justice. Surprisingly, extant empirical research on ageing confirms neither the hypothesis that SDR increases monotonically with age nor that SDR is a U-shaped function of age. Instead, we find an assortment of mixed results. Several studies have, indeed, found that SDR increases with age (Green et al. 1994; Read and Read 2004; Liu et al. 2016; Seamen et al. 2016; Vanderveldt 2016), although studies are equally likely to find that SDR decreases with age (Harrison et al. 2002; Trostel and Taylor 2001; Löckenhoff et al. 2011; Eppinger et al. 2012; Halfmann et al. 2013; Bixter and Rogers 2019) or that there is no relationship between the two (Chao et al. 2009; Roalf et al. 2011; Rieger and Mata 2015). This variability persists if we exclude studies with small sample sizes (N < 268) or reduce cultural variation by including only studies with sample populations in the United States. Only one study has found an SDR that is a U-shaped function of age (Read and Read 2004), and only one study has found an SDR that is a U-shaped function of health and survival expectations (Chao et al. 2009). Several other studies contradict these results (Green et al. 1994; Harrison et al. 2002, Trostel and Taylor 2001; Löckenhoff et al. 2011; Rieger and Mata 2015; Liu et al. 2016; Seamen et al. 2016).

148

Empowering Future People by Empowering the Young?

A limitation of these studies is that many elicit discount rates over timescales shorter than a year, significantly reducing the hypothesized effect of a shorter life expectancy on one’s SDR. But even in the three studies with a time horizon of at least 10 years, we find no univocal takeaway (Green et al. 1994; Read and Read 2004; Vanderveldt 2016). The best explanation for this variability appears to be that SDR is a highly multifaceted phenomenon that is mediated by numerous factors other than age, including wealth, retirement, and political ideology, and which varies in direction and magnitude from one decision context to another. We therefore need more nuanced research on SDR to isolate and measure the various psychological processes and determinants which underlie its complex structure. This explanation is supported by a 2002 review of three decades of literature on intertemporal choice, which finds subjective discount rates ranging from negative integers to infinity (Frederick et al. 2002). The study concludes that there is no empirical support for the idea that intertemporal choice should be modelled with a single discount rate that is consistent across choice situations. The argument that a smaller additional life expectancy leads to a higher discount rate appears to have assumed too naïve a picture of decision psychology. We need a better way to find out what, if any, relationship holds between age and short-term policy preference.⁶

2.2 Values The most direct way to assess the relationship between age and short-term policy preference is to look at the policy preferences that people actually have, expressed in voting. The common view described in section I implies that younger people will tend to have less short-termist policy preferences. Once again, the empirical literature presents us with mixed findings. The effect of age on short-term policy preference appears to be minimal or non-existent. Cohort effects, social cohesion between voters and other groups, ideological identity, and policy uncertainty appear to be much bigger drivers of short-termism in voter behaviour. As such, simply apportioning greater political power to the young will do little to reduce political short-termism. Two recent studies of 305 Swiss and 82 international referenda offer strong prima facie support to the idea that age is correlated with preferences for more short-termist policy. These studies, conducted by Gabriel Ahlfeldt and colleagues, categorize the answers to referenda questions according to the generational interests that they most promote (young vs elderly), analysing the extent to which younger and older voters ⁶ An additional obstacle for the move from intertemporal trade-offs in personal losses and gains to voting behaviour is the vote–buy gap (Norwood et al. 2019; Paul et al. 2019). Across a range of policy areas, such as green energy, for-profit prisons, and caged eggs, citizens regularly take political action that is in apparent conflict with their consumer behaviour: raising their taxes to fund green energy while failing to source their own electricity from windmills, banning prison practices which they support with their banking, and even banning products which they regularly consume. It appears that looking at people’s personal consumer and financial habits is a bad way to predict the policies they will support.

Ageing without Ageism?

149

support referenda decisions that are in their generational self interest, and come away with the strong conclusion that referenda ‘voters make deliberate choices that maximize their expected utility conditional on their stage in the lifecycle’ (Ahlfeldt et al. 2018, 2019). They find that younger voters ‘tend to be less conservative, attach higher priorities to the protection of the environment, and are more supportive of policies that, in relative terms, benefit the young’. In the Swiss study, these values are shown to swing free from cohort effects as well as the effects of status quo habituation. If the findings of Ahlfeldt and colleagues are accurate and generalizable to most countries and times, then they provide some support for the idea that older voters in fact support more short-termist policies. However, their findings are of limited significance. A key limitation is that many of the policy areas in which Ahlfeldt and colleagues find generational conflict are areas of intratemporal generational conflict rather than intertemporal generational conflict. For decisions about issues such as end-of-life care, school spending, sports facilities, retirement, unemployment, and transportation (the majority of the referenda studied), generational differences in attitudes likely correspond to people’s different preferences in their current stage of life rather than their different preferences over longer timescales. Finding that working people want better unemployment benefits does not indicate that they have longer time horizons in view but rather that they are at greater present risk of unemployment. The studies’ findings of significance for our purposes, then, are exclusively the effects of age on environmental and energy policy. In the international study, Ahlfeldt et al. (2019) find that age is not an unambiguous determinant of voting decisions on environmental legislation.⁷ In the Swiss study, which is much larger and better controlled, Ahlfeldt and colleagues find that age is a significant determinant of proenvironmentalist attitudes, to the point that a 20-year-old voter is 10 per cent more likely to vote favourably to the environment than an 80-year-old voter (Ahlfeldt et al. 2018: 16). The other major study on the effects of ageing on support for more short-termist policy comes from Alan Jacobs and J. Scott Matthews, who survey 1,213 votingage US citizens about their preferences on hypothetical US Social Security policy (Jacobs and Matthews 2012). The study asks participants to vote on a Social Security reform proposal that would impose taxes and benefit cuts for the next 5 or 40 years in order to prevent taxes and benefit cuts that were much larger when the period had ended. Jacobs and Matthews introduce two experimental manipulations, varying both the timing of the policy benefits and the causal complexity of the reform. Strikingly, Jacobs and Matthews find that time preference is not a major driver of short-termism in policy preferences. Measurements of participants’ subjective discount rate in fact inversely correlated with participants’ willingness to forego more distant future benefits. Importantly, Jacobs and Matthews find no discernible effect of age on participant decisions about long-term policy investment. Instead, they find ⁷ Findings on the relationship between age and support for green energy in this study are omitted because the particular referenda analysed are also included in the 2018 study.

150

Empowering Future People by Empowering the Young?

that the primary drivers of short-term political decision-making are uncertainties about whether (i) policies will, in fact, have their intended long-run effects; and (ii) future political leaders will act on the commitments we make today rather than reneging.

2.3 Implications The existing data on the relationship between age and short-termist political decision-making is ambiguous. No stable relationship has been found between age and SDR, and the direct effect of age on short-termism appears relatively small (10 per cent less likely to support green policy over the lifespan) or non-existent. Empowering the young by merely allocating them more formal political power may yet help combat political short-termism. In some countries and times, a small percentage increase in green policy support may be enough to shift energy policy. Depending on cohort effects, some groups of young people may be especially orientated towards the long term, making the contingent effect of proposals to empower the young much stronger. But given that young people’s voting behaviour does not appear to be significantly less short-termist than that of older people, simply giving young people additional votes is not a robust way to ameliorate political short-termism.

3. A forward-looking assembly On present evidence, we have little reason to believe that the young will use additional votes or legislative seats in a more prudent way than the elderly. But in this section, I argue that there are novel and powerful ways to harness young people’s greater remaining life expectancy for the advantage of future generations which do not rely on younger people having less short-termist policy preferences. The basic proposal is to create a novel youth assembly, a permanent, soft-power institution whose members are randomly selected from among the young, which rewards assembly members for successful policymaking 30 years in the future, based on its later effects. Youth, here, is being exploited centrally to extend the time horizon over which assembly members can expect to reap future rewards.

3.1 Futures assemblies ‘Citizens’ assemblies’ have been employed for consultation and informationgathering purposes throughout the world. These randomly selected groups of citizens provide deliberative and non-binding advice to the government in consultation with recognized experts. One of the most high-profile initiatives was Ireland’s

Ageing without Ageism?

151

100-member Citizens’ Assembly, which was established in 2016 and tasked with considering questions related to abortion, fixed-term parliaments, referenda, population ageing, climate change, and gender equality. The deliberations of the Irish assembly provoked a referendum to remove Ireland’s constitutional ban on abortion and substantially shaped Ireland’s Climate Action Plan (Coleman et al. 2019). The success of the Irish assembly and of citizens’ assemblies around the world reveals the promise of citizens’ assemblies tasked with the explicit mandate to represent future generations, or ‘futures assemblies’, for ameliorating short-termism. A general futures assembly, constituted by a stratified random sample of the general population, would have numerous features that predict success at combating shorttermism (John and MacAskill 2021). Being an unelected and publicly funded body, a futures assembly would be insulated from perverse election and fundraising incentives that pressurize elected officials to focus on near-term, visible issues that can help them gain re-election. Being randomly selected, it would be statistically representative of the general population and not chosen or excessively influenced by elected officials. And citizens’ assemblies have a demonstrated aptitude in laboratory and real-world experiments for reducing the deleterious effects of partisanship on careful, long-term deliberation (Fishkin and Luskin 2005; List et al. 2013; Fishkin et al. 2017). In the most recent major assembly, the Climate Assembly UK, 98 per cent of assembly members claimed to have understood almost everything that those in their deliberation groups had said, and 94 per cent felt respected by their fellow participants under disagreement (with none feeling disrespected) (Climate Assembly UK 2020). Finally, citizens’ assemblies are more informed than ordinary voters due to their deliberations with experts, reducing the deleterious effect of policy uncertainty on short-term policy support as found by Jacobs and Matthews’ (2012) Social Security reform experiment. Most importantly, a general futures assembly may need no incentive to reflect carefully on the interests of future generations beyond an explicit mandate to do so. Some limited evidence from the Kochi University Research Institute for Future Design suggests that when parents are explicitly asked to cast votes on behalf of their children, they vote for different parties than they normally would vote for in a sizable minority of cases (Aoki and Vaithianathan 2012). This is a promising sign that those who are asked explicitly to represent the younger generation do not simply use this opportunity to promote their own agenda but rather aim to promote the interests of the young and thereby adopt longer time horizons for political decision-making. This is further supported by evidence that actors within institutions tend to be compelled to follow norms and perform roles that are consistent with the established culture of their institution (Goodin 1995; Steiner et al. 2005: 127; MacKenzie 2016). Put simply, people who are asked explicitly to vote on behalf of another group seem to do so. So, a general futures assembly, tasked with representing future generations and giving non-binding advice to the government, would likely do well at ameliorating political short-termism. But it is possible to improve upon the assembly in two ways: first, by better aligning the incentives of the deliberating body with the interests of

152

Empowering Future People by Empowering the Young?

future generations and, second, by making its non-binding advice more difficult for the government to ignore.

3.2 Mechanism design An explicit mandate may be sufficient to motivate futures assembly members to adequately consider future generations in their recommendations.⁸ But there remain concerns of value drift, irrational time discounting, and capture by political elites and industry. One promising and underexplored mechanism for aligning incentives with the interests of future generations involves retrospective accountability. The most central problem of representing the interests of future generations in government is that of making political actors accountable to future generations. Future generations cannot vote in our elections, nor can they sanction or protest the decisions of their forebears. Retrospective accountability solves the accountability problem by rewarding policymakers many years into the future in proportion to the effects of their policy on the long run. A simple mechanism of retrospective accountability would involve empowering a body of future auditors—say, 30 years from now—to decide on the pension bonus of the decision-makers today based on how successfully these decision-makers promote the interests of future people. This would provide decisionmakers today with a positive financial incentive to look to the future—at least 30 years from now—when making any decisions. Such a mechanism would yield a significant advance on the time horizons of present institutions. A more sophisticated retrospective accountability mechanism (and the one I favour) would exploit strategic iteration of this mechanism to extend the time horizons of government much further again. On the iterated variant, the future auditors who decide on the later bonuses of present decision-makers themselves face a financial incentive to look again into the future. For their own financial situation will be tied to the evaluations of the next generation of auditors, who will determine their pension bonuses.⁹ To get a nice retirement bonus, future auditors have an incentive to evaluate present decision-makers in accordance with the preferences of the next generation of auditors, and so present decision-makers have an incentive to satisfy the preferences of the auditors two generations—60 years—from now. And so iterated, until we have extended the horizons of government to the longest time period relevant for political decision-making. On the simplest implementation of such accountability measures, each assembly decides on the bonus of the assembly 30 years prior. Because a futures assembly is a soft-power institution, with no formal powers of censure, a second obstacle it must overcome is ignorance by elected officials. If elected officials do not seriously consider the advice of the assembly, the latter will ⁸ These mechanism design issues are discussed at much greater length in John (2022: Chapter 2). ⁹ For a discussion of related formal models of intergenerational bargaining, see Heath (1997), who develops and extends a model by Gauthier (1986).

Ageing without Ageism?

153

have no power whatsoever. To overcome this barrier, futures assemblies should have two key features. First, they should be empowered to require reading and response from the legislature, ensuring that their advice is actually read. Second, they should be designed to be highly public and high-status institutions. All of their deliberations and reports should be public-facing, with a strong media team and minimal institutional complexity to ensure that the institution is well understood by the public. The assembly should be well paid and highly informed by experts and should be constituted by a demographically stratified random sample of the population. These features together will help to ensure that the assembly has high perceived legitimacy so that the neglect of the institution by elected officials will be unpopular among voters. Tempting though it may be, we should, at the given political moment, resist giving futures assemblies any stronger formal powers than this since the major reason that future-orientated institutions are repealed is that they have too much formal power (Jones et al. 2018). Citizens’ assemblies avoid repeal through their soft-power approach.

3.3 A novel youth assembly Two central considerations favour the adoption of a futures assembly constituted by the young (i.e. eligible voters under 40 and perhaps younger) rather than by the general populace. The first and simpler reason is that younger people can generally expect to live longer, allowing for much greater time horizons for a retrospective accountability mechanism. The older the assembly members are, the sooner they will need to receive a bonus for it to be valuable and the less valuable to them it will be. The second and more speculative reason is rooted in the finding that group deliberation creates greater empathy and solidarity between participants (Grönlund et al. 2017). There is evidence that such social cohesion makes voters more likely to act on the preferences of the larger group (Berkman and Plutzer 2004). Such social cohesion formed part of the explanation, in section II, for the minimal effect of age on short-term policy preference. If group deliberation succeeds in partly breaking down their cohesion with other ideological identities and interest groups and causes assembly members to form a more strongly youth-based political identity, they may, in turn, be more inclined to support the (long-term) interests of the young than ordinary voters.

4. Conclusion Political short-termism costs the global economy many billions and perhaps trillions of dollars annually and leads directly to millions of deaths from disasters and suboptimal resource allocation. This chapter has considered one popular set of proposals for ameliorating political short-termism rooted in the plausible thought that younger

154

Empowering Future People by Empowering the Young?

people will be more motivated to consider the long term in political decision-making given their longer remaining life expectancy. It has been shown that this prima facie plausible thought is severely lacking in empirical support. Younger people are not significantly more motivated to consider the long term in their voting behaviour, and this greatly weakens the case for simple systems of formally empowering the young. However, there are promising signs that alternative mechanisms for empowering the young would more significantly ameliorate short-termism, such as those incorporating retrospective liability. I have defended a ‘Youth Futures Assembly’, which incorporates such a mechanism and can be implemented to significant long-term beneficial effect. Further experimental evidence about the policy preferences of political surrogates, the incentive effects of retrospective accountability, and the relationship between group cohesion and policy preference could significantly strengthen or weaken the case for such a youth assembly, as could further, experimental evidence on the relationship between age and policy preference and on alternative mechanisms for ameliorating political short-termism. As the discussion in this chapter has shown, given the high costs of short-termism and the severity with which it plagues modern institutions, investigation into such matters are vital to the future livelihoods of people everywhere, from those who are our contemporaries to the myriad heirs of posterity.

Acknowledgements This chapter was greatly improved by feedback from Greg Bognar, Adam Gibbons, and Axel Gosseries.

References Ahlfeldt, Gabriel M., Wolfgang Maennig, and Steffen Q. Mueller. 2018. ‘The Generation Gap in Direct Democracy’. CEPR Discussion Paper. https://ideas.repec.org/p/cep/ cepdps/dp1552.html. Ahlfeldt, Gabriel M., Wolfgang Maennig, and Malte Steenbeck. 2019. ‘Direct Democracy and Intergenerational Conflicts in Aging Societies’. Journal of Regional Science 60 (1): 129. Aoki, Reiko, and Rhema Vaithianathan. 2012. ‘Intergenerational Voter Preference Survey—Preliminary Results’. CIS Discussion Paper Series. https://ideas.repec.org/p/ hit/cisdps/539.html. Arnold, R. Douglas. 1990. The Logic of Congressional Action. Yale University Press. New Haven, CT. Beck, Nathaniel. 1982. ‘Does There Exist a Political Business Cycle: A Box-Tiao Analysis’. Public Choice 38 (2), 205.

Ageing without Ageism?

155

Beckstead, Nick. 2013. ‘On the Overwhelming Importance of Shaping the Far Future’. PhD dissertation, Rutgers University. Berkman, Michael B., and Eric Plutzer. 2004. ‘Gray Peril or Loyal Support? The Effects of the Elderly on Educational Expenditures’. Social Science Quarterly 85 (5): 1178. Bidadanure, Juliana. 2016. ‘Youth Quotas, Diversity, and Long-Termism’. In Institutions for Future Generations, ed. Iñigo González-Ricoy and Axel Gosseries. Oxford: Oxford University Press, 266–281. Binder, Sarah A. 2006. ‘Can Congress Legislate for the Future?’ Research Brief, December. https://www.nyu.edu/content/dam/nyu/brademasCenter/documents/Research/ publications-legislating-future-binder.pdf. Bixter, Michael T., and Wendy Anne Rogers. 2019. ‘Age-Related Differences in Delay Discounting: Immediate Reward, Reward Magnitude, and Social Influence’. Journal of Behavioral Decision Making 32 (4): 471. Caney, Simon. 2016. ‘Political Institutions for the Future: A Five-Fold Package’. In Institutions for Future Generations, ed. Iñigo González-Ricoy and Axel Gosseries. Oxford: Oxford University Press, 135–155. Chao, Li-Wei, Helena Szrek, Nuno Sousa Pereira, and Mark V. Pauly. 2009. ‘Time Preference and Its Relationship with Age, Health, and Survival Probability’. Judgment and Decision Making 4 (1): 1. Climate Assembly UK. 2020. ‘The Path to Net Zero: Climate Assembly UK Full Report’. https://www.climateassembly.uk/recommendations/index.html. Coleman, Martha, Laura Devaney, Diarmuid Torney, and Pat Brereton. 2019. ‘Ireland’s World-Leading Citizens’ Climate Assembly. What Worked? What Didn’t?’ Climate Home News, 27 June. https://www.climatechangenews.com/2019/06/27/irelandsworld-leading-citizens-climate-assembly-worked-didnt. Chu, Cyrus C.Y., Hung-Ken Chien, and Ronald D. Lee. 2010. ‘The Evolutionary Theory of Time Preferences and Intergenerational Transfers’. Journal of Economic Behavior & Organization 76 (3): 451. Eppinger, Ben, Leigh E. Nystrom, and Jonathan D. Cohen. 2012. ‘Reduced Sensitivity to Immediate Reward during Decision-Making in Older than Younger Adults’. PloS One 7 (5): 1. Fishkin, James S., and Robert C. Luskin. 2005. ‘Experimenting With a Democratic Ideal: Deliberative Polling and Public Opinion’. Acta Politica 40 (3): 284. Fishkin, James S., Roy William Mayega, Lyn Atuyambe, Nathan Tumuhamye, Julius Ssentongo, Alice Siu, and William Bazeyo. 2017. ‘Applying Deliberative Democracy in Africa: Uganda’s First Deliberative Polls’. Daedalus 146 (3): 140. Frederick, Shane, George Loewenstein, and Ted O’Donoghue. 2002. ‘Time Discounting and Time Preference: A Critical Review’. Journal of Economic Literature 40 (2): 351. Gauthier, David. 1987. Morals by Agreement. Clarendon Press. Oxford. González-Ricoy, Iñigo, and Axel Gosseries, eds. 2016a. Institutions for Future Generations. Oxford: Oxford University Press.

156

Empowering Future People by Empowering the Young?

González-Ricoy, Iñigo, and Axel Gosseries, eds. 2016b. ‘Designing Institutions for Future Generations: An Introduction’. In Institutions for Future Generations, ed. Iñigo González-Ricoy and Axel Gosseries. Oxford: Oxford University Press, 3–23. Goodin, Robert E. 1995. Utilitarianism as a Public Philosophy. Cambridge: Cambridge University Press. Gosseries, Axel. 2016. ‘Generational Sovereignty’. In Institutions for Future Generations, ed. Iñigo González-Ricoy and Axel Gosseries. Oxford: Oxford University Press, 98–116. Greaves, Hilary, and William MacAskill. 2021. ‘The Case for Strong Longtermism’. Global Priorities Institute Working Paper Series No. 5-2021. Green, Leonard, Astrid F. Fry, and Joel Myerson. 1994. ‘Discounting of Delayed Rewards: A Life-Span Comparison’. Psychological Science 5 (1): 33. Grönlund, Kimmo, Kaisa Herne, and Maija Setälä. 2017. ‘Empathy in a Citizen Deliberation Experiment’. Scandinavian Political Studies 40 (4): 457. Halfmann, Kameko, William Hedgcock, and Natalie L. Denburg. 2013. ‘Age-Related Differences in Discounting Future Gains and Losses’. Journal of Neuroscience, Psychology, and Economics 6 (1): 42. Halstead, John. 2017. ‘High Stakes Instrumentalism’. Ethical Theory and Moral Practice 20 (2): 295. Harrison, Glenn W., Morten I. Lau, and Melanie B. Williams. 2002. ‘Estimating Individual Discount Rates in Denmark: A Field Experiment’. American Economic Review 92 (5): 1606. Healy, Andrew, and Neil Malhotra. 2009. ‘Myopic Voters and Natural Disaster Policy’. American Political Science Review 103 (3): 387. Heath, Joseph. 1997. ‘Intergenerational Cooperation and Distributive Justice’. Canadian Journal of Philosophy 27 (3): 361. Jacobs, Alan M., and J. Scott Matthews. 2012. ‘Why Do Citizens Discount the Future? Public Opinion and the Timing of Policy Consequences’. British Journal of Political Science 42 (4): 903. John, Tyler. 2022. ‘Chronopolitanism: Political Inclusion and the Distant Future’. PhD dissertation, Rutgers University. John, Tyler, and William MacAskill. 2021. ‘Longtermist Institutional Reform’. In The Long View, ed. Natalie Cargill and Tyler M. John. FIRST Strategic Insight Ltd., 2021. Johnson, Dominic, and Simon Levin. 2009. ‘The Tragedy of Cognition: Psychological Biases and Environmental Inaction’. Current Science 97 (11): 1593. Jones, Natalie, Mark O’Brien, and Thomas Ryan. 2018. ‘Representation of Future Generations in United Kingdom Policy-Making’. Futures 102 (1): 153. Karnein, Anja, and Dominic Roser. 2015. ‘Saving the Planet by Empowering the Young?’ In Youth Quotas and Other Efficient Forms of Youth Participation in Ageing Societies, ed. Jörg Tremmel, Antony Mason, Petter Haakenstad Godli, and Igor Dimitrijoski. Springer, 77–92.

Ageing without Ageism?

157

List, Christian, Robert C. Luskin, James S. Fishkin, and Iain McLean. 2013. ‘Deliberation, Single-Peakedness, and the Possibility of Meaningful Democracy: Evidence from Deliberative Polls’. Journal of Politics 75 (1): 80. Liu, Lu-lu., Xing-jie Chen, Ji-fang Cui, Jing Wang, Yi-bing Zhang, David L. Neumann., David H.K. Shum, Ya Wang, and Raymond C.K. Chan. 2016. ‘Age Differences in Delay Discounting in Chinese Adults’. Personality and Individual Differences 90 (1): 205. Löckenhoff, Corinna E., Ted O’Donoghue, and David Dunning. 2011. ‘Age Differences in Temporal Discounting: The Role of Dispositional Affect and Anticipated Emotions’. Psychology and Ageing 26 (2): 274. MacAskill, William. 2019. ‘Age-Weighted Voting’. Effective Altruism Forum, 12 July. https://forum.effectivealtruism.org/posts/b7BrGrswgANP3eRzd/age-weightedvoting. MacAskill, William. 2022. What We Owe the Future. London: OneWorld Publications. MacKenzie, Michael K. 2016. ‘Institutional Design and Sources of Short-Termism’. In Institutions for Future Generations, ed. Iñigo González-Ricoy and Axel Gosseries. Oxford: Oxford University Press, 24–48. Makridis, Christos, and Jonathan Hartley. 2020. ‘The Cost of Covid-19: A Rough Estimate of the 2020 US GDP Impact’. Mercatus Special Edition Policy Brief, 6 April. https://www.mercatus.org/publications/covid-19-policy-brief-series/cost-covid-19rough-estimate-2020-us-gdp-impact. Mayhew, David. 1974. Congress: The Electoral Connection. New Haven: Yale University Press. Multihazard Mitigation Council. 2017. ‘Natural Hazard Mitigation Saves 2017 Interim Report: An Independent Study’. Washington, DC: National Institute of Building Sciences. https://www.fema.gov/sites/default/files/2020-07/fema_ms2_interim_ report_2017.pdf. Norwood, Franklin Bailey, Glynn T. Tonsor, and Jayson Lusk. 2019. ‘I Will Give You My Vote but Not My Money: Preferences for Public versus Private Action in Addressing Social Issues’. Applied Economic Perspectives and Policy 41 (1): 96. Ord, Toby. 2020. The Precipice: Existential Risk and the Future of Humanity. London: Bloomsbury. Paul, Andrew S., Jayson L. Lusk., F. Bailey Norwood, and Glynn T. Tonsor. 2019. ‘An Experiment on the Vote-Buy Gap with Application to Cage-Free Eggs’. Journal of Behavioral and Experimental Economics 79 (1): 102. Piper, Kelsey. 2020. ‘Young People Have a Stake in Our Future. Let Them Vote’. Vox. https://www.vox.com/future-perfect/2019/9/10/20835327/voting-age-youth-rightskids-vote. Read, Daniel, and N.L. Read. 2004. ‘Time Discounting over the Lifespan’. Organizational Behavior and Human Decision Processes 94 (1): 22. Rieger, Matthias, and Rui Mata. 2015. ‘On the Generality of Age Differences in Social and NonSocial Decision Making’. Journals of Gerontology Series B: Psychological Sciences and Social Sciences 70 (2): 200.

158

Empowering Future People by Empowering the Young?

Roalf, David R., Suzanne H. Mitchell, William T. Harbaugh, and Jeri S. Janowsky. 2011. ‘Risk, Reward, and Economic Decision Making in Aging’. Journals of Gerontology Series B: Psychological Sciences and Social Sciences 67 (3): 289. Seaman, Kendra L., Marissa A. Gorlick., Kruti M. Vekaria, Ming Hsu., David H. Zald, and G.regory R. Samanez-Larkin. 2016. ‘Adult Age Differences in Decision Making across Domains: Increased Discounting of Social and Health-Related Rewards’. Psychology and Aging 31 (7): 737. Shrank, William H., Teresa L. Rogstad, and Natasha Parekh. 2019. ‘Waste in the US Health Care System: Estimated Costs and Potential for Savings’. Journal of the American Medical Association 322 (15): 1501. Steiner, Jürg, André Bächtiger, Markus Spörndli, and Marco R. Steenbergen. 2005. Deliberative Politics in Action: Analyzing Parliamentary Discourse. Cambridge: Cambridge University Press. Tarsney, Christian. 2019. ‘The Epistemic Challenge to Longtermism’. Paper presented at the First Workshop in Global Priorities Research, Oxford, March 2019. Trostel, Philip A., and Grant A. Taylor. 2001. ‘A Theory of Time Preference’. Economic Inquiry 39 (3): 379. Tufte, Edward R. 1978. Political Control of the Economy. Princeton: Princeton University Press. United Nations Office for Disaster Risk Reduction. 2015. ‘Global Assessment Report on Disaster Risk Reduction 2015’. https://www.preventionweb.net/english/hyogo/gar/ 2015/en/home/GAR_2015/GAR_2015_1.html. US Council of Economic Advisers. 2014. The Cost of Delaying Action to Stem Climate Change. Washington, DC: White House. Van Parijs, Philippe. 1998. ‘The Disfranchisement of the Elderly, and Other Attempts to Secure Intergenerational Justice’. Philosophy & Public Affairs 27 (4): 292. Vanderveldt, Ariana. 2016. ‘Decision Making in Older Adults: A Comparison of Delay and Probability Discounting across Ages’. PhD dissertation, Washington University. Weber, Elke U. 2006. ‘Experience-Based and Description-Based Perceptions of LongTerm Risk: Why Global Warming Does Not Scare Us (Yet)’. Climatic Change 77 (1): 103.

11 COVID-19, Age, and Rationing Greg Bognar

1. Introduction From the spring of 2020, as successive waves of the COVID-19 pandemic hit different countries, hospitals found themselves short of ventilators, intensive care unit (ICU) beds, and qualified medical personnel to take care of patients. They had to ration their resources, deciding who was given a chance to recover and who was left to die. In order to help hospital staff make these decisions, bioethicists, medical associations, and other professional bodies issued guidelines and recommendations for the criteria of admission of patients to ICU support. While these documents agreed to a remarkable extent on the criteria and their justification, there were also some marked differences—most notably, a deep disagreement over the role that age and life expectancy can play in the selection of patients. In March 2020, the Italian Society of Anesthesia, Analgesia, Resuscitation and Intensive Care (SIAARTI) issued a set of recommendations to help hospitals make decisions about admission to intensive care units. SIAARTI argued that it may be necessary to set an age limit on ICU admission in order to give priority to those who have a greater probability of survival and greater life expectancy. People over the limit would not be eligible for ICU treatment.1 To many, this seemed an unacceptable form of ageist discrimination; to others, a tragic but necessary and ethically correct response to a grave public health emergency. Other guidelines, while agreeing with SIAARTI in other recommendations, rejected the use of age. In this chapter, I am going to look at the role of age and life expectancy in the rationing of medical resources in emergency situations, drawing on some of the publicly available COVID-19 critical care recommendations. I will begin by giving a rough sketch of how triage procedures work, their ethical aims, and the reason age arises as a problem for them. I illustrate the disagreement over age by a simplified toy example and then by some real-life examples from COVID-19 critical care recommendations. I conclude by presenting a cautious defence of using age and life expectancy in the rationing of medical resources in public health emergencies. To forestall misunderstanding, it should be noted that this chapter relies exclusively on published guidelines and recommendations and not clinical practice. At 1 For a discussion of the SIAARTI recommendations, see Craxì et al. (2020); Vergano et al. (2020).

Greg Bognar, COVID-19, Age, and Rationing. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Greg Bognar (2023). DOI: 10.1093/oso/9780192894090.003.0012

160

COVID-19, Age, and Rationing

this point, we do not know, and will probably never fully know, how doctors, nurses, and hospital administrators actually made life-and-death rationing decisions in the ‘fog of war’ of fighting a novel virus. But the recommendations of bioethicists and medical associations tell us how people who had thought about these issues carefully believed these decisions should be made. They are the evidence of the ethical responses that different societies made to a major public health emergency.

2. The ethical aims of triage Triage is a kind of rationing procedure for matching patients to the available resources in conditions of extreme scarcity. Traditional triage, a practice invented for battlefield medicine in the Napoleonic Wars, involved sorting the wounded into three groups: those who are likely to survive their injuries without immediate medical attention, those who are unlikely to survive even with immediate medical attention, and those for whom immediate medical attention may make the difference between life and death. Physicians focused on the last group in order to save as many lives as possible. Traditional triage is a procedure for establishing eligibility for a scarce resource according to narrow criteria: there must be both a high chance that the patient would die without the resource and a high chance that the patient would survive with the resource. Medical attention both has to be necessary and have a good prospect of success. The ethical aim of traditional triage is to maximize the number of lives saved. Evidently, this aim is based on benefit maximization—the principle that you ought to do as much good with the available resources as possible. But traditional triage also has a specific, narrow interpretation of benefit: lives. When you focus on lives, you ignore all the other characteristics of patients—for instance, you ignore their chances of long-term survival and care only about their immediate prospects. Chance of shortterm survival is used as the criterion that establishes eligibility for the resource. Thus, triage procedures involve a criterion of eligibility and an interpretation of benefit. Maximizing the number of lives saved is not the only possible ethical aim of triage. One alternative is to interpret benefit as the amount of life rather than the number of lives. In this case, the probability of long-term survival becomes a relevant consideration because treating patients who are likely to survive for a long time maximizes the amount of life saved, while treating patients who are likely to survive only for a short time does not. In practice, amount of life can be conveniently measured in years and long-term survival by life expectancy at the given age. Since, other things being equal, younger patients have better chances of long-term survival, maximizing the amount of life usually implies giving priority to younger patients. Interpreting the unit of benefit as life years gives an indirect role to age. The interpretation of benefit can be even broader. You might think that besides the quantity of life, its quality is also relevant. The ethical aim of triage procedures might be not only to preserve lives but also to preserve lives that go better, at least

Ageing without Ageism?

161

with respect to health. One common method for measuring health-related quality of life is to use quality-adjusted life years or QALYs. On this view, benefit is interpreted as QALYs. I should explain. The construction of QALYs begins by assigning quality of life values to different health outcomes on a scale between 0 and 1, where 0 represents an outcome that is just as bad as death and 1 represents full health. For instance, suppose that in one health state patients have problems with performing normal daily activities and they have some moderate pain; the quality-adjusted weight of this health state might be 0.76. In another health state, patients have some problems with walking, self-care, and other daily activities, and they are in moderate pain and moderately depressed. The quality-adjustment factor in this case might be 0.52. Assume that both of these health states last a full year. In this case, the QALYs associated with the first are 0.76 and the QALYs associated with the second are 0.52. QALYs can be added up. If a patient stays in the first health state for two years, their QALYs are 1.52; if, in the third year, their condition worsens and they end up in the second state, their QALYs are 2.04 (0.76 for two years, 0.52 for one).2 A triage procedure that interprets benefit as QALYs directs you to maximize the number of expected QALYs. Maximizing the amount of life in terms of life years or life expectancy makes age indirectly relevant. Maximizing QALYs makes quality of life directly relevant. You can take a step further. QALYs can be age-weighted: a QALY that goes to a younger patient can be given more weight than a QALY that goes to an older patient. For instance, if a life-saving intervention would provide either a young patient or an old patient with 10 additional years of life in full health, then the 10 QALYs might be multiplied by a greater positive number (representing the age-weights) if the intervention is provided to the young patient than if it is provided to the old patient. Maximizing age-weighted QALYs gives priority to younger and healthier people. These four interpretations of benefit can have very different practical implications. For illustration, consider the following simple example. Suppose you have only two available ventilators in the ICU, but there are four patients who need immediate ventilator support. Without it, the patients will die. You must decide which two of the four patients to admit. The four patients differ in the following ways. Patient A is 30 years old. If her life is prolonged, she can expect to live another 45 years. During most of these years, she can expect to be in good health, with some deterioration at the end. Thus, suppose that her expected QALYs are 40. If those QALYs are age-weighted, such that the QALY value in each year that she expects to live is multiplied by a weight that diminishes as she gets older, the result is 35. Her expected 40 QALYs become 35 age-weighted QALYs. Patient B is 70 years old. His life expectancy is 15 years. He’s in poor health, thus his expected QALYs are only 8, and, because he is old, his expected age-weighted QALYs are only 4. At 65, patient C is younger but she can look forward only to another 10 2 For a more detailed introduction, see Bognar and Hirose (2022).

162

COVID-19, Age, and Rationing Table 11.1 Four patients

Age Life expectancy Expected QALYs Expected age-weighted QALYs

A

B

C

D

30 45 40 35

70 15 8 4

65 10 9 5

40 10 7 6

years. She’s also in better health, thus her 10 years are equivalent to 9 QALYs. Her 9 QALYs yield 5 age-weighted QALYs. Finally, patient D is only 40, but he has a condition that is expected to kill him in 10 years. His health during these years will be reasonably good. His expected QALYs are 7, equivalent to 6 age-weighted QALYs on account of his young age. The information about the four patients is summarized in Table 11.1. If your ethical aim is to maximize the number of lives saved, the information in Table 11.1 is irrelevant. You can, at most, save two patients, and there is no moral reason to favour any of them. Therefore, you should give each of them an equal chance. You can select the two patients for ICU support randomly by holding a lottery; that is, if demand for a life-saving resource exceeds supply, maximizing the number of lives saved requires random selection. If your ethical aim is to maximize the amount of life, then the second row in Table 11.1 becomes relevant. Random selection is not permissible any more since only admitting A and B maximizes the number of life years saved. The role of age in this triage procedure is indirect: even though both C and D are younger than B, B has a greater life expectancy despite being older. Thus, maximizing life years does not always entail saving the younger. If quality of life is a relevant consideration, your choice might be made in terms of expected QALYs. In this example, QALYs can be maximized by selecting A and C. However, if age is also a relevant and independent consideration, then you can take it into account by age-weighting expected benefits. Hence, your ethical aim would be to maximize expected age-weighted QALYs. In this example, you would then select patients A and D. Depending on how benefit is interpreted—and thus the ethical aim of the triage procedure—you get different results. They range from random selection to choosing A and B, A and C, and A and D. Some people might object that maximization is not the appropriate moral principle to apply in the kind of emergency situations that we have considered. They might argue that considerations of equity or justice should not be ignored. For instance, when deciding whom to admit to the ICU, you should give priority to those patients who are worse off. But this proposal runs into the same issue that I have highlighted: it needs to give an interpretation of benefit. It needs to have an account of what it is in terms of which patients can be more or less badly off. The problem is not avoided by substituting ‘giving priority to the worse off ’ for ‘benefit maximization’.

Ageing without Ageism?

163

For instance, suppose that to be worse off is interpreted as having lived less. This is not an implausible interpretation: those who have already had a long life have had the chance to enjoy it and carry out their plans to a greater extent than those who have had less time. On this interpretation, A and D should be admitted to the ICU: they are the youngest. But you could also argue that the worse off are those who have the most to lose if they are not selected, either in terms of remaining years of life (hence, A and B should be chosen), in terms of quality-adjusted life years (A and C), or ageweighted QALYs (A and D). Or, perhaps, the worse off are those who have the fewest years of life left (C and D), the fewest QALYs (B and D), or the fewest age-weighted QALYs (B and C). The way you interpret benefit makes a difference to how you make your rationing choices. As soon as you go beyond interpreting benefit as the number of lives saved (regardless of duration, quality, or any other personal characteristic) and as soon as you move away from simple maximization, things quickly become complicated.

3. Criteria of eligibility As waves of infections during the COVID-19 pandemic peaked, hospitals around the world found themselves short of ventilators and ICU beds. In anticipation and then in response, governments, professional bodies, and bioethicists issued guidelines and recommendations describing the fundamental values underlying ICU admission decisions and the ethical principles to guide those decisions. In most cases, these guidelines are not legal documents. They are more like non-binding professional recommendations. At first glance, the guidelines from various countries look similar. They identify similar underlying values and propose similar sets of principles. But there are differences, and those differences often reveal important ethical disagreements. In the following, I will focus mainly on the role of age and life expectancy. It goes without saying that it is impossible to discuss here any of the guidelines in detail.3 But, generally speaking, the various guidance documents propose three sets of principles: non-discrimination, justice (sometimes called equality or equity), and benefit maximization. Earlier, I distinguished between criteria of eligibility and the interpretation of benefit as components of a triage procedure. Many of the non-discrimination principles in the recommendations can be interpreted as negative criteria of eligibility—that is, as prohibitions on determining eligibility on the basis of certain characteristics or considerations. For instance, the Australian and New Zealand Intensive Care Society (ANZICS) states that ‘access to intensive care must not be based on irrelevant and discriminatory considerations such as gender, sexual orientation, religion, disability, social status, personal connections, wealth, citizenship, insurance status, 3 For useful overviews, see Antommaria et al. (2020); Jöbges et al. (2020); Ehni et al. (2021).

164

COVID-19, Age, and Rationing

ethnicity or race’. In the guidance issued by the Swiss Academy of Medical Sciences, one of the fundamental ethical principles is that resources must be allocated ‘without unjustified unequal treatment on grounds of age, sex, residence, nationality, religious affiliation, social or insurance status, or disability’. And the recommendations issued jointly by eight German medical societies contends that ‘social characteristics’ must not play any role in establishing eligibility.⁴ The recommendations also propose positive criteria of eligibility. Only those patients who have a good chance of survival should be eligible for intensive care treatment. ‘Good chance of survival’ is understood as a good chance of short-term survival—usually, as a good chance of being discharged from the hospital but sometimes as a good chance of being alive after one year. In the German recommendations, this criterion is called clinical prospect of success. We can think of COVID-19 critical care triage as a two-stage process. In the first stage, criteria of eligibility are used to narrow the pool of patients to those who should be considered for admission to intensive care. Non-discrimination principles prohibit the selection to be based on morally irrelevant personal characteristics. Clinical prospect of success requires that the selection be based on probability of short-term survival. In effect, prospect of short-term survival is the only criterion. Accordingly, factors like long-term survival and quality of life are ignored. This works in cases when the number of patients match the available resources after the eligibility criteria are applied. But, in many cases, triage may require a second stage because the number of patients exceeds the available resources. It is at this second stage that benefit may need to be interpreted more narrowly. At this stage, age and life expectancy might become relevant. If this is the correct way to interpret ICU admissions criteria in a public health emergency like the COVID-19 pandemic, then it can be understood why it is not unreasonable to object to age limits on ICU admissions. As I mentioned in the Introduction, in its March 2020 recommendations, SIAARTI argued that it may be necessary to set an age limit on ICU admissions. To many people, this looked like a deeply unjust form of discrimination; to others, an unfortunate, but unavoidable and reasonable way to manage extreme scarcity. But one can argue that the real problem is that the age limit was proposed as a criterion of eligibility. People over the limit would have been ruled out from the start (they would have had no claim to ICU treatment at all), regardless of their other characteristics. For instance, assuming that the limit is at 70, patient B in my example from section 2 would have been rejected even though he had the second-greatest life expectancy among the four patients. An age limit as a criterion of eligibility would also be inconsistent with the principle, included in many recommendations, that no patient should be ruled out categorically from the start.⁵ Negative criteria of eligibility, as the basis for categorical

⁴ See ANZICS (2020) for the Australian and New Zealand recommendations, SAMW (2020) for the Swiss, and Marckmann et al. (2020) for the German ones. ⁵ See, e.g. Auriemma et al. (2020); White and Lo (2020).

Ageing without Ageism?

165

exclusion, should be restricted to clinical prospect of success. Age cannot, in itself, be a factor in determining prospect of success, even though it may co-vary with other, relevant factors. As the British Medical Association acknowledges in its guidance notes: A simple ‘cut-off’ policy with regard to age or disability would be unlawful as it would constitute direct discrimination. A healthy 75-year-old cannot lawfully be denied access to treatment on the basis of age. However, older patients with severe respiratory failure secondary to COVID-19 may have a very high chance of dying despite intensive care, and consequently have a lower priority for admission to intensive care. (BMA 2020)

Thus, even if age is often correlated with clinical prospect of success, it cannot provide a ground for exclusion itself. On the other hand, it is not clear why age should be ruled out from the second stage of the triage process, when benefit may need to be interpreted more narrowly. In the recommendations issued by the German medical societies, ‘calendar age’ is ruled out from playing any part in COVID-19 triage because, they claim, this would conflict with equality.⁶ Presumably, by equality, the medical societies mean equal treatment at a point in time because the refusal to consider calendar age may lead to very unequal outcomes. For instance, consider the idea that equality should be interpreted as equality of whole lives.⁷ Equal treatment at a point in time can lead to very unequal outcomes when entire lives are considered. To see the point, consider again my example from section 2. Suppose you have to decide whether to treat A or B. A is 30 years old and can expect to live for another 45 years if she is treated; B is 70 years old and can expect to live for another 15 years if he is treated. According to the German recommendations, age cannot be considered in the decision because that would amount to unequal treatment. Thus, it might be that B gets the treatment, and, as a result, he survives to 85 while A dies at 30. This outcome is certainly much more unequal than treating A, who would then survive until 75 while B dies when he is 70. If you care about equality between lives, the prohibition on taking calendar age into account is wrong. The prohibition of giving any role to age also conflicts with so-called ‘life-cycle’ considerations. ANZICS suggests that ‘if a situation arises where patients are similarly ranked in terms of clinical priority, some legitimate instances where it may be ethically justifiable to consider other determinants for prioritisation include [. . .] that younger patients who have lived through fewer life stages are prioritised over older patients’ (ANZICS 2020). (Clinical priority is what is elsewhere called short-term survival or clinical prospect of success.) The Critical Care Society of Southern Africa (CCSSA) recommends that if patients in a priority group are tied, then the decision should be made according to age: younger patients, who have had less opportunity ⁶ See Marckmann et al. (2020: S116). ⁷ See also Chapter 6 in this volume, in which Axel Gosseries discusses the ‘lifetime’ view.

166

COVID-19, Age, and Rationing

to live through life’s stages, should get priority. They recommend the following life stages: ages 12–40, 41–60, 61–75, and over 75. As the Society argues: the ethical justification for incorporating the life-cycle principle is that it is a valuable goal to give individuals equal opportunity to pass through the stages of life—childhood, young adulthood, middle age, and old age. [. . .] Younger individuals receive priority because they have had the least opportunity to live through life’s stages. (CCSSA 2020)

In both cases, it is clear that life-cycle considerations are not criteria of eligibility. They come into play only in the second stage of triage, when demand for ICU support exceeds supply. Moreover, the ‘life-cycle principle’ is not a purely benefit maximizing principle, even if it implicitly assumes that younger people have a greater life expectancy; that is, it assumes that they have a good chance to live through their remaining life-cycles. Rather, the ‘life-cycle principle’ belongs to the family of equal opportunity principles. It is similar to views that give priority to achieving ‘complete lives’ or ‘fair innings’. The basic idea of such views is that it is valuable if people reach all stages of a normal human life, and everyone should have a fair go, or equal opportunity, to do so. While some of these views recommend strict age limits (for instance, arguing that by 75 one has gone through all life stages), others introduce an age-weighting function that gives diminishing value to later years (or QALYs, if quality of life is also taken into account).⁸ Although there are good reasons to reject age limits as criteria of eligibility for ICU admission, there are also good reasons not to reject age and life expectancy when further selection is necessary among eligible patients. Doing so conflicts with ‘complete lives’ equality, equality of opportunity, and age-weighted benefit maximization.

4. Interpretation of benefit There is broad agreement that during a public health emergency, when resources are scarce, benefit maximization is the primary moral principle. There is less agreement, however, on how ‘benefit’ should be interpreted. In the simplest case, maximizing benefits can be just saving as many lives as you can. But as scarcity becomes more extreme, hard choices have to be made. You can turn to random selection (or something similar, like first-come, first served). Or you can interpret benefit more narrowly. This is what several COVID-19 triage recommendations do. ⁸ See Harris (1985); Williams (1997); or Persad, Wertheimer, and Emanuel (2009). For a critical assessment of the latter, see Kerstein and Bognar (2010). For a view that incorporates an age-weighting function, see Bognar (2015).

Ageing without Ageism?

167

For instance, the Canadian Medical Association declares that ‘in the context of a pandemic, the value of maximizing benefits is most important’ and explains that it requires that we ‘should aim both at saving the most lives and at maximizing improvements in individuals’ post-treatment length of life’. But the Association also states that maximizing improvements in post-treatment length of life—that is, maximizing life years—is merely a ‘subordinate aim’ that ‘becomes relevant only in comparing patients whose likelihood of survival is similar’.⁹ It seems that on these proposals, the recommendation is to adopt two interpretations of benefit, with little justification or explanation of why one is to be preferred over the other. But perhaps the intention is to adopt two principles for ICU admission: maximizing the number of lives saved and maximizing the amount of life saved, with the former principle taking priority over the latter—such that the second principle comes into play only when the first one cannot make a clear selection. In other words, maximizing life years, which requires taking into account life expectancy (and thus, indirectly, age), can be used only as a kind of tie-breaker. But, if you think about it, the idea of adopting the two principles with a subordinate role to maximizing life years makes little sense. To see why, return to my earlier example. Suppose you have four ventilators for the four patients A, B, C, and D. You maximize the number of lives saved, evidently, by assigning each one of the patients to a ventilator. But suppose, as in my original example, that you only have two ventilators. Maximizing the number of lives saved would now entail that you ought to select two of the patients randomly. (Assume that first-come, first-served, which is a controversial principle for many reasons, does not apply in this case. All the patients arrive at the same time.) However, suppose now that you have adopted saving the most life years as a subordinate principle. It gives you a clear recommendation: allocate the two ventilators to patients A and B. There is no tie that needs to be broken, given that you already had a principle all along to select the two patients. If you had not already adopted saving the most life years, you would have selected the two patients randomly. Maximizing the amount of life saved is not a subordinate principle to maximizing the number of lives saved. It does not merely break ties. In other words, what is different between the situation when you have four ventilators and when you have two is the degree of scarcity. You can either adopt saving as many lives as you can, in which case you would select randomly when scarcity gets sufficiently extreme, or you adopt saving as many years of life as you can, in which case you should use it regardless of the degree of scarcity—you can use it when you have four ventilators available just as well as when you have two. (If you still have ties, you can break them with random selection). Thus, it makes little sense to say that you have two principles when, in fact, only one is doing all the work. You only need the principle of maximizing the amount of life saved. ⁹ Quoted from Emanuel et al. (2020) in CMA (2020).

168

COVID-19, Age, and Rationing

Why is this point not recognized more broadly? I think, in the case of COVID19 at least, part of the explanation is the unequal distribution of mortality risk.1⁰ Suppose, for a moment, that we are faced with a pandemic in which people of all ages are exposed to the same risk of infection and death, and the ages of patients who need ICU treatment are distributed evenly. Arguably, maximizing the amount of life saved by giving priority to younger patients would be the salient moral principle. We would not demote it to a subordinate, ‘tie-breaking’ role. But, in the case of COVID19, the overwhelming majority of patients who needed ICU support (especially in the first waves) were older, and thus considerations of life expectancy and age were less salient. If that is correct, it provides an important lesson for preparing for the ethical response to the next pandemic. The following proposal emerges from these considerations. ICU admissions criteria for public health emergencies should combine non-discrimination principles and clinical prospect of success, as criteria of eligibility, with an interpretation of benefit as life years. The ethical aim of triage in such situations should be the maximization of the amount of life saved—that is, saving as many life years as possible. This entails giving priority to those whose life expectancy is greater. Since life expectancy is in part a function of age, the principle gives an indirect role to age. It will normally give priority to younger people. One objection to this proposal might be this. Take A, B, and C from my example. A’s life expectancy is 45 years, B’s is 15 years, and C’s is 10 years. If you save A, you get 45 years; if you save B and C, you get 25 years. Thus, it seems that you ought to save A. Maximizing life years may entail saving fewer lives. But this objection misunderstands the nature of the kind of situations for which the proposal applies. In triage, we try to match patients to resources. If you have only one ventilator, you would choose A. But if you have two, you would choose A and B. You would not deny resources from the additional patient. And if you have three ventilators, you would give them to all the three patients. Other things being equal, a proposal to maximize the amount of life entails maximizing the number of lives saved. There are two further reasons to interpret benefit in terms of life years and adopt the principle of maximizing the amount of life saved. First, the proposal can find some support in established ethical theories. Maximizing the amount of life saved is obviously compatible with utilitarianism. The view can also be considered as a practical implementation of whole-life egalitarianism: if the ‘currency’ to be equalized is years of life, it will lead to more egalitarian outcomes. Finally, the view can correspond to ‘life-cycle’, ‘complete life’ or ‘fair innings’ views in so far as it will tend to give priority to those who have experienced fewer stages of life. The second reason is that the proposal is supported by empirical studies of people’s moral views in resource allocation problems.11 It corresponds to widely held social

1⁰ For the distribution of COVID-19 mortality risk according to age, see Ioannidis, Axfors, and Contopoulos-Ioannidis (2020). 11 A point also made by Altman (2020).

Ageing without Ageism?

169

views (at least, in countries where we have empirical results). For instance, in a study in Australia, respondents considered saving four 20-year-olds as important as saving ten 60-year-olds; in a study in the United States, respondents thought that saving one 20-year-old is roughly equivalent to saving seven 60-year-olds; and in a study in Sweden, saving one 30-year-old was considered to be equivalent to saving thirty-four 70-year-olds.12 The proposal of maximizing the amount of life saved corresponds to these common moral judgements.

5. Conclusion In this chapter, I offered a philosophical reconstruction of the triage process for ICU admission during the COVID-19 pandemic on the basis of some of the publicly available guidelines and recommendations. I argued that we can think of a triage procedure as the combination of three components: a moral principle (typically, benefit maximization); an interpretation of benefit; and a set of criteria of eligibility for the scarce resource. I examined the role of age and life expectancy in this framework. I argued that age should not be considered a criterion of eligibility but that it can have an indirect role when benefit is interpreted as amount of life (in the form of life years). I proposed that maximizing the amount of life saved should be the ethical aim of rationing in public health emergencies. Should we go further? Should we also include quality of life, for instance, in the form of QALYs? Or should we even maximize age-weighted QALYs? Although I think these questions raise difficult moral problems, addressing them is beyond the scope of this chapter.

References Altman, Matthew C. 2020. ‘A Consequentialist Argument for Considering Age in Triage Decisions during the Coronavirus Pandemic’. Bioethics 35: 356–363. Antommaria, Armand H. Matheny, Tyler S. Gibb, Amy L. McGuire et al. 2020. ‘Ventilator Triage Policies during the COVID-19 Pandemic at US Hospitals Associated with Members of the Association of Bioethics Program Directors’. Annals of Internal Medicine 173: 188–194. ANZICS (Australian and New Zealand Intensive Care Society). 2020. ‘Guiding Principles for Complex Decision Making during Pandemic COVID-19’. ANZICS. https://www. anzics.com.au/wp-content/uploads/2020/04/ANZI_3367_Guiding-Principles.pdf. 12 See Nord et al. (1996) for the Australian study, Cropper, Aydede, and Portney (1994) for the US one, and Johannesson and Johansson (1997) for the Swedish one. See also Lewis and Charny (1989); Busschbach, Hessing, and Charro (1993); Nord (1999); Tsuchiya (1999); and Tsuchiya, Dolan, and Shaw (2003).

170

COVID-19, Age, and Rationing

Auriemma, Catherine L., Ashli M. Molinero, Amy J. Houtrow, Govind Persad, Douglas B. White, and Scott D. Halpern. 2020. ‘Eliminating Categorical Exclusion Criteria in Crisis Standards of Care Frameworks’. American Journal of Bioethics 20: 28–36. BMA (British Medical Association). 2020. ‘COVID-19—Ethical Issues’. Guidance Note, BMA. https://www.bma.org.uk/media/2226/bma-covid-19-ethics-guidance.pdf. Bognar, Greg. 2015. ‘Fair Innings’. Bioethics 29: 251–261. Bognar, Greg, and Iwao Hirose. 2022. The Ethics of Health Care Rationing: An Introduction (2nd edn). New York: Routledge. Busschbach, Jan J. V., Dick J. Hessing, and Frank Th. de Charro. 1993. ‘The Utility of Health at Different Stages in Life: A Quantitative Approach’. Social Science & Medicine 37 (2): 153–158. CCSSA (Critical Care Society of Southern Africa). 2020. ‘Allocation of Scarce Critical Care Resources during the COVID-19 Public Health Emergency in South Africa’. CCSSA, 2 April. https://criticalcare.org.za/wp-content/uploads/2020/04/ Allocation-of-Scarce-Critical-Care-Resources-During-the-COVID-19-PublicHealth-Emergency-in-South-Africa.pdf. CMA (Canadian Medical Association). 2020. ‘Framework for Ethical Decision Making during the Coronavirus Pandemic’. CMA. https://policybase.cma.ca/en/permalink/ policy14133. Craxì, Lucia, Marco Vergano, Julian Savulescu, and Dominic Wilkinson. 2020. ‘Rationing in a Pandemic: Lessons from Italy’. Asian Bioethics Review 12: 325–330. Cropper, Maureen L., Sema K. Aydede, and Paul R. Portney. 1994. ‘Preferences for Life Saving Programs: How the Public Discounts Time and Age’. Journal of Risk and Uncertainty 8: 243–265. Ehni, Hans-Jörg, Urban Wiesing, and Robert Ranisch. 2021. ‘Saving the Most Lives— A Comparison of European Triage Guidelines in the Context of the COVID-19 Pandemic’. Bioethics 35: 125–134. Emanuel, Ezekiel J., Govind Persad, Ross Upshur et al. 2020. ‘Fair Allocation of Scarce Medical Resources in the Time of COVID-19’. New England Journal of Medicine 382 (21): 2049–2055. Harris, John. 1985. The Value of Life. London: Routledge & Kegan Paul. Ioannidis, John P. A., Cathrine Axfors, and Despina G. Contopoulos-Ioannidis. 2020. ‘Population-Level COVID-19 Mortality Risk for Non-Elderly Individuals Overall and for Non-Elderly Individuals without Underlying Diseases in Pandemic Epicenters’. Environmental Research 188: 109890. Jöbges, Susanne, Rasita Vinay, Valerie A. Luyckx, and Nikola Biller-Andorno. 2020. ‘Recommendations on Covid-19 Triage: International Comparison and Ethical Analysis’. Bioethics 34: 948–959. Johannesson, Magnus, and Per-Olov Johansson. 1997. ‘Is the Valuation of a Qaly Gained Independent of Age? Some Empirical Evidence’. Journal of Health Economics 16: 589–599.

Ageing without Ageism?

171

Kerstein, Samuel J., and Greg Bognar. 2010. ‘Complete Lives in the Balance’. American Journal of Bioethics 10: 37–45. Lewis, P. A., and M. Charny. 1989. ‘Which of Two Individuals Do You Treat When Only Their Ages Are Different and You Can’t Treat Both?’ Journal of Medical Ethics 15: 28–32. Marckmann, Georg, Gerald Neitzke, Jan Schildmann et al. 2020. ‘Decisions on the Allocation of Intensive Care Resources in the Context of the COVID-19 Pandemic: Clinical and Ethical Recommendations of DIVI, DGINA, DGAI, DGIIN, DGNI, DGP, DGP and AEM’. Medizinische Klinik—Intensivmedizin und Notfallmedizin 115: 115–122. Nord, Erik. 1999. Cost–Value Analysis in Health Care: Making Sense Out of Qalys. Cambridge: Cambridge University Press. Nord, Erik, Andrew Street, Jeff Richardson, Helga Kuhse, and Peter Singer. 1996. ‘The Significance of Age and Duration of Effect in Social Evaluation of Health Care’. Health Care Analysis 4: 103–111. Persad, Govind, Alan Wertheimer, and Ezekiel J. Emanuel. 2009. ‘Principles for the Allocation of Scarce Medical Interventions’. The Lancet 373 (9661): 423–431. SAMW (Swiss Academy of Medical Sciences). 2020. ‘COVID-19 Pandemic: Triage for Intensive-Care Treatment under Resource Scarcity’. SAMW, revised version 3.1. https:// www.samw.ch/de/Ethik/Themen-A-bis-Z/Intensivmedizin.html. Tsuchiya, Aki. 1999. ‘Age-Related Preferences and Age Weighting Health Benefits’. Social Science & Medicine 48: 267–276. Tsuchiya, Aki, Paul Dolan, and Rebecca Shaw. 2003. ‘Measuring People’s Preferences Regarding Ageism in Health: Some Methodological Issues and Some Fresh Evidence’. Social Science & Medicine 57: 687–696. Vergano, Marco, Guido Bertolini, Alberto Giannini et al. 2020. ‘SIAARTI Recommendations for the Allocation of Intensive Care Treatments in Exceptional, Resource-Limited Circumstances’. Minerva Anestesiologica 86 (5): 469–472. White, Douglas B., and Bernard Lo. 2020. ‘A Framework for Rationing Ventilators and Critical Care Beds during the COVID-19 Pandemic’. Journal of the American Medical Association 323 (18): 1773–1774. Williams, Alan. 1997. ‘Intergenerational Equity: An Exploration of the “Fair Innings” Argument’. Health Economics 6: 117–132.

12 Ageism in Assisted Reproduction Francesca Minerva

1. Women and reproduction: Legal, social, and biomedical advancements Over the past century, women in the Western world have taken more control over their bodies as well as over their sexual and reproductive lives. These changes were prompted by societal and cultural developments as well as by medical and technological advancements. For instance, feminism and the sexual liberation movement, among other factors, highlighted the distinction between sex and reproduction, between being a woman and being a mother. New legal measures (such as the legalization of abortion) and new medical tools (such as the contraceptive pill and assisted reproduction technologies—ARTs) made it possible to achieve a separation between sex and reproduction that allows women to both have sex without getting pregnant and to get pregnant without having sex. The possibility of choosing when to start a pregnancy has granted women more freedom to plan their lives according to their preferences, rather than according to the physiological processes they happen to experience. For instance, contraceptives have made it possible to postpone or avoid pregnancy and therefore to invest more time in professional or personal projects rather than in child rearing; on the other hand, ARTs have helped women to get pregnant at an age when female fertility has declined. Although a lot of progress has been made, and woman enjoy more reproductive freedom and bodily autonomy than in the past, they still experience major reproductive disadvantages compared to men. In particular, women older than 40 have much less reproductive freedom than their male counterparts. The current debate on female disadvantages is focused on figuring out how society can increase equality between genders through, among other things, interventions aimed at supporting working mothers or at preventing the penalization of pregnant women in the workplace. These are surely issues worth considering since they have a major impact on women’s work–life balance and on the overall quality of their lives. However, there are some features of female biology that constitute a disadvantage regardless of the social, economic, and cultural context one lives in. Such disadvantages cannot be alleviated by any social or cultural intervention but only by medical and technological ones. In this chapter, I focus on the conjunction of two biological Francesca Minerva, Ageism in Assisted Reproduction. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Francesca Minerva (2023). DOI: 10.1093/oso/9780192894090.003.0013

Ageing without Ageism?

173

aspects of being a woman that are inherently disadvantageous: reproduction and ageing.1

2. Legal limits to reproductive freedom Female fertility declines at a much faster pace than male fertility. Female fertility peaks in the early twenties and starts to decline after the age of 27, speeding up quite rapidly between the ages of 35 and 40. Around the age of 50, women hit menopause, and by the age of 60 it is extremely rare for a woman to get pregnant naturally (van Kooij et al. 1996; Spandorfer et al. 2007). In contrast, male fertility starts to decline after the age of 40, but men never become completely infertile as they keep producing sperm throughout their adult life. Aside from the biological obstacles to getting pregnant, older women can face a number of legal obstacles; in particular, restrictions that make it difficult for them to even access ARTs. In 18 (out of 33) European countries, including Belgium, Denmark, and the Netherlands, access to ARTs for women between 45 and 51 is denied by law, meaning that even when such women are willing to pay privately for ARTs, they are not allowed to use them. However, these same countries don’t put any legal restriction on male access to ARTs, even when men are older than 45. Only Portugal imposes an upper age limit for men who want to access ARTs (the age limit being 60 rather than 45) and two countries, Finland and Sweden, recommend against providing ARTs to men older than 56 (but this recommendation is not legally enforced). The reason behind the upper age limits for female access to ARTs is not explicitly stated by the legislators. But, at least in the case of the Italian law regulating access to in vitro fertilization (IVF) (law 40/2004), we know that the upper age limit of 51 was chosen because, at the time this law was approved, the oldest woman known to have conceived naturally was 51 years old. (In 2016, Mrs Venuso gave birth to a healthy baby she had conceived without the use of IVF at the age of 61 (Iuliano 2016), so logically the Italian law should modify the upper age limit up to 61.) Of course, the fact that something happens naturally does not imply that it is morally good or should be legally permissible, so in the following paragraphs the main arguments against pregnancy at advanced age are discussed.

3. Arguments against allowing older women to access ARTs The main objections to allowing the use of ARTs by older women are based on the assumptions that late pregnancies can damage the pregnant woman and/or the prospective child. 1 This is not the same as saying that there are no aspects of male biology that are disadvantageous or that women are overall more biologically disadvantaged than men.

174

Ageism in Assisted Reproduction

Let us first consider the objection based on medical concerns for the mother’s health. According to Caplan and Patrizio (Caplan and Patrizio 2010), for instance, older women should not be allowed to access ARTs because pregnancy in older women can be risky: it has been proved that older women are more likely to suffer from health issues during pregnancy, and they often suffer from gestational diabetes or hypertension (Klemetti et al. 2011). Older women who get pregnant at an advanced age should be adequately informed about the health risks that the pregnancy poses to them and should be allowed to terminate it if they believe that the pregnancy could be too dangerous for them. But once they have been adequately informed and have autonomously decided to continue with the pregnancy or to continue with ARTs, they should be allowed to do what they believe is in their best interest. If they decide that, overall, becoming a mother is worth the risks, perhaps because they believe that the psychological benefit of becoming a mother outweighs the potential health problems, there is no reason why they shouldn’t be allowed to do what they argue to be in their best interest. We don’t usually prevent people from engaging in potentially dangerous activities such as parachuting, nor do we prevent them from refusing treatments that could save their life (patients are usually allowed to refuse a treatment). One of the pillars of the Western health-care system is respect for the patient’s autonomy, even when this means that doctors have to let a patient die of conditions that could be easily cured (as happens when a Jehovah’s witness refuses a blood transfusion) or have to help them to die (as happens in countries where euthanasia and assisted suicide are legal). Given the importance that, as a society, we attribute to the patient’s autonomy and self-determination, it is hard to understand why women over the age over 51 (in some countries, over the age of 45) are so often thwarted in their attempts to access reproductive treatments. One possible explanation is that the main concerns are not raised by the health risks of the woman but rather by the best interest of the prospective child, as discussed below. Numerous studies have suggested that advanced maternal age increases the risk of both chromosomal (such as trisomy 21) and non-chromosomal abnormalities (such as cardiac and renal problems) (Hollier et al. 2000). However, recent studies have challenged these results with respect to non-chromosomal conditions (Frederiksen et al. 2018): For instance, a study found that advanced maternal age ‘is associated with an overall decreased risk for major anomalies’ (Goetzinger et al. 2017) and another study found ‘no association between the incidence of birth defects of unknown aetiology and advancing maternal age’ (Baird et al. 1991). We need more research in order to find conclusive evidence, but, until such evidence is provided, it is unwarranted to argue that older women shouldn’t have children because of the much higher rate of health issues for the prospective baby. Chromosomal abnormalities, such as trisomies 21, 18, and 13, do seem to be linked to advanced maternal age (Loane et al.2013), and the possible increase in the total number of children born with such chromosomal abnormalities is often cited in the case against late motherhood.

Ageing without Ageism?

175

Before prenatal screening became available, there was no chance of detecting chromosomal abnormalities before birth and, in some cases, for a long time after birth. Nowadays, women of all ages have access to genetic screening and testing that can accurately detect most chromosomal abnormalities by, at the latest, the twentieth week of pregnancy. For women who use assisted reproduction, pre-implantation tests are usually available so that at least some health issues can be detected before the embryo transfers. Indeed, although the average maternal age has increased consistently over the past 30 years, the percentage of children born with chromosomal conditions has not gone up, thanks to pre-implantation screening in the case of IVF and pre-natal screening in the case of natural conception. One could argue that, regardless of the availability of pre-natal screening, the higher risk of producing a child with a chromosomal abnormality should be enough reason to prevent older women from reproducing through IVF, not only because it is in the best interests of prospective children to be born without a disability but also because society has an interest in having a healthy population. This argument, though, is used quite inconsistently because people with congenital conditions such as deafness, blindness, and dwarfism are not prevented or deterred from reproducing even when, given the high hereditability of their conditions, it is very likely that their children will inherit the disability. If people with highly hereditable congenital conditions can have access to ARTs, then there is no reason why women older than 45 or 51 shouldn’t have access to ARTs as well. Moreover, many people with disabilities don’t consider their disability something that makes their lives not worth living, and, indeed, they are often not concerned about passing it on to their children. It is beyond the scope of this chapter to discuss how different disabilities might affect the quality of life of different people, but it is worth noticing that the issue of disability is not as decisive as often assumed, as suggested by the recent developments in disability studies (Barnes 2016). Another argument against allowing older women to have children is based on the idea that young parents are more fit to take care of children because they have more energy. However, the best evidence available doesn’t fully support this hypothesis, at least when mental energy is concerned: for instance, a study found that women who have children in their 50s through ARTs do not report lower levels of physical and mental functioning compared to younger women and display lower levels of parental stress, leading the authors to conclude that there continues to be a lack of data justifying the restriction of access to ART services based on maternal age alone (Steiner and Paulson 2007). Another study found that older maternal age was associated with fewer behavioural, social, and emotional difficulties in children at 7 and 11 years of age but not at age 15. Older mothers reported less frequent use of physical sanctions at 7 and 11 years of age and less use of verbal sanctions at 7, 11, and 15 years of age (Trillingsgaard and Sommer 2018). So, it seems that older mothers have a more relaxed approach to parenting, perhaps because, as suggested by another study, parents under 35 experience a long-term drop

176

Ageism in Assisted Reproduction

in happiness and well-being after having children, but the same doesn’t happen to those having children after the age of 35 (Myrskylä and Margolis 2014) . A study using a large sample of Swedes born between 1960 and 1991 concluded that individuals born to older mothers, including those at the oldest ages, are taller, remain longer in the educational system, are more likely to attend university, and perform better on standardized tests than their siblings who were born when their mothers were younger (Barclay and Myrskylä 2016). Let us now consider the objection to advanced-aged maternity that children born to older women could become orphans at a young age. It is true that all of us would like to have our parents around for as long as possible and that the younger the parents, the more likely it is that they will accompany us for longer. However, age is not the only factor to take into account when considering the potential lifespan of a person. People with risky jobs (such as fire-fighters or formula one drivers) or dangerous hobbies (such as parachuting), and those affected by conditions (such as Huntington’s disease) that shorten one’s life-span are also more likely to leave their children without a parent at a relatively young age. So, if the objection concerning the possibility of leaving a child without parents were as strong as opponents of IVF for older women claim, then similar bans on reproduction should be imposed on people with risky jobs and hobbies and on those affected by conditions that are likely to shorten their life span. Since these people are not legally prevented from accessing ARTs, it seems that the ban on access to ARTs for women older than 45–51 cannot be justified using this argument. But let’s assume, for the sake of argument, that despite all this evidence suggesting that children born to older parents are at least as well off as those born to younger ones, there are still good reasons to believe that having younger parents is better than having older parents. Even if it were better, overall, for children to have young parents, the counterfactual for the child born to older parents is not being born to younger parents but rather not being born at all. So, for example, if my parents had had a child 10 years before they had me, that child would be a different person, not an older version of me. And, most likely, I would have not been born at all. So, even if it would have been better for any generic child of my parents to be born to younger parents, the argument doesn’t apply to a specific child my parents had, that is, my sister or me. Since the child they didn’t conceive 10 years before they conceived me doesn’t exist, and people who don’t exist have no interests (including an interest in existing), nobody has been harmed by my parents’ decision to have the two specific children they brought into existence, although perhaps other children would have had a better life because they were born to younger parents.2 Thus, arguments in favour of upper age limits would need to prove that the quality of life of people born to older parents is so low that it would have been better for these people not to be born at all. The burden of proof is on them since, even if there were evidence supporting the claim that people born to younger parents have 2 For a more detailed discussion of this argument based on the non-identity problem, see Parfit (1984).

Ageing without Ageism?

177

a higher quality of life overall (and we have shown there is very little empirical evidence supporting this view), there is no reason to assume that not being brought into existence is better than being brought into existence by older parents.

4. Why is this a problem worth discussing? One might object that the legal restrictions to accessing ARTs for women aged over 45 is not an issue worthy of consideration. After all, the large majority of women have children in their late twenties (in Europe, the average age at which women have their first child is 29), so the percentage of women trying to conceive in their forties and fifties (or at an even older age) is quite small. Given that there aren’t many women in their late forties or fifties struggling to get pregnant, the argument goes, the fact that they are denied access to ARTs can hardly be considered a problem requiring ethical discussion and, even less, legal intervention. I suggest that this is not the only way to look at this situation: the fact that fewer women try to conceive in their forties and fifties than in their twenties and thirties cannot be taken as proof that all or most women actually want to have children in their twenties and thirties, and that they would not make different choices if their fertility window was wider than it currently is. Indeed, there is some evidence that an increasing number of women want to have children at a later time. For instance, the fact that the percentage of UK women having children in their forties has steadily increased over the years3 could support the hypothesis that the desire to have children doesn’t necessarily decline at the same pace as fertility does. It is often assumed that most women have children before they turn 35 because they have an inbuilt clock that makes them want to have children before it’s too late. However, there are other plausible explanations why women usually reproduce before the age of 35. It is possible, for instance, that rather than developing a sudden desire for a baby, women feel urged to get pregnant because they are aware of their incipient infertility, but they would choose to postpone becoming a mother if their fertility window were wider. It is possible that if women had the option of having more time to devote to work (or hobbies, or personal achievements, or even just fun) without fear of having to give up on becoming mothers altogether, they would actually choose to wait for longer before having children. I am not suggesting, here, that mothers and fathers always suffer negative professional or personal consequences because of their decision to have children. I am just making the simple observation that children require time, energy, and money and that their needs cause the shifting of these resources away from the parents towards the children for at least a few decades after birth (when parents become old and the children are mature, it is often the case that 3 From the Office for national Statistics (UK). https://www.ons.gov.uk/peoplepopulationand community/birthsdeathsandmarriages/conceptionandfertilityrates/bulletins/conceptionstatistics/ 2016#conception-rates-decreased-at-all-ages-except-for-women-aged-40-and-over.

178

Ageism in Assisted Reproduction

the resources are shifted back from the children towards the parents). If we look at the overall lifetime of an individual, it becomes apparent that, the longer one can postpone the moment when they invest such resources in children, the longer they can work towards other goals. I am also not arguing, here, that everybody should have goals other than raising children or that having the raising of children as one’s primary goal is a bad thing or worse than having other priorities. I am instead making the much more modest claim that having the option of investing more of one’s youth in goals other than parenthood is a good thing for people who have this preference and possibly for the children and society at large as well if this results in happier parents and children.

5. Medical solutions: Social freezing, delaying menopause, and ectogenesis So far, I have argued that we should remove the upper legal age limits that prevent women older than 45 from accessing ARTs (Goold 2005; Caplan and Patrizio 2010; Smajdor 2011; Ekberg 2014). Although removing these legal limits is a necessary first step to reducing ageist discrimination against women, the change would not, by itself, alleviate the disadvantage experienced by women because of their fast decline in fertility. If upper age limits to access ARTS after 45 were removed, the chances of older women getting pregnant after 45 (without using an egg donor) would still be small: removing the upper age limit would improve the situation of the 1–2 per cent of women who can get pregnant through IVF (using their own eggs) after 45 and of those who are willing to use an egg donor. However, this would not solve the problem for the large percentage of women (98–99 per cent) who can’t become pregnant after 45 unless they use a donor’s eggs. In order to tackle this problem, we need to look at medical and technological tools that can either alter biology (for instance, by tackling menopause), or obviate the need for interventions at older ages (for instance, by providing social freezing for free to everyone).

5.1 Social freezing Cryopreservation of oocytes is a relatively new medical tool that allows a woman to preserve eggs for an indefinite amount of time so that if, for any reason, she doesn’t have eggs at a later point in her life, she won’t need to use an egg donor in order to conceive through IVF. Couples can also cryopreserve embryos, but in most European countries this practice is allowed only if an IVF cycle produced more embryos than could be implanted; the spare ones are then cryopreserved to be implanted at a later time.⁴ Oocytes, sperm, and embryo cryopreservation can all be considered ⁴ In many European countries, couples are not allowed to create embryos with the goal of cryopreserving them for later use, whereas in some parts of the United States this form of cryopreservation is legally permitted.

Ageing without Ageism?

179

forms⁵ of ‘social freezing’, that is, technological interventions aimed at giving time to people who don’t want to, or can’t, procreate immediately but do want to increase their chances of procreating at some future time. Cryopreservation of eggs and embryos is quite expensive and, given that it is not subsidized by public health systems (in contrast to IVF treatments for infertility below a certain maternal age, which often are subsidized), the entire cost is paid privately. Since such treatments cost between 3,000 and 10,000 euros, it is not surprising that many people cannot afford it. Some big tech companies, such as Google and Facebook, have started offering social freezing benefits to their female employees, but this policy was not warmly received. For instance, it has been pointed out that female employees might be ‘encouraged’ (or perhaps subtly compelled) by their employer to freeze their eggs and postpone pregnancy even if it was in their interest to have children early in life. There is a general preoccupation that social freezing would make women less autonomous and more vulnerable to the expectations of their employers, who have an interest in women focusing on their career rather than child-rearing (Harwood 2009). Technical solutions to social problems like social freezing are often opposed on the basis that they might end up delaying or completely crowding out other solutions. Highly desirable jobs (usually those that are highly paid and confer a high social status) can be extremely difficult to get, and people competing for these jobs often have to deal with long office hours, no holidays, and no distractions. Having a baby while competing against many other qualified candidates is not likely to help women to get to the top of their law firm or company. This is in part due to lack of adequate social resources to support parents to take care of children, but it is also in part due to the fact that women have to carry a foetus for nine months and that if they decide to breastfeed, they will be their child’s only source of nutrition for about six months. This means that for at least a few months, they will need to work less, and, unfortunately, this can have a negative impact on their career. Although society can do a lot more than it currently does to help women, it can’t remove the burden of pregnancy and breastfeeding. However, if women were encouraged to use social freezing, they could postpone having children until the moment after they obtained the job they want, which they are more likely to get if they don’t take a break for a pregnancy. It seems deeply unfair that women are penalized in the workplace for something we all, as a society, have a strong interest in, that is, creating the next generation. But claiming that the problem of women having to choose between work and motherhood is exclusively a social problem and therefore that it should only be addressed through social strategies, is not going to help. Denying that parenthood is also tougher on women for biological reasons is not going to solve the problem, and it will divert societal resources away from finding a solution to the disadvantages women face (Firestone 1970 [2000]). Instead of looking at social freezing as if it were just another means to oppress women and erode their autonomy, we should welcome it as a tool that can help ⁵ Medical freezing, instead, refers to the choice of freezing gametes or embryos for people who are about to start treatments, such as chemotherapy, that might affect their future fertility.

180

Ageism in Assisted Reproduction

women to overcome the disadvantage of having a reproductive system that ages too quickly (Goold and Savulescu 2009). Indeed, society should provide all women with free social freezing in their early twenties so that they can secure a large number of high-quality eggs for the future. With this ‘biological insurance’ against the passing of time, they would be free to choose the path they prefer: they might choose to have children when they are young, or they might choose to have children in their forties or even later. Respecting women’s autonomy also implies not assuming that their desires are always exclusively the result of social pressure and almost never of their free choice. Moreover, denying any biological contribution to the disadvantages women face, especially when it comes to older women, won’t help either women or society: women will not have what they really want, and society will waste resources on measures that can’t really achieve what they are trying to achieve.

5.2 Delaying menopause Another medical intervention that could help women overcome their biological disadvantages would be developing treatments that can delay menopause. Women are born with all of their egg follicles (about 300,000 per ovary), and through their lives they lose about 1,000 follicles per month, so that by the age of 40 only a small percentage of these follicles are still available. The rapid loss of follicles is the reason why women’s fertility decreases and eventually they become infertile around the age of 50. Although a treatment aimed at delaying or eliminating menopause could have an enormous positive impact on women, there have been no breakthroughs in this area. It’s not unreasonable to suspect that the lack of research is due not only to how difficult it would be to achieve the desired results but also to the fact that menopause (and ageing in general) are not seen as medical problems in the first place. On the one hand, the fact that menopause is a natural phenomenon may explain why medical research has not taken any interest in it, as many see natural phenomena as intrinsically good. On the other hand, the fact that ageing and menopause don’t raise concerns related to injustice or inequality may explain why philosophers have not taken much interest in them. However, there is no reason to believe that something natural that happens to everyone, regardless of race, social class, or education is not morally problematic. There are some things, such as ageing and dying, that are inherently bad even though they are natural and affect everyone. Indeed, if anything, there is something especially tragic about these common and ineluctable experiences since there is no way anybody can escape them, and therefore they lower the wellbeing of every person who has not died young (which is, of course, yet another kind of tragedy). Natural and ‘egalitarian’ problems such as ageing and menopause should be tackled by medical research at the very least because of their psychological and medical side effects. Consider menopause: if the number of follicles that get lost every month

Ageing without Ageism?

181

were simply halved, women would stay fertile for much longer, and they probably wouldn’t need to rely on ARTs in order to conceive at an older age. Not only would they be free from the anxiety induced by the clicking of the biological clock, but they would also be spared all of the other health issues normally associated with menopause (which, for some women, include osteoporosis, osteoarthritis, hair loss, weight gain, insomnia, depression, and anxiety). It seems that research aimed at pausing or reversing menopause could be one of the greatest revolutions in the history of reproductive medicine, yet it’s a very underdeveloped area of research. One could argue that the kind of radical change to women’s biology I am suggesting would be yet another way for ‘the patriarchy’ to exploit the bodies of women well beyond the age to which they are currently exploitable. If it were possible for women to procreate well past their current reproductive age, they would perhaps be expected to have more children or to have children later on in their life, even if their actual preference would be to have them at a young age. It is, of course, possible that many of the desires stemming from the development of new technologies are not ‘genuine’ desires but rather by-products of societal pressure and technological advancement. In this sense, the desire to have children at increasingly older ages could be influenced by new expectations set up by the new technologies themselves. In general, it is very hard to distinguish between a genuine desire from a ‘societally induced’ one, and separating them sharply might not be possible. We are social animals—we live in relation to other people, and we are necessarily sensitive to the preferences and expectations of others. However, when social pressure is kept at a reasonably low level, when individual autonomy is promoted and respected as much as possible, and education is sufficiently widespread, giving people more options is usually a good thing.

6. Conclusion In this chapter, I have argued that upper age limits to female access to ARTs should be removed so as to allow women older than 51 to use IVF in order to become mothers. I have also argued that new and future medical tools, such as social freezing and treatments aimed at delaying menopause, should be introduced and provided for free in order to solve the problem of the quick loss of fertility in women. If we want to achieve equality between men and women, we need to start tackling not only disadvantages brought about by social factors but also those stemming from biological ones.

References Baird, Patricia A., A. Dessa Sadovnick, and Irene M. Yee. 1991. ‘Maternal Age and Birth Defects: A Population Study’. The Lancet 337 (8740): 527–530.

182

Ageism in Assisted Reproduction

Barclay, Kieron, and Mikko Myrskylä. 2016. ‘Advanced Maternal Age and Offspring Outcomes: Reproductive Aging and Counterbalancing Period Trends’. Population and Development Review 42(1): 69–94. Barnes, Elizabeth, 2016. The minority body: A theory of disability. Oxford University Press. Caplan, Arthur L., and Pasquale Patrizio. 2010. ‘Are You Ever Too Old to Have a Baby? The Ethical Challenges of Older Women Using Infertility Services’. Seminars in Reproductive Medicine 28 (4): 281–286. Ekberg, Merryn Elizabeth 2014. ‘Assisted Reproduction for Postmenopausal Women’. Human Fertility 17 (3): 223–230. Firestone, Shulamith 1970 [2000]. The Dialectic of Sex: The Case for Feminist Revolution. New York: William Morrow and Company. Frederiksen, Line Elmerdahl, Andreas Ernst, Nis Brix et al. 2018. ‘Risk of Adverse Pregnancy Outcomes at Advanced Maternal Age’. Obstetrics & Gynecology 131 (3): 457–463. Goetzinger, Katherine R., Anthony L. Shanks, Anthony O. Odibo, George A. Macones, and Alison G. Cahill 2017. ‘34: Advanced Maternal Age and the Risk of Major Congenital Anomalies: Survival of the Fittest?’ American Journal of Obstetrics & Gynecology 210 (1): S23. Goold, Imogen 2005. ‘Should Older and Postmenopausal Women Have Access to Assisted Reproductive Technology?’ Monash Bioethics Review 24 (1): 27–46. Goold, Imogen and Julilan Savulescu. 2009. ‘In Favour of Freezing Eggs for Non-Medical Reasons’. Bioethics 23 (1): 47–58. Harwood, Karey 2009. ‘Egg Freezing: A Breakthrough for Reproductive Autonomy?’ Bioethics 23 (1): 39–46. Hollier, Lisa M., Kenneth J. Leveno, Mary Ann Kelly, Donald D. MCIntire, and F. Gary Cunningham. 2000. ‘Maternal Age and Malformations in Singleton Births’. Obstetrics & Gynecology 96 (5): 701–706. Iuliano, Lorenzo 2016. ‘Io, mamma in Campania felice a 61 anni: Non ho paura’. Il Mattino, 21 September. https://www.ilmattino.it/caserta/mamma_campania_61_ anni_intervista_castel_volturno-1978437.html. Klemetti, Reija, Jennifer J. Kurinczuk, and Maggie Redshaw. 2011. ‘Older Women’s Pregnancy Related Symptoms, Health and Use of Antenatal Services’. European Journal of Obstetrics & Gynecology and Reproductive Biology 154 (2): 157–162. Loane, Maria, Joan K. Morris, Marie-Claude Addor et al. 2013. ‘Twenty-Year Trends in the Prevalence of Down Syndrome and Other Trisomies in Europe: Impact of Maternal Age and Prenatal Screening’. European Journal of Human Genetics 21 (1): 27–33. Myrskylä Mikko, and Rachel Margolis. 2014. ‘Happiness: Before and After the Kids’. Demography 51 (5): 1843–1866. Parfit, Derek 1984. Reasons and Persons. Oxford: Oxford University Press. Smajdor, Anna 2011. ‘The Ethics of IVF over 40’. Maturitas 69 (1): 37–40.

Ageing without Ageism?

183

Spandorfer, Steven D., Kristin Bendikson, Kate Dragisic, Glenn Schattman, Owen K. Davis, and Zev Rosenwaks. 2007. ‘Outcome of In Vitro Fertilization in Women 45 Years and Older Who Use Autologous Oocytes’. Fertility and Sterility 87 (1): 74–76. Steiner, Anne Z., and Richard J. Paulson. 2007. ‘Motherhood after Age 50: An Evaluation of Parenting Stress and Physical Functioning’. Fertility and Sterility 87 (6): 1327–1332. Trillingsgaard, Tea, and Dion Sommer. 2018. ‘Associations between Older Maternal Age, Use of Sanctions, and Children’s Socio-Emotional Development through 7, 11, and 15 Years’. European Journal of Developmental Psychology 15 (2): 141–155. van Kooij, Roelof J., Caspar WN Looman, Johannes DF Habbema, Marinus Dorland, and Egbert R. te Velde. 1996. ‘Age-Dependent Decrease in Embryo Implantation Rate after In Vitro Fertilization’. Fertility and Sterility 66 (5): 769–775.

13 An Education Resource Account for Early School Leavers Andrée-Anne Cormier and Harry Brighouse

The age at which school ceases to be compulsory varies among jurisdictions. In England, it is legal to leave school at 16, though until 18 one is required to stay in full-time education, for example, at a college; start an apprenticeship or traineeship; or spend 20 hours or more a week working or volunteering while in part-time education or training. The limited availability of apprenticeship and training opportunities makes school the default. In Wales, no such conditions apply: 15-year-olds can legally leave school on the last Friday in June if they will turn 16 before the next school year begins. In California, school is compulsory until the age of 18, with the exception that students can leave after they turn 16 as long as they have passed the California High School Proficiency Exam and their parents consent to them leaving. In Canada, the age is 16 or 18, depending on the province. In Argentina, Germany, Israel, and the Netherlands, schooling of some kind is compulsory until the age of 18. In this chapter, we defend the view that schooling should not be compulsory past the age of 16. We offer two arguments. First, requiring reluctant 16–18-year-olds to stay in school is more costly than it is beneficial in terms of the relevant values at stake. Second, there are good reasons to legally allow 16-year-olds to make their own decisions about whether to stay in the formal education system. But we anticipate— indeed, hope for—a rather obvious objection: that students who leave school at 16 thereby forego the considerable direct investment in the development of their human capital that is tied to participation in formal schooling. Governments invest universally in human capital development via formal schooling up to the age at which students either leave legally or drop out; after that, investment is selective and patchy in most countries. However, many students who choose to leave school before 18 will be ready to take advantage of formal education later—perhaps not even very much later—in their lives. We argue that those students are entitled to ear-marked funding to complete their formal schooling or to enrol in work training programmes at any age. More specifically, we propose that an education resource account (ERA) be established for all those students.

Andrée-Anne Cormier and Harry Brighouse, An Education Resource Account for Early School Leavers. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Andrée-Anne Cormier and Harry Brighouse (2023). DOI: 10.1093/oso/9780192894090.003.0014

Ageing without Ageism?

185

1. Mandatory schooling until age 16 and an educational resource account Our policy proposal is twofold: (1) End mandatory schooling at 16: Reduce the age at which school attendance ceases to be legally required to 16. Students would be permitted to cease attending school at the end of the school year in which they turn 16. For students who wish to complete secondary school, they could do so. (2) Establish an ERA to pay for further education: Establish an inflation-linked ERA for each student who decides to leave high school at 16 or 17 and without graduating.1 The funds in the account would amount to 100 per cent of average per-pupil spending, pro-rated for the number of school days prior to their expected graduation date that they stop attending school. The fund could be used for tuition at accredited educational institutions at any time in the student’s life. The age of 16 is, to some extent, an arbitrary choice. We are open to the idea that, even in wealthy countries, schooling should be optional even for some students younger than 16. But 16 is now the lowest age at which schooling is compulsory in most wealthy democracies.2 We want to argue for this as the maximum limit: schooling should never be compulsory above that age. For many jurisdictions, including Germany, parts of Canada, California, England, but not Wales or the Netherlands, observing our limit would require a change in the law. Of course, even in jurisdictions where students are currently required to stay in school until the age of 18, not all do so. Actual drop-out rates are difficult to monitor accurately, but in California and Wisconsin, both states which require school attendance until 18, the graduation rates were 82.7 per cent and 88.6 per cent, respectively for 2016–2017.3 As things stand, considerable resources are expended attempting (with varying degrees of success) to keep the non-graduating 17.3 per cent or 11.4 per cent of students in school. On our proposal, some of these resources would, instead, be saved towards educational spending when those students, voluntarily, return to formal education to which, as things stand, they are not entitled. We shall present two distinct but complementary kinds of argument for the first part of the proposal. In section 2, we argue that, for some non-trivial number of students, formal schooling is not well calibrated to their needs when they are 16–18 and that this problem is not easily solvable and generates significant costs. In section 3, 1 See David Miliband (1990) for a previous version of the ERA proposal, which inspired this one. The ERA should be distinguished from a basic capital grant, which would provide all young adults with a cash grant with which they could do whatever they wanted. See Ackerman and Alstott (1999); Ackerman et al. (2006). Pikkety (2020) also endorses a version of the basic capital grant. In contrast with the capital grant proposal, the ERA earmarks money for a particular purpose, thus resembling more closely Anne Alstott’s proposals for a caregiver resource account (Alstott 2005). 2 It also seems to be the threshold of choice for advocates of lowering the voting age. 3 See World Population Review (2021).

186

An Education Resource Account for Early School Leavers

we argue for allowing 16-year-olds to make their own decisions. Despite that, reasons relating to both the public good and fairness support reserving funds so that students can take up formal schooling later in their lives. In section 4, we present those reasons. Bear in mind that we are offering this proposal as part of what philosophers now tend to call non-ideal theory.⁴ In non-ideal theory, as opposed to ‘ideal theory’, the theorist accepts various unjust, or otherwise non-ideal, features of societies as parameters which they are not targeting for criticism or change, even though they may be critical of them and want them to change. To take the obvious point: we believe that a good deal of the material inequality that characterizes the societies for which we are making this proposal is unjust. But we are assuming that it will not change much, and we are not assuming that our proposal will eliminate it or even significantly ameliorate it. We abstain from comment about what an education system would be like in a society that was distributively perfectly just. The proposal is a way of improving, from the point of view of justice, the education system within societies that are unjust on various dimensions.

2. An argument from efficiency (broadly conceived) This part of our argument turns on an empirical conjecture. Forcibly keeping students in school past the age of 16 is inefficient. Managers and teachers exert considerable time, energy, and effort to keep over-16-year-olds who would otherwise leave in school. All things considered, it would be better to expend that time, energy, and effort on the learning of others. In saying that a higher school leaving age is inefficient, we are not suggesting that economic efficiency, narrowly understood, is the only criterion for making education policy decisions. The just distribution of burdens and benefits should be the central goal of government action. But achieving it efficiently is better than achieving it inefficiently. We conjecture that, as long as something like an ERA is in place, reducing the school leaving age would free up resources to spend on educating others, while not making the leavers, on average, educationally worse off. The morally relevant costs should not be understood in simple financial terms. Costs and benefits should be understood in terms of students’ learning, effects on the school ethos, time, energy, and the psychological burdens associated with managing students who do not want to be there, as well as schools’ ability to attract and retain highly competent teachers. We articulate the reasons for making the empirical conjecture below, but we can only vindicate it by appeal to limited evidence from social science. That is because existing, high-quality social science provides only a fraction of the morally relevant data we would need—at least, in the main languages that we read (English and French). ⁴ On the distinction between ideal and non-ideal theory, see, e.g. Stemplowska and Swift (2012).

Ageing without Ageism?

187

The group of students who would, currently, choose to leave school if they were no longer required to be there after age 16 is heterogeneous. Some struggle academically due to disabilities or learning difficulties. Others have unpropitious families or peers and environments, which make it difficult for them to focus on school. Yet others are just not well served by school: some do well academically, and others do not, but they are bored or irritated at that age by the constraints of school. Other students want to pursue alternative projects (they want to travel or pursue unconventional careers). Undoubtedly, a non-trivial number of students are, by the age of 16, alienated from formal schooling.⁵ They are unlikely to be learning much, if anything, and they are likely to engage in some of the most disruptive behaviour.⁶ Disruptive students are costly to secondary schools. The effects of disruptive student behaviour are not limited to the learning of the disruptive student. For smallscale disruptions (e.g. a student arriving late or refusing to be quiet when the teacher asks the class to concentrate on what another student has been saying), the loss of attention from the teacher and the loss of instructional time may exhaust the learning-relevant aspects of the event. But other disruptions (e.g. issuing a threat to, or swearing at, a teacher or another student, banging on a desk, or loud whispering) can be distracting for other students beyond the time that the teacher is engaged in stopping the behaviour. In a classroom with persistent disruptions, students can be distracted simply by anticipation of the next disruptive behaviour. And not all disruptions occur in the classroom—audible occurrences outside the classroom during instructional time can impinge on learning, as can experiences, or merely reasonable fear, of being unsafe during recess.⁷ Schools can, and do, reduce these costs to learning, and this is costly: maintaining school discipline makes demands on the time of deans of students, assistant principals, security staff, and school psychologists. Schools in the United States commonly contract with Police Departments for dedicated police personnel. Fully evaluating the cost of behaviour management would take into account not only often large fractions of the salaries of the inhabitants of aforementioned roles, the instructional time lost to discipline, and learning lost to infractions but also the loss of quality of teachers and the cost of replacing them. In a study of teachers leaving New York City schools,

⁵ Of 306 17-year-olds surveyed in Quebec about the reasons why they decided to leave school, most (46.9 per cent) listed school-related reasons: a lack of interest in school, boredom, demotivation, or problems with teachers. The other two main reasons listed were the desire or need to work (28.6 per cent) and personal or familial problems (10.7 per cent). See Saysset et al. (2007). Economists Messacar and Oreopoulos (2012) similarly note that students in the United States ‘often report leaving because they are not motivated or inspired to work hard, or because classes are not interesting enough’ (2012: 8). ⁶ Data from Quebec, for instance, indicates that 15-year-old boys on the path to dropping out engage in more misconducts than other students and their behaviour tends to be more disruptive for the school organization as a whole compared to other students or girls at risk of dropping out. Students who dropped out in general are more than twice more likely than other students to have been expulsed from school or to have visited the direction’s office for disciplinary reasons. See Saysset et al. (2007). ⁷ Violence at school is widespread. One survey on school violence in Canada conducted by Mission Research for the CBC reveals that 4 in 10 boys between the ages of 14 and 21 have been physically assaulted at school, 1 in 5 with a weapon, and 12 per cent have been sexually assaulted. See Common et al (2019); Gorbeil (2019). School crime and violence is a serious problem in the United States as well. For recent statistics, see Wang et al. (2019).

188

An Education Resource Account for Early School Leavers

Boyd et al (2011) found that student discipline issues came second only to lack of administrator support (itself, often discipline-related) among school-based factors explaining why teachers left a school. Another study found that new teachers were 8 per cent more likely to exit the profession if they were threatened by a student and, more generally, that alternative work opportunities and wage differentials have less effect on teacher attrition than work environment (Gilpin 2011). No studies are available concerning those people who are repelled from the profession in the first place because they believe that student discipline will be an unreasonably significant part of their job. Not all disruptive students are alienated from school and some disruptive students would not choose to leave. We conjecture, though, that a substantial proportion of those students whose behaviour disrupts the smooth functioning of school life would decide not to attend school past age 16, which, in turn, would reduce disruption costs overall. One US study found that raising the compulsory age to 18 increased crime and violence rates within public high schools. On average, 11.5 more crimes per school were perpetrated yearly after the law changed (Gilpin and Pennig 2012). This matters a great deal from the point of view of learning and achievement as well as students’ daily lived experience. Crime rates and danger at school profoundly affect educational outcomes for all students.⁸ We could not find data on the effects of raising the school leaving age on smaller-scale disruptions. But if students on the path to dropping out tend to be uninterested in school and in so far as uninterested students tend to engage in disruptions of all kinds, the negative effects of raising the legal drop out age on learning and the quality of the school environment might be more pervasive than we know. We should also consider the effects of lowering the compulsory age on students’ and schools’ behaviour. Knowing that one doesn’t have to be at school itself may change behaviour: a student who is alienated from school and resents having to be there or thrives on the attention of officials who, they know, are required to keep them in school, might behave differently in the absence of compulsion.⁹ And knowing that students aren’t required to attend might lead schools to change their own behaviour. They would continue to have a financial incentive to attract students and lowering the school leaving age would subtly shift power to the potential leavers. ‘What would keep these students in school willingly?’ is a very different question for schools to ask than ‘How do we enforce the requirement that they attend school?’ Additionally, schools that are not required to expend so much management time, energy, and

⁸ See Gilpin and Pennig (2012: 17) for an overview of the evidence in supporting this claim. ⁹ It is also possible that some, or even many, students who are currently disruptive in school would actually stay in school and would be less disruptive if attendance were voluntary. There’s an obvious reason for this: disruptive behaviour is contagious, and some currently disruptive students who are on the road to dropping out would, in the absence of disruptive peer influence, be more disposed, and find it easier, to comply with the standards of the school. For discussions of peer influence in the classroom, see Cooley (2010); Sacerdote (2011); Burke and Sass (2013).

Ageing without Ageism?

189

attention on managing behaviour would be more able to experiment and innovate with curricular and pedagogical strategies to improve educational provision.1⁰ Even if students are all well-willed and well-behaved, some resources must be devoted to maintenance of discipline. Our proposals would not eliminate the need for behaviour management protocols, or the staffing to implement them, or leadership that maintains a culture in which they work well. But the associated costs would be lower if there were fewer and less serious disruptions, and we have provided reasons to believe that reducing the legal drop-out age to 16 is likely to contribute to reducing those costs. Before proceeding, we want to emphasize that this is only one part of our proposal and that we have offered only a partial defence of it. The argument we have proposed so far may not be weighty enough on its own decisively to support lowering the school leaving age to 16. The case becomes stronger when the first argument is combined with a second different kind of argument—which we offer in section 3. Then, in section 4, we argue for the ERA, which would insure the educational interests of the students who drop out and the interests of society in having them educated.

3. An argument from anti-paternalism In this section, we defend a moderately anti-paternalist stance with respect to 16year-olds, suggesting that they might reasonably be granted the choice about what to do.11 Opponents of lowering the compulsory age of schooling to 16 might argue that teenagers are often incompetent decision-makers, that they tend not to know what’s best for them, and/or that, in any case, we should limit their freedom to act as they see fit for their own good. In fact, proponents of raising, or maintaining, the school leaving age to 18 sometimes offer explicitly paternalistic arguments (Messacar and Oreopoulos 2012). Many teenagers though, have quite reasonably concluded, by the time they turn 16, that their particular school is not well suited to them. Whereas they may also have wrongly judged that formal education is not for them, many such students have good reasons to believe that the actual experiences they will have at school in the short term will not be valuable.12 However, independently of whether leaving school at 16 is the best decision teenagers could make for themselves, we think there is a strong presumption in favour of letting them make their own decisions, at least from a legal standpoint. The law should not require them to stay in the formal education system against their will. 1⁰ Some studies of the cost of behaviour management make low estimates because they neglect the labour costs associated with enforcement. See, e.g. Swain-Bradway et al. (2017). See Foster et al. (2007) for a careful explanation of how to measure the true costs of behaviour management systems and why the real labour costs are often neglected. 11 See also Chapter 3 in this volume. 12 It is possible that releasing students from a miserable schooling experience early would reduce the chances of them drawing the false inference that because their school doesn’t work for them, school in general wouldn’t work for them.

190

An Education Resource Account for Early School Leavers

Why let 16-year-olds decide whether or not to stay in school? Consider this. On the one hand, we think that younger children are the proper subjects of paternalistic treatment in education and beyond. On the other hand, we think that most adults in most contexts are unfitting subjects of paternalistic treatment—unlike children, they should be free to make their own decisions (and mistakes) about how to live and what to do, in light of what they think is good for them or valuable. So, the question is: why treat older teenagers more like adults than children? To answer this, we need to say more about what makes paternalism towards adults morally objectionable. To treat someone paternalistically is to limit their autonomy or freedom to act as they see fit or simply intervene in their life based on considerations having to do with their well-being. Most moral and political philosophers think that it is at least presumptively wrong to limit the autonomy of adults in order to promote their own good. That doesn’t mean that adults should have unlimited freedom to do what they please: we can limit freedom based on considerations of justice and other third-party grounds. The idea is simply that it is normally objectionable to make decisions for others, or act against their will, specifically ‘for their own good’.13

3.1 Two types of argument against paternalism Explanations for why paternalism is wrong fall into two broad categories: welfarebased and respect-based justifications. The first main welfare-based justifications, famously proposed by John Stuart Mill, say that paternalism is morally objectionable because people generally are the best judges of what’s good for them. This justification relies on a disputable empirical conjecture. Recent scientific evidence suggests that most of us are relatively poor decision-makers. We have limited attention and self-control and a range of cognitive biases that cause us routinely to make irrational decisions (Thaler and Sunstein 2009; Arieli 2008; Kahneman 2011). This suggests that we aren’t especially good judges of what’s in our best interests and that we may not always be the best judges either.1⁴ A more plausible welfare-based justification against paternalism is that autonomous decision-making is itself a component of well-being.1⁵ Even if people aren’t excellent at judging what’s in their best interests or making prudentially great decisions, they may benefit from the choices they make being their own. 13 There are a few competing accounts of exactly how paternalism should be understood (for instance, philosophers debate whether paternalism necessarily involves restricting freedom or whether it necessarily involves being motivated by a negative judgement about the other person’s ability to decide for themselves). See Quong (2011: Chapter 3) for a helpful overview of alternative accounts of paternalism. We don’t need to take a side in these debates here. The arguments we offer in this section are independent of the specific account of paternalism one adopts. 1⁴ In fact, the recent empirical research in social psychology and behavioural economics has led some philosophers to argue that paternalism, including coercive paternalism, is indeed often justified. See Conly (2013). 1⁵ Note, however, that Riley (2018) offers a more nuanced interpretation of Mill’s ‘best judges’ argument according to which ‘a competent individual, though not always prudent, is the best judge of her own good as she conceives it’. See Riley (2018: 162).

Ageing without Ageism?

191

In contemporary moral and political philosophy, respect-based opposition to paternalism is more common than welfare-based opposition. The idea is that people should decide what is good for them and how they want to live because it is their life. We each have authority over our own life because we are equals. The view here is that respecting someone’s moral and political equality requires recognizing that they have ultimate authority to judge and pursue their own ends. If so, paternalism (towards adults) is always presumptively wrong since it involves intervening in the life of another person based on our own idea of what is good for them. This involves treating them as if they were not our equals and, as such, disrespecting their moral status. Both these accounts of paternalism’s objectionability support our claim that we should generally respect the choices of 16-year-olds about their schooling. Notice that, in all cases, what makes someone an unfitting subject of paternalistic treatment is possession of the relevant capacities. People count as equals and as having authority over their own lives only if they possess the capacities that allow them to make their lives their own in a meaningful sense; and they only benefit all-things-considered from choice-making if they are sufficiently good at choosing. The relevant capacities typically include practical and moral reasoning and the capacity for having a minimally robust sense of identity. If the autonomy threshold for paternalism is set too high, then paternalism will be justified overly broadly because adults are not as competent at decision-making as we like to think. No matter where the threshold is set, there will be individual differences between and within age groups, and at least some 16-year-olds won’t meet the threshold. Yet, if the threshold is set low enough to avoid justifying paternalism toward adults, most 16-year-olds will meet it and many by a greater margin than many adults. In fact, not only is it the case that some 16-year-olds are more competent decision-makers than some adults, but also there is evidence suggesting that the differences between older teenagers and adults in terms of cognitive abilities are insignificant (Steinberg et al. 2009). But even if teenagers are statistically less competent than adults at making rational decisions, a very large number remains competent enough and many more so than adults, who we think should not be paternalized. So, unless we are prepared to accept more paternalism towards adults than most liberalegalitarian philosophers find tolerable, we have reason to accept at least a general presumption in favour of respecting the choices of 16-year-olds.

3.2 Against legal paternalism We will consider three possible objections to this view. First, suppose that most 16-year-olds are competent to make, and would benefit from making, their own decisions, while some are not. It is more plausible to conclude that we should individualize our treatment of teenagers than that all 16-year-olds should be treated paternalistically. Some 16-year-olds should be paternalized, whereas others shouldn’t, and we should be careful to evaluate each case separately.

192

An Education Resource Account for Early School Leavers

Second, one could argue, following Franklin Hall (2013) and Fowler (2014), that even if most 16-year-olds are as competent as many adults who we think shouldn’t be paternalized, it is nonetheless justified to paternalize 16-year-olds but not adults because the stage of life one occupies makes a moral difference. We can both maximize the development of autonomy and treat everybody equally over the course of their life by educationally paternalizing individuals when they are younger (even those who already possess well-developed capacities for autonomy) but refrain from paternalizing them when they are older. This seems to provide a reason against allowing 16-year-olds to make their own choices about staying at school or not. Third, some 16-year-olds have low capacities for autonomy, and even if they are sufficiently autonomous in the morally relevant sense, some will make catastrophic choices for themselves, which risk making them miserable in the short and/or long term. In some cases, the choice to leave school at 16 (e.g. to engage in criminal activity) will be one such decision. So, we should paternalize 16-year-olds in order to avoid that scenario. We have sympathy for all of these objections. But they don’t undermine our proposal that the law should not force 16-year-olds to remain in the formal schooling system against their will for paternalistic reasons. Anti-paternalism, even towards adults, has its limits. When a competent adult contemplates a bad, or indeed catastrophic, decision and can’t be reasoned out of it, their loved ones or others are sometimes justified in paternalizing them, especially if they know them well and care about them. But they have reason to adopt the least coercive approach possible. Forcing someone (not) to do something is (generally) worse than encouraging, incentivizing, or nudging them (not) to do that thing. That said, the fact that paternalism may sometimes be justified in interpersonal relations doesn’t imply that the law should be paternalistic. Not only is the law coercive, but it is also a blunt, impersonal apparatus. The law cannot make fine-grained judgements about the appropriateness of paternalizing particular individuals with respect to particular decisions in specific contexts. The law paints a broad brush of how people should be treated based on more or less arbitrary criteria such as age. So, what we need to evaluate are the costs associated with alternative legal generalizations. The choice is stark. We paternalize all 16- and 17-year-olds by forcing them to stay in school, at the cost of disempowering many competent individuals who have good reason to prefer leaving school in the short term and may benefit from having that option in various other ways. Or we treat them as capable of leading their own lives, at the cost of legally allowing some to make bad choices or choices they aren’t fully competent to make. The second option is superior, especially if we consider that there are other ways to help students than coercing them to stay in school by law. These may include softer and more tailored paternalistic interventions such as giving paternalistic guidance and incentives. Adults such as teachers, parents, family members, and community leaders are better positioned than the law to do that well. That’s simply because knowing the individual helps to figure out how (best) to promote their interests or whether treating them paternalistically is appropriate at all.

Ageing without Ageism?

193

We should also keep in mind that schools, families, and communities have indeed a responsibility to ensure that teenagers who wish to leave school are at least properly informed about the possible consequences of their choice and about their alternatives. Providing at-risk students with relevant information is generally not excessively difficult or costly for teachers or other educators to do. Nor is it paternalistic. On the contrary; it is empowering them to make better choices for themselves. The same is true for providing students with opportunities to deliberate with others about what to do, to seek advice from educators and counsellors they can trust. Schools should do those things and school policies should support schools in so doing. But even when schools fail in this respect, that does not suffice to justify legally paternalizing 16-to17-year-old students by forcing them (including those who are autonomous enough to make their own decisions and those who would benefit from leaving school in the short term) to stay in a system that actually disserves and fails them.

4. Case for establishing an ERA Suppose the school leaving age is reduced. Many of those who, as a result, ceased to attend school would be worse off, especially if they never return to formal education—and most of those are, predictably, among the most disadvantaged students (Wood et al. 2017). Members of racialized groups, linguistics minorities, and students from low-income households are overrepresented in the population of students who drop out before the end of high school (Saysset et al. 2007; Messacar et al. 2012). Scholars in favour of requiring schooling until age 18 point to sobering statistics about the life prospects of individuals who drop out of school before 18. As Messacar and Oreopoulos report: Among recent dropouts in the United States, 16 percent are unemployed and 32 percent live below the poverty line; those with jobs earn an average of only $12.75 per hour [. . .] Labor-market outcomes remain bleak throughout life [. . .] [S]ocial outcomes are worse for dropouts compared to any other education attainment group. Thirty-three percent of recent female dropouts have given birth while a teenager, 13 percent of male and female dropouts are separated or divorced, 32 percent report being unhealthy, and 22 percent report being unhappy. (Messacar and Oreopoulos 2012)

Any plausible theory of educational justice would support prioritizing benefit to the disadvantaged students (Schouten 2012). Does this count against our proposal to reduce the school leaving age to 16? The second part of our proposal, the establishment of an ERA, helps to address this objection. As things stand, if a student drops out of school before 18, they are forgoing their one chance for a guaranteed free education. The case for the ERA is simply this: it would be unjust, and socially suboptimal, to force the students who leave school to forfeit the resources that would otherwise be spent on their schooling. The ERA secures educational resources for students who decide to leave school early

194

An Education Resource Account for Early School Leavers

and guarantees an opportunity for them to return to education, deliberately, when they are ready. The precise administrative details of the ERA would depend on context. For example, in England, where education funding is a matter for central government, the funds would probably be held and administered by central government, although the amount of funding would be sensitive both to the spending in the locale where the student is at school and to their specific circumstances. In the United States, where states authorize school funding (much of which is raised locally), the funds would probably be held and administered by states or even by school districts. Again, though, the amounts should be sensitive to local expenditure and individual circumstances. Ideally, the fund would be administered at the least local level of government feasible (to enable mobility) but be as sensitive as possible to the needs of the individual student (to enable generosity to students who are the least well served by the system). Accredited institutions should not be restricted to traditional academic institutes of further education but would include trade schools and institutions with apprenticeship programmes. Why should the ERA be an educational resource account? Wouldn’t it be more consistent with the anti-paternalism we have defended in section 3 just to give students a cash grant instead? The answer has two parts. First, our anti-paternalism is moderate. We really do think that education is valuable enough for the students that it is worth not creating a short-term incentive for them to quit, which a cash payment might do. Second, though, the case for government spending on education is not just that education is good for the individual person who gets educated. It also hinges on the fact that their education enables them better to contribute to society, something that others have a legitimate interest in them doing. Messacar and Oreopoulos observe: Several studies also link a region’s proportion of dropouts to its overall prosperity. Individuals earn higher wages if they work in regions with fewer dropouts, irrespective of their own level of educational attainment (e.g., Ciccone and Peri, 2006; Moretti, 2004; Rauch, 1993). Crime rates are lower (Lochner and Moretti, 2004), and civic participation is higher (Dee, 2004; Milligan, Moretti, and Oreopoulos, 2004). For these reasons, the high school dropout rate is sometimes used as a quality measure of schools and an appraisal of the skill level of the future national workforce (Heckman and LaFontaine, 2010). (Messacar and Oreopoulos 2012: 5–6)

There are also documented broader consequences of compulsory education that make democracies more effective by increasing political interest and involvement (Milligan et al., 2004). (Messacar and Oreopoulos 2012: 10)

Messacar and Oreopoulos take this as evidence in favour of schooling being compulsory to age 18. But it is, in fact, just evidence that people are better off when their

Ageing without Ageism?

195

neighbours are better educated. Using the ERA instead of a cash pay-out reflects this fact. Early school leavers don’t, as individuals, have a claim on the money that would have been spent on them: they have a claim on the education system being designed in a way that educates them well over the long term without unduly paternalizing them in the short term. Establishment of the ERA improves the design of the system. In our description of the ERA, we said that the funds should be available throughout the life of the student. Given the third-party interest in having a well-educated population, and given the strong evidence that earlier educational interventions are more efficient at producing human capital, would it be better to place a term limit on the funds so they expire, or are gradually reduced at, say, age 30, or 35? Experimental social science might find which policy is more efficient, but, unless the effects are large, a lifetime entitlement seems more humane and signals the value of lifelong learning.

5. Conclusion We defended two policy proposals in this chapter, which, we argued, should be adopted together. First, schooling should not be mandatory past the age of 16 for both efficiency and anti-paternalistic reasons. Second, an ERA should be established for school leavers for basic reasons of justice and because it benefits the community as a whole. We anticipate some readers wondering why we don’t just advocate improving schools to enable them to accommodate the full range of students: why not just make schools more engaging and interesting? We think schools should be better than they are, for everyone! But, as a general matter, improving schools is very difficult indeed, even on a piecemeal basis.1⁶ The technologies for widespread improvement through government action are not well understood—if they were, schools would be much better than they are. As it is, government initiatives for school improvement have, at best, incremental benefits and take time to succeed. In the meantime, we would rather free over-16s from school, while providing them with the lifeline of a guarantee of educational investment later in their lives. And, of course, we conjecture that our proposal would have an improving effect by freeing up the resources currently consumed by students for whom school is not a good option. Readers may also wonder what our proposal implies about the use of age limits in law and public policy in general. There are many other activities which arguably should be regulated with age limits, such as paid work, sexual activity, drinking, marriage, military service, and voting. We don’t take ourselves to have offered a strategy about how to handle age limits in all cases. We think that schooling is a distinct domain of state intervention. In the case of the other activities mentioned, the question is what minimum age is appropriate to acquire a permission to do something. With the exception of voting, the most important moral consideration in these 1⁶ For excellent explanations of why school improvement is so difficult see Elmore (2004); City (2008); Payne (2008).

196

An Education Resource Account for Early School Leavers

cases is probably that of protecting the vulnerable more than benefiting the parties concerned. By contrast, schooling is a huge imposition, which dramatically structures the entire lives of students and which must be justified, at least in part, by its value for the person receiving it. The less confident we are that the recipient and others will actually benefit, the less confident we should be that we are justified in imposing it.

Acknowledgements The authors are grateful to the editors for extensive and valuable feedback and to Madelyn Aaronson and Arjan Heir for editorial assistance.

References Ackerman, Bruce, and Anne Alstott. (1999). The Stakeholder Society. New Haven, CT: Yale University Press. Ackerman, Bruce, Anne Alstott, and Philippe Van Parijs. (2006). Redesigning Distribution. London: Verso. Alstott, Anne. (2005). No Exit: What Parents Owe Their Children and What Society Owes Parents. Oxford: Oxford University Press. Arieli, Dan. (2008). Predictably Irrational: The Hidden Forces That Shape Our Decisions (rev. and exp. edn). New York: HarperCollins E-books. Boyd, Donald, Pam Grossman, Marsha Ing, et al. (2011). ‘The Influence of School Administrators on Teacher Retention Decisions’. American Educational Research Journal 48 (2): 303–333. Burke, Mary A., and Tim R. Sass. (2013). ‘Classroom Peer Effects and Student Achievement’. Journal of Labor Economics 31 (1): 51–82. City, Elizabeth. (2008). Resourceful Leadership: Tradeoffs and Tough Decisions on the Road to School Improvement. Cambridge, MA: Harvard Education Press. Common, David, Anu Singh, and C aitlin Taylor. (2019). ‘“I Thought He Was Dead”: CBC Survey Reveals 4 in 10 Boys Are Physically Assaulted at School’. CBC News, 24 October. https://www.cbc.ca/news/canada/school-violence-marketplace-1.5224865. Conly, Sarah. (2013). Against Autonomy: Justifying Coercive Paternalism. New York: Cambridge University Press. Cooley, Jane. (2010). ‘Classroom Peer Effects’. In The New Palgrave Dictionary of Economics, ed. Steven N. Durlauf and Lawrence E. Blume. London: Palgrave Macmillan, pp. 1–7. Elmore, Richard. (2004). School Reform from the Inside Out: Policy, Practice, and Performance. Cambridge. MA: Harvard Education Press. Foster, E. Michael, Deborah Johnson-Shelton and Ted K. Taylor. (2007). ‘Measuring Time Costs in Interventions Designed to Reduce Behavior Problems among Children and Youth’. American Journal of Community Psychology 40(1–2): 64–81.

Ageing without Ageism?

197

Fowler, Timothy. (2014). ‘The Status of Child Citizens’. Politics, Philosophy & Economics 13 (1): 93–113. Franklin-Hall, Andrew. (2013). ‘On Becoming Adult: Autonomy and the Moral Relevance of Life’s Stages’. Philosophical Quarterly 63 (251): 223–247. Gilpin, Gregory A. (2011). ‘Reevaluating the Effect of Non-Teaching Wages on Teacher Attrition’. Economics of Education Review 30 (4): 598–616. Gilpin, Gregory A., and Luke A. Pennig. (2012). ‘Compulsory Schooling Laws and InSchool Crime: Are Delinquents Incapacitated?’ Center for Applied Economics and Policy Research, CAEPR Working Papers, 2012-005. Gobeil, Mathieu (2019). ‘Plus du tiers des jeunes Canadiens victimes de violence à l’école’. Radio-Canada. https://ici.radio-canada.ca/nouvelle/1357099/violence-ecoleprimaire-secondaire-canada-physique-sexuelle-sondage-etude Kahneman, Daniel. (2011). Thinking, Fast and Slow. New York: Farrar, Straus and Giroux. Messacar, Derek and Philip Oreopoulos. (2012). ‘Staying in School: A Proposal to Raise High School Graduation Rates’. The Hamilton Project. Miliband, David. (1990). Learning by Right: An Entitlement to Paid Education and Training. London: Institute for Public Policy Research. Payne, Charles. (2008). So Much Reform, So Little Change: The Persistence of Failure in Urban Schools. Cambridge, MA: Harvard Education Press. Piketty, Thomas. (2020). Capital and Ideology. Cambridge, MA: Harvard University Press. Quong, Jonathan. (2011). Liberalism without Perfection. New York: Oxford University Press. Riley, Jonathan. (2018). ‘Mill’s Absolute Ban on Paternalism’. In The Routledge Handbook of the Philosophy of Paternalism (1st edn), ed. Kalle Grill and Jason Hanna. New York: Routledge, pp. 153–169. Sacerdote, Bruce. (2011). ‘Peer Effects in Education: How Might They Work, How Big Are They and How Much Do We Know Thus Far?’ In Handbook of the Economics of Education, Vol. 3, ed. Eric A. Hanushek, Stephen Machin, and Ludger Woessmann. Amsterdam: Elsevier, pp. 249–277. Saysset, Valérie, Lise Giroux, and Ève-Marie Castonguay. (2007). ‘Décrochage et retard scolaires: Caractéristiques des élèves à l’ˆage de 15 ans’. Gouvernement du Québec Ministère de l’Éducation, du Loisir et du Sport. Schouten, Gina. (2012). ‘Fair Educational Opportunity and the Distribution of Natural Ability: Toward a Prioritarian Principle of Educational Justice’. Journal of Philosophy of Education 46 (3): 472–491. Steinberg, Laurence, Elizabeth Cauffman, Jennifer Woolard et al. (2009). ‘Are Adolescents Less Mature than Adults?: Minors’ Access to Abortion, the Juvenile Death Penalty, and the Alleged APA “Flip-Flop”’. American Psychologist 64 (7): 583–594. Stemplowska, Zofia, and Adam Swift. (2012). ‘Ideal and Nonideal Theory’. In The Oxford Handbook of Political Philosophy, ed. David Estlund. New York: Oxford University Press, pp. 373–392.

198

An Education Resource Account for Early School Leavers

Swain-Bradway, Jessica, Sarah Lindstrom, Catherine Bradshaw et al. (2017). ‘What Are the Economic Costs of Implementing SWPBIS in Comparison to the Benefits from Reducing Suspensions?’ Positive Behavioral Interventions & Supports: 1–7. Wang, Ke, Yongqiu Chen, Jizhi Zhang et al. (2020). ‘Indicators of School Crime and Safety: 2019’. National Center for Education Statistics, U.S. Department of Education and Bureau of Justice Statistics. https://nces.ed.gov/pubs2020/2020063.pdf. Wood, Laura, Sarah Kiperman, Rachel C Esch et al. (2017). ‘Predicting Dropout Using Student- and School-Level Factors: An Ecological Perspective’. School Psychology Quarterly: The Official Journal of the Division of School Psychology, American Psychological Association 32 (1): 35–49. World Population Review. (2021). ‘High School Graduation Rates by State 2021’. World Population Review. http://worldpopulationreview.com/states/high-schoolgraduation-rates-by-state.

14 Differentiating Retirement Age to Compensate for Health and Longevity Inequality? Vincent Vandenberghe

The rise of average healthy life expectancy calls for an overall increase in retirement age.1 However, many observers suggest that the latter should be differentiated across socio-demographic categories to achieve equality. The request stems from the observation that same-age older individuals differ a lot in terms of health and, what is more, that the differences systematically correlate with their socio-demographic background and also remaining longevity. Historically, in most retirement systems, a uniform age has been used to proxy for poor health and the subsequent loss of work capacity and the need for replacement earnings. But now comes this proposition to adopt a slightly more refined proxying strategy: using several retirement ages to better match the distribution of health status and work capacity across socio-demographic groups. Many observers and stakeholders call for the abolition of the uniform retirement age and its replacement with a differentiated retirement age policy. A first sight, the proposal makes perfect sense. But we will show in this chapter that, in a world of systematic retirement age differentiation by socio-demographic group, there would still be a lot of what health economists call type-F errors (failure of treatment, i.e. no retirement for people in poor health) and type-E errors (excessive treatment, i.e. people in good health going for retirement).

1. Introduction The increase in life expectancy is arguably the most remarkable by-product of medical progress and economic growth. Since the end of the nineteenth century, advanced economies have been gaining roughly 2.4 years of longevity every decade (Oeppen et al. 2002). But this trend—in combination with lower fertility—translates into population ageing. And this has far-reaching economic and socio-political consequences. Different things could adjust to counterbalance the contraction of the working-age 1 Pension economists regularly recommend indexing the retirement age on the population’s life expectancy. Denmark and the Netherlands have enacted reforms that make these indexations automatic. Vincent Vandenberghe, Differentiating Retirement Age to Compensate for Health and Longevity Inequality?. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Vincent Vandenberghe (2023). DOI: 10.1093/oso/9780192894090.003.0015

200

Differentiating Retirement Age to Compensate for Health

population and the rise of old-age dependency. They comprise higher female participation in the labour force (at least in the countries where it remains low), slightly longer hours of work, less unemployment, more flexible work arrangements,2 or even shorter initial education (Gosseries and Vandenberghe 2016). But so far, the most common form of adjustment consists in raising the effective3 age of retirement. Researchers at the Organisation for Economic Co-operation and Development (OECD) have shown that indexing retirement age on (rising) life expectancy could stabilize old-age dependency ratios, preventing dramatic tax increases to finance payas-you-go pensions, or a general reduction of the level of pensions (Oliveira Martins et al. 2005). And stricter retirement policies implemented since the mid-1990s have proved effective at increasing employment rates (Atalay and Barrett 2015), although from a historically low level (Costa1998). However, one concern often raised is whether such policies are fair, as older workers may differ significantly in their health status, residual work capacity, and remaining life expectancy.⁴ What is more (and this, in a sense, is good news for the advocates of retirement age differentiation), these differences are not distributed randomly but are systematically correlated with socio-demographic traits. And these are easy to observe (education, gender, income percentile)⁵ and could thus be used as a basis for differentiation. The question we ask more specifically in this chapter is: what would it take in terms of lowering (or raising) the retirement age to ensure that all socio-demographic groups can expect to retire in similar health⁶ and, by extension, can expect to live the same number of years in retirement? The overall normative perspective underpinning this exercise is egalitarian.⁷ But note that when stressing the equality of expected outcomes, we hint at ex ante egalitarianism (Diamond 1967), while some of our results about people’s health at the moment of their retirement signal the likely importance of ex post egalitarianism (Fleurbaey et al. 2016; Ponthiere 2020).⁸ The ex ante perspective considers the average outcome (here, health and, by extension, longevity) that someone can expect to attain at the moment they retire (given the

2 To attract into paid employment people who otherwise stay completely out of the labour force. 3 Which can be significantly lower than the legal age of retirement. ⁴ There is strong evidence that ill health at 50 is correlated with a shorter life span/early death. De Nardi et al. (2016) show that lifespan is 3.3 years shorter for those with bad health than for those with good health, while Pijoan-Mas et al. (2014) show that the equivalent numbers are 5.6 for men and 4.7 for women at age 50. ⁵ Economists would also argue that many of these are not that easy to ‘manipulate’, minimizing the risk strategic behaviour to be allowed to retire earlier. ⁶ Or, equivalently, can expect to work only as long as their health-driven work capacity stays above a certain threshold, which is the same for every group (Vandenberghe, 2021b). ⁷ It can also be connected to the objective of ‘actual fairness’ commonly used by pension specialists. Strictly speaking, actuarial fairness means that an individual’s lifetime pension contributions and benefits should equate. But, in more utilitarian terms, some euros of accrued pension benefits have caused more disutility (due to more arduous jobs, for instance). And poor health at an older age means that people’s ability to enjoy pension benefits may be significantly reduced. What is more, in its strict financial sense, there is no actual fairness if remaining life is considerably shorter for some individuals. ⁸ Ex ante vs ex post should be understood here primarily in their temporal sense. The ex ante term points at expectations or predictions before people retire, while the ex post term logically refers to the realization of health at that age.

Ageing without Ageism?

201

socio-demographic group they belong to), while the ex post standpoint is that of their actual and realized health on the day they retire.⁹ And there can be a significant discrepancy between the two. Statisticians invariably adopt an ex ante perspective. For instance, they may have computed that you and your colleague should still be in perfect health at the age of 67, given the socio-economic category you belong to, but ex post, when both of you reach that age, it turns out your respective health statuses diverge significantly. The chapter tries to inform the policy debate. We compute realistic estimates of the degree of retirement age differentiation needed to compensate1⁰ for health differences at an older age. To that end, we make extensive use of European Survey of Health, Ageing and Retirement in Europe (SHARE) panel data documenting the health status of large numbers of individuals aged 50+ in more than 20 countries. We first quantify the health gradient ∆H across European countries and within each of them across socio-demographic groups (i.e. gender, education) at typical retirement age. We then estimate the degree of retirement age differentiation ∆a that would be needed to equalize expected health at the moment of retirement. We will expose in detail here below the way this is done, but here is the key idea. Consider that, at a given uniform retirement age a, individuals belonging to two different sociodemographic groups display a sizeable health gradient ∆H. Consider, at the same time, that biological ageing invariably causes a decline of health at a rate say β. This almost trivial assumption has key implications in the context of this discussion. It means that advancing (or postponing) the age of retirement could be a way to compensate for same-age health differences across countries or socio-demographic groups, and thus equalize (expected) health and possibly longevity at the moment of retirement. We will see hereafter that one can relatively easily compute the retirement age differentiation as ∆a = ∆H/β. Note that we discuss the relevance of an ‘automatically’ and group-based differentiated retirement age policy that would aim at equalizing expected health at the moment of retirement. By ‘automatically’, we mean that the right to retire at a certain age would be granted just on the condition of belonging to a particular category or socio-demographic group. There would be no need to undergo screening and be subjected to individualized checks, as is the case to get disability benefits. Another critical point is that the retirement age differentiation exercise presented here assumes that work and retirement’s impacts on health are limited. Lowering or raising the retirement age (and thus varying the duration of careers) is supposed to have no significant effect on health (or longevity). We posit that the latter is primarily driven by age (our variable of differentiation and equalization) and other factors (genetic background, childhood health, and work circumstances correlated to age and education, etc.).11 This may seem to be a strong assumption, but a relatively ⁹ Fleurbaey et al. (2016) state, ‘At the end of the day, what matters is what people achieved, not what they expected to achieve.’ 1⁰ Positively or negatively. 11 And also, when considering differences across countries, macroeconomic variables such as gross domestic product (GDP)/head or capital intensity.

202

Differentiating Retirement Age to Compensate for Health

abundant economic literature backs it. Bassanini and Caroli (2015) review the many papers that have studied the impact of work on health and find very mixed evidence. As to the relationship between work/retirement and longevity, in their recent paper, Bozio et al. (2021) state that ‘if an impact of later retirement on mortality would be detectable, it would remain very small in magnitude’. Our key results point to the need for a very high degree of differentiation to equalize expected health across countries, with a retirement age ranging from 52 in Poland (POL) to 79 in Switzerland (CHE). The degree of retirement differentiation is also significant across socio-economic groups within countries. For instance, tertiaryeducated individuals should retire more than 10.6 years later than those with less than upper-secondary education attainment. But we also show that systematic retirement age differentiation would fail to match a significant portion of the ex post distribution of health status; that is, in a world of systematic retirement age differentiation by socio-demographic group, there would still be a lot of what health economists call type-F errors (failure of treatment, i.e. no retirement for people in poor health) and type-E errors (excessive treatment, i.e. people in good health going for retirement). This raises the question of which policy is better. Should policymakers go for ‘more elaborate’ but still proxy-based retirement policies consisting of socio-demographicgroup-based differentiated retirement age? Or should they stick to the historical norm of uniform retirement age, supplemented by disability benefits conditional on individual screening of health status? In what follows, we present the SHARE data on health and explain how we can compute health-equalizing differentiated retirement ages. We wrap up the chapter with a discussion of the possible policy implications of our results.

2. Data and analytical framework We use waves 1–2 and 4–7 (2004–2017) of the SHARE survey, a total of 238,363 observations (i.e. the number of individuals times the number of waves they participated to).12 All individuals in SHARE are 50 or older when interviewed for the first time. Data limitations of different sorts (missing values or variables, absence of repeated observations as the country participated only in one wave) explain that we retain only 20 out of the 29 participating countries in the analysis.13 SHARE contains a rich set of items about people’s physical health status and their mental and cognitive health. Here, we will rely only on the former. In Vandenberghe (2021b), we also consider mental health and cognition, but the results are qualitatively very similar to those reported here. Most health items are selfreported, and many are subjective because they correspond to how people perceive and self-assess their overall health status (Table 14.1). But SHARE questionnaires also 12 In SHARE, most respondents are observed consecutively in several survey waves. 13 Austria (AUT), Belgium (BEL), Croatia (HRV), Czech Republic (CZE), Denmark (DNK), Estonia (EST), France (FRA), Germany (DEU), Greece (GRC), Hungary (HUN), Israel (ISR), Italy (ITA), Luxembourg (LUX), the Netherlands (NLD), Poland (POL), Portugal (PRT), Slovenia (SVN), Spain (ESP), Sweden (SWE), and Switzerland (CHE).

Ageing without Ageism?

203

Table 14.1 Health items: Subjective health

AUT BEL CHE CZE DEU DNK ESP EST FRA GRC HRV HUN ISR ITA LUX NLD POL PRT SVN SWE

Poor general healtha

Long-term illnessb

Limitsc

Limitsd

Limitse

2.90 2.92 2.62 3.26 3.17 2.47 3.18 3.75 3.08 2.83 3.24 3.59 2.95 3.11 2.97 2.85 3.59 3.66 3.22 2.65

0.46 0.45 0.34 0.52 0.59 0.49 0.45 0.70 0.43 0.31 0.58 0.66 0.49 0.36 0.46 0.46 0.64 0.52 0.47 0.52

2.44 2.44 2.64 2.35 2.37 2.58 2.66 2.23 2.51 2.75 2.37 2.32 2.57 2.58 2.45 2.35 2.28 2.34 2.42 2.50

0.11 0.17 0.06 0.14 0.13 0.10 0.12 0.22 0.12 0.06 0.15 0.17 0.17 0.10 0.11 0.07 0.23 0.27 0.15 0.10

0.22 0.25 0.10 0.24 0.18 0.17 0.23 0.33 0.17 0.18 0.24 0.40 0.40 0.17 0.18 0.16 0.31 0.33 0.21 0.15

Note: a: average of 1(Excellent)-2(Very Good), 3(Good), 4 (Fair) and 5(Bad), European scale; b:Yes (1) No (0); c: limited in activities because of health (1. Severely limited, 2. Limited, but not severely, 3. Not limited); d: number of limitations with activities of daily living (1–6); e: limitations with instrumental activities of daily living (1–9). Source: SHARE 2004–2017.

explicitly refer to specific health conditions diagnosed by health professionals (heart attack, hypertension, cholesterol, stroke, diabetes, lung disease, cancer, and so on). SHARE interviewers also collect measurements like the maximum grip strength of respondents (Table 14.2). In what follows, we will make extensive use of physical ill-health indices. These are computed as the first principal component1⁴ of available health items (Table 14.1). Not surprisingly, an examination of the indices shows that the incidence of ill health goes up with age (Vandenberghe 2021a). However, there are substantial differences across countries. For instance, at the age of 67, the ill-health index in Switzerland (CHE), at −0.476, is much lower than in Estonia (EST), where it reaches 0.36.1⁵ There 1⁴ Principal component analysis is a technique for reducing data dimensionality, increasing interpretability while minimizing information loss. The values for the principal component reported later have been standardized internationally. This means that a one-unit change of the index corresponds to one standard deviation of the international distribution of the health index. 1⁵ These numbers represent fractions of the international standard deviation of ill health scores and can be considered sizeable health differences.

Table 14.2 Health items: Objective conditions

AUT BEL CHE CZE DEU DNK ESP EST FRA GRC HRV HUN ISR ITA LUX NLD POL PRT SVN SWE

Heart attacka

Hyper- Cholestensiona terola

Strokea

Diabetesa

Lung Cancera a disease

Ulcera

Parkinsonsa,b

Cataracta

Hip Mobilityc a fracture

Max. of grip strength measured

0.10 0.09 0.06 0.12 0.10 0.08 0.08 0.16 0.10 0.09 0.12 0.18 0.13 0.08 0.09 0.09 0.16 0.10 0.10 0.10

0.39 0.33 0.28 0.49 0.41 0.32 0.36 0.47 0.31 0.38 0.46 0.55 0.40 0.39 0.33 0.27 0.44 0.46 0.44 0.36

0.04 0.03 0.02 0.04 0.03 0.03 0.02 0.04 0.02 0.02 0.04 0.06 0.03 0.02 0.02 0.03 0.04 0.04 0.03 0.03

0.12 0.10 0.07 0.17 0.13 0.07 0.15 0.12 0.11 0.11 0.13 0.17 0.22 0.11 0.11 0.09 0.14 0.19 0.13 0.10

0.06 0.06 0.04 0.07 0.07 0.07 0.05 0.06 0.05 0.04 0.04 0.06 0.04 0.05 0.07 0.07 0.05 0.06 0.05 0.04

0.04 0.06 0.02 0.05 0.03 0.03 0.03 0.08 0.03 0.07 0.05 0.08 0.05 0.03 0.06 0.02 0.06 0.08 0.06 0.02

0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.01

0.07 0.05 0.06 0.08 0.07 0.06 0.07 0.08 0.05 0.05 0.05 0.04 0.09 0.05 0.07 0.05 0.06 0.08 0.06 0.08

0.01 0.02 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.03 0.01 0.01 0.02 0.01 0.01 0.02 0.02 0.02

35.15 35.88 35.72 34.91 36.78 37.78 30.51 34.42 34.24 33.32 35.19 32.83 30.26 33.11 35.26 36.68 34.24 30.34 35.36 36.25

0.22 0.31 0.15 0.25 0.19 0.23 0.28 0.20 0.24 0.28 0.21 0.21 0.36 0.23 0.33 0.18 0.23 0.41 0.26 0.17

0.04 0.05 0.04 0.05 0.07 0.05 0.04 0.05 0.05 0.02 0.06 0.05 0.05 0.03 0.07 0.05 0.04 0.06 0.05 0.06

1.24 1.28 0.64 1.38 1.19 0.82 1.30 1.73 1.15 1.38 1.81 2.00 1.29 1.22 1.21 0.89 1.90 1.94 1.53 0.88

Note: a: average response : has (1) has not (0) aforementioned condition; b: or dementia, senility; c: Mobility, arm function and fine motor limitations (0–10 sale). This variable is based on items: limitation with 1. walking 100 m, 2. sitting two hours 3. getting up from chair 4. climbing several flights of stairs 5. climbing one flight of stairs 6. stooping, kneeling, crouching 7 reaching or extending arms above shoulder 8. pulling or pushing large objects 9. lifting or carrying weights over 5 kilos 10. picking up a small coin from a table; d: 0–100 (measured by interviewer). Source: SHARE 2004–2017.

Ageing without Ageism?

205

are also differences in the intensity of the ill-health/age gradient. In other words, both the level of ill health and the ill-health/age relationship vary internationally. We deploy a two-stage estimation using the SHARE data. Stage one aims to identify, for each country j, the degree of retirement age differentiation around the age of 671⁶ that would ensure people retire with a level of (expected) ill health equal to the international average.1⁷ Suppose Hj67 represents the average ill-health index of respondents aged 67 in country j and H67 the average international health. In that case, there is an ill-health index gap in that country equal to the difference between these two terms. If βj67 represents the marginal effect of a year of age on the illhealth index,1⁸ then one can estimate the age of retirement, ensuring equalization of expected ill health as: aj = 67 – (Hj67 – H 67 )/βj67 .

14.1

Stage two proceeds along the same lines as stage one but with the aim of identifying the health-equalizing differentiated retirement age inside each country j for each socio-demographic group k. That intra-national differentiation occurs not around the age of 67, as in the formula above, but around the retirement age aj applicable to the country’s average citizen and computed at stage one. Essential with such a setting are estimates of (i) the ill-health index gaps across countries or socio-demographic groups within each country and (ii) the βs (i.e. the impact of biological ageing on health). As to the latter’s estimation, we resort to the panel dimension of SHARE data (SHARE consists of up to six waves, measuring individuals’ ill health every two or three years). In other words, the calculated βs only reflect the within-respondent deterioration of health over time. This eliminates many of the biases that may contaminate estimates based on cross-sectional data.

3. Results 3.1 Health-equalizing differentiated retirement ages Key results appear in Figure 14.1 and Table 14.3. They display the rather significant degree of retirement age differentiation that would be required to equalize expected 1⁶ Internationally, 67 is gradually becoming the new reference (OECD 2019). This said, we lower that age to 61, 63, and 65 when doing robustness analysis in Vandenberghe (2021a), and results remain qualitatively very similar to those reported here. 1⁷ From a normative standpoint, equalizing expected health (and indirectly longevity) across countries has limited intuitive appeal compared to a similar equalization across groups inside countries. Classical and contemporary political philosophers have focused on justice almost exclusively within countries (or similar politically integrated entities). Discussing the difference between global and more traditional theories of justice is beyond the scope of this chapter. We quantify what it takes to equalize health internationally and separately intra-nationally (see stage 2 below). The reader who thinks that equalization only makes sense intra-nationally should only consider stage 2 results, that is, our estimates of what it takes to equalize expected health across gender and education categories. 1⁸ Note the presence of subscript j, indicating that the marginal effect of ageing can vary from country to country and the superscript 67, indicating that it is calculated around the age of 67.

206

Differentiating Retirement Age to Compensate for Health

Mean retirement age($)/country

Male

Female

53.9 / HUN 52.4 / POL 55.5 / EST 65.1 / ISR 66.5 / LUX 63.4 / HRV 62.5 / CZE 64.6 / SVN 58.6 / PRT 65.7 / DEU 68.9 / AUT 67.5 / BEL 65.8 / ESP 68.6 / FRA 70.2 / GRC 74.9 / DNK 68.9 / ITA 72.8 / NLD 74.6 / SWE 79.4 / CHE 53.9 / HUN 55.5 / EST 52.4 / POL 62.5 / CZE 65.1 / ISR 58.6 / PRT 66.5 / LUX 63.4 / HRV 65.7 / DEU 64.6 / SVN 65.8 / ESP 67.5 / BEL 68.9 / ITA 68.6 / FRA 68.9 / AUT 70.2 / GRC 72.8 / NLD 74.6 / SWE 74.9 / DNK 79.4 / CHE

40

67 y. 45

50

55

60 65 Retirement age

1. ISCED3

Figure 14.1 Differentiated retirement ages equalizing (expected) ill health across and within countries Note: ISCED = upper secondary degree (International Standard Classification of Education). $: Stage one estimates of retirement ages equalizing expected ill health (men and women pooled). Source: SHARE 2004–2007.

ill health upon retirement. Focusing on cross-country differences, we see that Poland (POL) is the country where the age of retirement would have to be the lowest at 52.39. By contrast, it would have to be as high as 79.45 in Switzerland (CHE). By construction, these retirement age differences primarily reflect ill-health gaps among older people. And it is pretty interesting to note that the former—and presumably also the latter—essentially parallel GDP per capita differences. This result is confirmed by an analysis of the differentiated retirement ages. The higher GDP per head but also capital intensity,1⁹ the higher the health-equalizing retirement age (Vandenberghe 2021a). Also, within each country, additional differentiation of retirement age would be needed to account for the significant variations in health across socio-demographic groups. For instance, in Poland (POL), the retirement age should range from 43.97 to 60.67. And in Switzerland (CHE), we estimate that it should vary between 73.42 and 82.61. The combination of between- and within-country ill-health differences among elderly individuals results in (ill)health-equalizing retirement ages ranging from 40

1⁹ Capital intensity is the amount of physical capital (machines, computers, roads, or harbours) per worker. Alongside the well-known GDP per head, economists use it as a proxy of the degree of development and sophistication of an economy.

Table 14.3 Differentiated retirement ages equalizing ill health (international reference = 67): Between- and within-country differentiation

Differentiated retirement

Gap to international reference (67)

Differentiated retirement F

Differentiated retirement M

ISCED3

ISCED3

F-M gap

ISCED gap

Retirement Retirement Maximumminimum maximum minimum

AUT BEL CHE CZE DEU DNK ESP EST FRA GRC HRV HUN ISR ITA LUX NLD POL PRT SVN SWE

68.92 67.49 79.45 62.50 65.71 74.87 65.78 55.52 68.57 70.22 63.42 53.93 65.11 68.89 66.46 72.79 52.39 58.62 64.57 74.62

1.92 0.49 12.45 −4.50 −1.29 7.87 −1.22 −11.48 1.57 3.22 −3.58 −13.07 −1.89 1.89 −0.54 5.79 −14.61 −8.38 −2.43 7.62

67.82 65.45 77.91 61.42 64.38 73.67 68.15 52.78 67.70 67.92 63.61 58.64 64.65 68.53 66.05 72.10 50.43 61.45 65.32 74.36

69.50 70.66 80.79 65.36 66.58 75.46 71.76 56.38 70.95 74.56 64.76 57.61 65.09 73.67 68.22 77.82 55.17 62.49 66.77 76.57

64.11 65.03 79.11 60.94 64.00 71.69 66.95 51.62 68.37 70.84 60.12 43.66 53.61 71.79 53.81 71.89 49.70 61.37 61.07 73.58

71.90 73.44 80.66 63.85 65.81 75.46 72.94 57.01 71.08 75.73 67.49 62.37 67.57 75.19 74.71 78.86 55.13 60.96 66.36 75.78

72.50 73.50 82.61 71.30 69.94 79.24 75.40 60.51 73.41 77.10 66.68 66.81 74.08 74.03 76.13 82.72 60.67 65.14 72.87 80.35

64.81 61.38 73.42 51.75 59.65 72.80 61.23 42.02 64.46 65.36 58.94 40.00 53.87 62.08 56.50 67.77 48.77 54.33 60.25 70.29

68.81 66.41 79.92 65.12 63.13 70.54 74.47 51.02 67.08 67.83 67.84 53.69 69.30 70.37 62.08 76.85 43.97 69.18 61.96 76.20

69.83 68.55 80.38 67.39 70.34 77.68 68.75 65.31 71.56 70.56 64.06 82.23 70.78 73.13 79.57 71.67 58.56 60.84 73.76 76.60

−1.69 −5.21 −2.89 −3.94 −2.21 −1.79 −3.61 −3.60 −3.26 −6.64 −1.15 1.03 −0.44 −5.14 −2.17 −5.73 −4.74 −1.04 −1.44 −2.21

−6.70 −7.82 −5.23 −13.00 −8.31 −6.22 −7.98 −16.09 −6.07 −5.73 −5.84 −32.69 −18.69 −6.64 −22.70 −7.37 −10.38 −5.14 −12.65 −6.54

72.50 73.50 82.61 71.30 70.34 79.24 75.40 65.31 73.41 77.10 67.84 82.23 74.08 75.19 79.57 82.72 60.67 69.18 73.76 80.35

International reference

67

Source: SHARE 2004–2017.

64.11 61.38 73.42 51.75 59.65 70.54 61.23 42.02 64.46 65.36 58.94 40.00 53.61 62.08 53.81 67.77 43.97 54.33 60.25 70.29

8.38 12.12 9.19 19.55 10.69 8.70 14.16 23.29 8.96 11.74 8.90 42.23 20.47 13.10 25.76 14.96 16.71 14.85 13.51 10.06

208

Differentiating Retirement Age to Compensate for Health

(Hungary (HUN), low-educated females) to 82.72 (Netherlands (NLD), highly educated males). Finally, we find that, on average across the European Union, women should be allowed to retire 2.9 years earlier than men.2⁰

3.2 Limit to retirement age differentiation based on ex ante equalization In this section, we focus on what happens inside each country, and we examine and discuss the importance of what economists call the variance ‘within’ sociodemographic groups. So far, inside each country, we have essentially looked at the ‘between’ group variance in an attempt to differentiate retirement age. We have shown that ill health varies significantly (in the statistical sense) between groups at any given age beyond 50. And we have used these differences (in combination with group-specific age/ill-health gradients) to compute differentiated retirement ages, ensuring equalization of expected ill health across groups, contributing thus to the realization of ex ante equality across pensioners (Diamond 1967). But this amounts to focusing on the average characterizing different sociodemographic groups, ignoring the potentially huge dispersion ‘within’ each of them in terms of realized health and thus the full magnitude of the ex post distribution of conditions. From an ethical perspective, both ex ante and ex post egalitarianism are attractive (Fleurbeay et al. 2016), and it is beyond the scope of this chapter to rank them. But we would posit that ex post egalitarians should be concerned to observe that systematic retirement age differentiation as modelled above is still synonymous with what Cornia and Stewart (1993) call type-F and type-E errors. The first type, ‘failure of treatment’, corresponds to individuals suffering from ill health but who belong to the socio-economic group that, on average, fares relatively well and gets assigned a high retirement age. The second type of error—‘excessive’ treatment21—is just the symmetric case, that is, individuals whose health is expected to be relatively bad given the socio-economic group they belong to, and thus are allowed to retire early, but who are de facto in good shape. Figure 14.2 illustrates, for some of the countries from our data set, how difficult it is to avoid type-F and type-E errors and thus to achieve ex post equality. Both errors remain very frequent, whatever the country considered. The dotted lines and curves correspond to highly educated women, the solid ones to low-educated women. The vertical lines represent the average value of the ill-health index. And there is no doubt that highly educated females are, on average, in better health than their less-educated 2⁰ The result that women should be allowed to retire earlier based on their health is somewhat surprising as women have a higher life expectancy. But our result accords with a recurrent result of the morbidity/mortality literature (Case and Praxon 2005). Women have worse self-rated health and more hospitalization episodes than men from early adolescence to old age but are less likely to die at each age. 21 That some may probably consider as less problematic than type-F errors. The underlying question— which is beyond the scope of this chapter—is whether more weight should be put on avoiding type-F errors than on type-E errors.

209

Ageing without Ageism? BEL

–2

0 2 Ill-health index F: low education

0.0 0.2 0.4 0.6

Percentage

Percentage 0.0 0.2 0.4 0.6

DEU

4

–2

F: high education

0 2 Ill-health index F: low education

FRA

0 1 Ill-health index

F: low education

2

3

F: high education

0.0 0.2 0.4 0.6

Percentage –1

F: high education

POL

Percentage 0.0 0.2 0.4 0.6 –2

4

–2

–1

0 1 2 Ill-health index

F: low education

3

F: high education

Figure 14.2 Difficulty of differentiating ex post (importance of type-E and type-F errors). The case of low-educated versus highly educated females aged 55–65 in Germany (DEU), France (FRA), Belgium (BEL), and Poland (POL) Source: This chapter uses data from SHARE Waves 1, 2, 4, 5, 6, 7 (DOIs: 10.6103/SHARE.w1.710, 10.6103/SHARE.w2.710, 10.6103/SHARE.w4.710, 10.6103/SHARE.w5.710, 10.6103/SHARE.w6.710, 10.6103/SHARE.w7.711, 10.6103/SHARE.w8cabeta.001), see http:// BorschSupan2013 for methodological details. The SHARE data collection has been funded by the European Commission through FP5 (QLK6-CT-2001-00360); FP6 (SHARE-I3: RII-CT-2006-062193, COMPARE: CIT5-CT-2005-028857, SHARELIFE: CIT4-CT-2006-028812); FP7 (SHARE-PREP: GA No. 211,909, SHARE-LEAP: GA N\degree 227,822, SHARE M4: GA N\degree 261,982, DASISH: GA N\degree 283,646); and Horizon 2020 (SHARE-DEV3: GA N∘ 676,536, SHARE-COHESION: GA No. 870,628, SERIES: GA No. 654,221, SSHOC: GA No. 823,782) and by DG Employment, Social Affairs and Inclusion. Additional funding from the German Ministry of Education and Research, the Max Planck Society for the Advancement of Science, the US National Institute on Aging (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30 AG12815, R21 AG025169, Y1-AG-4553-01, IAG BSR06-11, OGHA 04–064, HHSN271201300071C), and from various national funding sources is gratefully acknowledged (http://www.share-project.org).

peers: their line is systematically more to the left. The curves depict the distribution of the ill-health index around the average. And there, the key message is that the distributions for highly versus low-educated women overlap. This means that there are many highly educated females with a high ill-health index (higher than the average for low-educated females). These would be denied early retirement despite their ill health. Similarly, many low-educated women have a low ill-health index (lower than the average for highly educated women). This hints at the possibility of many loweducated women in relatively good condition who would (illegitimately) be granted the right to retire early due to inaccurate targeting. One way to go beyond the visual evidence on display in Figure 14.2 is to resort to variance decomposition techniques. This amounts to estimating the share of total country-level ill-health variance explained by the socio-demographic categories used

210

Differentiating Retirement Age to Compensate for Health

above (gender, education). Results reported in full in Vandenberghe (2021a) show that such a share is small, often less than 5 per cent and never larger than 9 per cent.

4. Discussion This chapter explored the idea that retirement age should be differentiated across socio-demographic categories to better match older individuals’ heterogeneous health, diverging capacity to stay in paid employment, and differences in terms of remaining longevity. Using European SHARE data on health, we have computed the degree of retirement age differentiation that would be required to equalize health at the moment of retirement. Such a policy would be a way of systematizing earlier suggestions that pension reforms aimed at raising the retirement age should make an exception for workers with demanding occupations since health considerations may make it unreasonable to expect them to work longer. They also echo recent work on the fairness of retirement systems under unequal longevity (Ponthiere 2020). Health at an older age and remaining longevity are, indeed, highly correlated. Our conclusions are as follows. First, European older populations vary significantly in their health around the typical retirement ages. This is true across countries (with tentative evidence that higher GDP per capita translates into better health) and within countries, between sociodemographic groups, with lower-educated elderly individuals being less healthy than their same-age more highly educated peers (something that might also be related to income differences). Second, equalization can be achieved both across countries and within each country but requires extensive retirement age differentiation. To equalize expected health for the different socio-demographic groups forming their populations, most European countries would have to admit more than 10 years of difference between those groups with the worst and best health status. Third, there are limitations to what can be achieved by resorting to a differentiated retirement age policy to achieve ex ante health and longevity equality. SHARE data clearly show that such a policy would still be prone to ex post inequalities due to extensive type-F errors (failure of treatment, i.e. retirement rights not granted to people in poor health) and type-E errors (excessive treatment, i.e. rights granted to people in good health). The importance to be given to these errors, taken together, is a matter of ethical perspective.22 People who only care about ex ante equality would probably consider that not much should be done about the residual ‘withingroup’ differences, thus about these type-F and type-E errors. And they would likely embrace the differentiated retirement policy simulated here. But the ex post point of view seems to be difficult to ignore completely. Remember that it consists of paying

22 And so is also the relative importance to be given to type-F vs type-E errors. In the context of retirement, most observers would probably consider that granting early retirement to someone who is still in good health (type-E error) is less of a problem than denying that right to someone who is in poor health (type-F error). But, again, that discussion goes beyond the scope of this chapter.

Ageing without Ageism?

211

attention to these important but unaccounted ‘within socio-economic-group’ variations of health status. Retirement age differentiation would, almost inevitably, be based solely on ‘between-group’ statistical (thus ex ante estimated) differences and ignore the rest. But, as we have shown above, allowing retirement age to differ across six groups (three educational attainment levels for each gender) would account for at most 9 per cent of country-level health variance. If what matters socially is equalizing each individual’s health upon retirement (i.e. ex post equalization), then the gains from abandoning a uniform retirement age policy as a proxy for realized health and longevity appear limited. Of course, other policies than differentiated retirement age can be implemented. And some of our results are supportive of this option. For instance, the sheer magnitude of ex post health status differences across individuals may justify upstream public-health or related policies designed to combat health inequality, even at the early stages of life. Also, the importance of the unaccounted inter-individual health inequalities within our retirement groups probably calls for a less ex ante approach and a more individualized and ex post treatment of health differences. Would it be feasible? Probably not as part of a retirement policy. But maybe it could be done via disability insurance or, more precisely, the screening procedure that determines the eligibility of individuals to disability benefits. Contrary to the statisticians telling retirement policymakers what could be done by exploiting ex ante health differences across socio-demographic groups, the doctors in charge of the screening are (at least potentially) assessing each individual’s realized health. And this probably puts them in a position to achieve what theorists call ex post equality. In many countries, disability benefits are closely linked to oldage pension systems. And their role is to provide, on a case-by-case basis, ‘retirement’ opportunities to people who suffer from ill health but aren’t yet eligible for proper retirement money. And it is also expected that workers who receive disability benefits subsequently shift to the old-age pension system once they reach the official retirement age. This raises the question of which policy is best suited to account for health and related longevity inequalities. Should policymakers go for a more elaborate but still proxy-based retirement policy with several retirement ages? Or should they stick to what has been the historical norm, that is, a uniform retirement age, supplemented by disability benefits conditional on an individualized but time-consuming,23 possibly stigmatizing and also error-prone assessment of health status?

Acknowledgments The content of this text owes a lot to Vandenberghe (2021a). The research exposed here was financially supported by ARC Research project No. 18–23-088 Sustainable, Adequate and Safe Pensions (SAS Pensions) 2018–2023, funded by the Fédération Wallonie-Bruxelles. 23 And thus costly.

212

Differentiating Retirement Age to Compensate for Health

References Atalay, Kadir, and Garry F. Barrett. (2015). ‘The Impact of Age Pension Eligibility Age on Retirement and Program Dependence: Evidence from an Australian Experiment’. Review of Economics and Statistics 97 (1): 71–87. Bassanini, Andrea, and Eve Caroli. (2015). ‘Is Work Bad for Health? The Role of Constraint versus Choice’. Annals of Economics and Statistics 119: 13–37. Bozio, Antoine, Clémentine Garrouste, and Elsa Perdrix. (2021). ‘Impact of Later Retirement on Mortality: Evidence from France’. Health Economics, 12 March. https://doi. org/10.1002/hec.4240. Börsch-Supan, Axel, Martina Brandt, Christian Hunkler, Thorsten Kneip, Julie Korbmacher, Frederic Malter, Barbara Schaan, Stephanie Stuck, Sabrina Zuber, on behalf of the SHARE Central Coordination Team (2013), Data Resource Profile: The Survey of Health, Ageing and Retirement in Europe (SHARE), International Journal of Epidemiology, 42(4), 992–1001. Case, Anne, and Christina Paxson. (2005). ‘Sex Differences in Morbidity and Mortality’. Demography 42 (2): 189–214. Cornia, Giovanni A., and Frances Stewart. (1993). ‘Two Errors of targeting’. Journal of International Development 5 (5): 459–496. Costa, Dora L. (1998). ‘The Evolution of Retirement: An American Economic History, 1880-1990, National Bureau of Economic Research, Inc, Cambridge, MA, p 226 . De Nardi, Mariacristina, Eric French, and John Bailey Jones. (2016). ‘Savings after Retirement: A Survey’. Annual Review of Economics 8(1): 177–204. Diamond, Peter A. (1967). ‘Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparison of Utility: Comment’. Journal of Political Economy 75: 765–765. Fleurbaey, Marc, Marie-Louise Leroux, Pierre Pestieau, and Grégory Ponthiere. (2016). ‘Fair Retirement under Risky Lifetime’. International Economic Review 57 (1): 177–210. Gosseries, Axel, and Vincent Vandenberghe. (2016). ‘Augmenter l’ˆage de la retraite: La seule réponse possible au vieillissement?’ Le Soir, 29 January, p. 22. OECD. (2019). ‘Pensions at a Glance’, OECD, Paris, p.180. Oeppen, Jim, and James W. Vaupel. (2002). ‘Broken Limits to Life Expectancy’. Science 296 (5570): 1029–1031. Oliveira Martins, Joaquim, Frédéric. Gonand, Pablo- Antolín, Christine de la Maisonneuve, and Kwang- Yeol Yoo. (2005). ‘The Impact of Ageing on Demand, Factor Markets and Growth’. Organisation for Economic Co-operation and Development (OECD) Publishing, Economics Department Working Papers 420. Pestieau, Pierre, and Maria Racionero. (2016). ‘Harsh Occupations, Life Expectancy and Social Security’. Economic Modelling 58 (C): 194–202. Pijoan-Mas, Josep, and José-Víctor Rios-Rull. (2014). ‘Heterogeneity in Expected Longevities’. Demography 51 (6): 2075–2102. Ponthiere, Grégory (2020). ‘A Theory of Reverse Retirement’. Journal of Public Economic Theory 22: 1618–1659. https://doi.org/10.1111/jpet.12458.

Ageing without Ageism?

213

Vandenberghe, Vincent. (2021a). ‘Differentiating Retirement Age to Compensate for Health Differences’. IZA Journal of Labor Policy 11 (1): 1–34. Vandenberghe, Vincent (2021b). ‘Health, Cognition and Work Capacity beyond the Age of 50: International Evidence on the Extensive and Intensive Margin of Work’. International Labour Review 160: 271–31.

15 Ageing in Place and Autonomy Is the ‘Age-Friendly’ City Initiative Too Elderly-Friendly? Kim Angell

1. Introduction Imagine that you are an urban planner with a finite amount of public money to spend on developing a city’s environment. What policies should you support? According to an increasingly influential approach in urban development, you should make the city age-friendly (Buffel et al. 2014: 53–54). An age-friendly city gives its elderly inhabitants the opportunity to ‘age in place’ by modifying its built and social environments in ways that offset the biological and psychological effects of aging (Scharlach 2016: 325; WHO 2017a: 40). By enabling people to sustain their favoured projects and practices where they currently live, an age-friendly city promotes personal autonomy (FRA 2018: 5; WHO 2020: 8); it enables ‘older people to retain the maximum control over their lives’ (WHO 2017b: 12). The age-friendly city initiative (henceforth, the AFC initiative) has received much attention from social scientists. Empirical studies range from accounts of national or regional differences in what people regard as age-friendly (Moulaert and Garon 2016) to how age-friendly proposals might be efficiently implemented (Greenfield et al. 2013; Buffel et al. 2014). Much less has been written on the initiative’s desirability. Addressing that gap, this chapter analyses whether the AFC initiative’s own appeal to personal autonomy can justify its elderly-friendly policies. The chapter proceeds as follows. I first elaborate the parts of the AFC initiative that I shall focus upon, including its appeal to autonomy. I then turn to the philosophical literature. After outlining a suitable account of personal autonomy, I consider what it implies for the desirability of the AFC initiative’s policies. I argue that older and younger city inhabitants may, in principle, have similar autonomy-based claims to favourable urban development. On the plausible assumption that resources will, in practice, be scarce, I offer some prioritization recommendations for our imagined urban planner before I briefly conclude.

Kim Angell, Ageing in Place and Autonomy. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Kim Angell (2023). DOI: 10.1093/oso/9780192894090.003.0016

Ageing without Ageism?

215

2. What is an age-friendly city? Since the World Health Organization (WHO) launched the AFC initiative in the 1990s and 2000s, more than 1,000 cities and communities worldwide have become part of its Global Network for Age-Friendly Cities and Communities—each committing to adopting age-friendly policies. The AFC initiative covers eight domains where a city might be age-friendly: (i) outdoor spaces and buildings, (ii) transportation, (iii) housing, (iv) social participation, (v) respect and social inclusion, (vi) civic participation and employment, (vii) communication and information, and (viii) community support and health services (WHO 2007, 2016, 2017a). Although cities may focus only on a subset of domains (Scharlach and Lehning 2016: 219), most such initiatives remain quite comprehensive, aiming to minimize ‘physical barriers to aging well’, to enhance ‘social engagement’, and to optimize ‘multidimensional health and well-being’ (Scharlach and Lehning 2016: 203–204). Taken together, the eight domains reach well beyond the traditional core of urban planning: a city’s built environment. For reasons of space, I delimit my analysis to domains (i)–(iii)—those pertaining to that core. Here are some central policy examples. In domain (i): making public spaces and areas of commerce easily accessible, for example, retrofitting with universal design features, better street lighting, auditory street crossing devices; making important amenities locally available (creating or maintaining ‘20-minute neighborhoods’); equipping outdoor places with benches and public toilets; and preserving built features that support the elderly’s sense of belonging to and identification with the area (WHO 2017a: 22–23). In domain (ii): making public transportation services accessible (adding bus stops, routes and vehicles with universal design features); broadening pavements to motivate and facilitate walking; and providing on-demand shuttle bus services (WHO 2017a: 33–34). In domain (iii): providing affordable public housing in various sizes; offering support for modifications to and maintenance of private homes; and arranging for intergenerational flat shares (WHO 2017a: 48–50). So much for the content of age-friendly policies. How are they justified? As mentioned, an age-friendly city promotes the ability of its elderly people to age in place. Having such an opportunity (ostensibly) matters because individuals have a morally significant interest in personal autonomy, in having control over their own lives. The WHO writes: age-friendly environments ‘ensure that older people age safely, continue to develop personally and contribute to their communities while retaining autonomy and health’ (2020: 8); such environments enable people ‘to be and do what they value throughout their lives’ (2020: 9). Similarly, the European Union underlines the importance of being able to live ‘a dignified life, defined by choice, control and autonomy and participation, whatever one’s age’ (FRA 2018: 5). For present purposes, I shall adopt the AFC initiative’s view that autonomy has considerable value. The elderly’s interest in autonomy is thus a good candidate for grounding their moral ‘claim’ to age in place. When the elderly have such a claim, it is of genuine (although not necessarily overriding) importance that we satisfy it. If

216

Ageing in Place and Autonomy

the AFC initiative’s policies do enable the elderly to age in place, the urban planner seems justified in recommending them.1 However, whether those policies are indeed justified from the viewpoint of autonomy remains an open question. The reason is the potential mismatch between the AFC initiative’s concrete policies and its purported justification in autonomy. While the former almost invariably focus on the elderly’s autonomy, the latter presents selfgovernance as valuable throughout people’s lives. If autonomy matters throughout people’s lives, the elderly’s autonomy-based claim to age in place might be overridden by the competing autonomy-based claims of younger inhabitants. Depending upon the prevalence and strength of such competing claims, the city might have reason to supplement, or even replace, the AFC initiative’s ‘elderly-friendly’ policies, either with ‘young-friendly’ ones (which promote young people’s autonomy) or ‘all-age-friendly’ policies (which promote people’s autonomy regardless of age).2 In sections 3–5, we explore those issues. We first consider what it means to have personal autonomy. We then consider whether the elderly’s interest in autonomy grounds their purported claim to age in place. Finally, to determine whether the AFC initiative’s elderly-friendly policies are justified, we consider whether younger inhabitants have competing autonomy-based claims.

3. Personal autonomy: A broad account The philosophical literature on personal autonomy is comprehensive, and I cannot do justice to the various views on offer here. I shall, instead, focus on what I call a ‘broad’ account of autonomy, one which seems congenial to the AFC initiative’s justification of its policies. According to that broad account, to have personal autonomy a person must: (a) have certain mental abilities for decision-making, including the ability to rationally fit means to ends and to critically assess the ends themselves (i.e. her desires, values, projects) with a view to ‘authenticate’ them as her own; (b) have access to an adequate set of options from which to choose her projects in life; and (c) enjoy non-interference with her chosen projects. Those conditions are necessary and jointly sufficient for autonomy.3 The three-fold account is ‘broad’ because it includes aspects of a person’s psychology—condition (a)—as well as her environment—conditions (b)–(c). Some favour a ‘narrow’ account of autonomy, which focuses strictly upon a person’s mental abilities (see, e.g. Taylor 2005). Proponents of narrow views may claim that ‘an impoverished person in the slums of Chennai’ has the same ‘capacity to make decisions on the basis of his ends and values’ as ‘an affluent Englishman living in a leafy suburb of London’, although ‘owing to his greater wealth it is more likely that the 1 Policies might, of course, be justified for other reasons than their ability to promote claims. I bracket that here. 2 Whether the candidate policies have the intended effects is a large empirical question that I cannot address. I therefore assume that the policies do promote the relevant interests. 3 For views along such lines, see Raz (1986: 154–155, 372–378); Oshana (2006).

Ageing without Ageism?

217

Englishman would be able to exercise his autonomy to realise his goals’ (Taylor 2005: 153). For present purposes, I take no stand in the debate between broad and narrow accounts of autonomy. Note, however, that on a narrow understanding of autonomy, the AFC initiative’s (allegedly autonomy-based) justification of its elderly-friendly policies might be dead on arrival. After all, those policies aim to promote the elderly’s autonomy by modifying the (built) environment in ways that enable them to pursue their ends. Because an elderly person’s ability for means–ends rationality and critical assessment of ends may remain the same regardless of whether her ends are realized or frustrated, the AFC initiative’s policies may tend to have little, if any, (narrow) autonomy-promoting effect. In contrast, from the perspective of the broad account, the city’s decision to modify its built environment or to stand back and leave the elderly to their own devices may have very significant effects upon their autonomy. Because the broad account connects a person’s autonomy to aspects of her environment, it might provide a more charitable interpretation of the AFC initiative’s appeal to autonomy. I shall therefore use the broad account when assessing the desirability of the initiative’s elderly-friendly policies.⁴ I want to emphasize four aspects of the broad account, as I presently understand it. First, a person’s interest in the three conditions for autonomy plausibly extends throughout her life. As a matter of fact, ‘[p]eople usually control their lives not by deciding once and for all what to do for the rest of their lives. Rather they take successive decisions, [. . .] sometimes reversing earlier decisions, sometimes further implementing them, and often dealing with matters unaffected by the earlier decisions’ (Raz 1986: 374, n. 1). We could, of course (implausibly), expect a person to choose a detailed plan of life while young and stick to it throughout her life. If so, condition (b) would seem to matter only in that initial choice situation. As I understand the present account, however, an autonomous person is able to continuously revise her projects. Without that ability she would not be in control of her life. Retaining access to adequate options therefore matters, even to a person who, as it happens, continually reaffirms her initial choices. Second, although a person must retain access to adequate options, the content of those options may vary, both over time within a person’s life and between persons. What is adequate for a young inhabitant who has grown up in the city may change as she becomes older. Similarly, what is adequate for a young city dweller might not be adequate for people from rural areas, whose projects, family, and social ties are located there. Third, a person’s own choices typically influence what counts as a worthwhile life—and hence which options will be adequate going forward. That way of exercising control over one’s future ends is an important part of what it means to be a self-governing agent, one who controls and shapes her own life (Raz 1986: 387).

⁴ Even if proponents of narrow accounts cannot defend the AFC initiative’s policies in the name of autonomy, they might, of course, defend them on other grounds, for example, in the name of (distributive) fairness.

218

Ageing in Place and Autonomy

Fourth, a person’s autonomy (typically) matters regardless of the content of her chosen options. She may autonomously choose a life of religious devotion, for example, or an atheist conception of the good. From the viewpoint of autonomy, her interests in mental ability, access to adequate options, and non-interference have the same moral significance either way.⁵ To sum up, to live an autonomous life, a person must be mentally able, have access to adequate options, and enjoy non-interference with the options she chooses. Importantly, and as echoed by the AFC initiative’s appeal to autonomy, a person’s interest in those three conditions extends throughout her life. Because living autonomously matters regardless of age, we have initial reason to doubt that people’s autonomybased claims in the city, whatever they turn out to be, will follow any strict age-based pattern. To explore this further, however, let us now start with the claims of the elderly.

4. Claims of the elderly to age in place We have seen that people have a three-fold interest in autonomy: in non-interference, access to adequate options, and mental ability. Does that interest ground the elderly’s purported claim to age in place? As we shall see, while neither interest component succeeds on its own, they might make a decent case when taken together.

4.1 Non-interference and negative claims to stay Many believe that a person has a claim to stay in a particular place if that is necessary for them to pursue their self-chosen projects (Waldron 2004; Lefkowitz 2015; Moore 2015; Stilz 2019). Imagine Sally, an elderly inhabitant who regularly pursues various social, cultural, and other activities in her current neighbourhood. Sally’s interest in non-interference grounds a claim not to be removed from that area and to freely access the locations where those activities take place. That might include its shops, restaurants, cinemas, museums, parks, places of worship and so on. Although she might perform similar types of activities elsewhere (in other neighbourhoods, cities, or countries), the fact that her self-chosen projects revolve around her particular neighbourhood makes a moral difference. It gives third parties autonomy-based reason against displacing Sally; their doing so would interfere with the specific life she has chosen. The individual claim to stay is normally defended with regards to relatively large places, such as a specific region, city, or district (Stilz 2013; Moore 2015). However, the autonomy-based ground for such claims may, in principle, work, I believe, also for larger (e.g. countries or continents) and for smaller (e.g. neighbourhoods or streets) ⁵ This points towards the connection between autonomy and the liberal idea of state neutrality (Sneddon 2013: 152–164).

Ageing without Ageism?

219

places. The size of the relevant place is determined by the content of the claimant’s projects. Some people’s projects might be satisfiable if they can stay in a specific city. Others might need to stay in a particular neighbourhood, building, or even room. At any rate, their claim to stay may, in principle, be supported by the same interest in non-interference. Is there a necessary connection between the weight of a person’s claim to stay and the time she has spent in the relevant place? I think not. Although people may tend to have weightier claims in places where they have lived for a long time, the increased weight has nothing to do with time itself. Instead, their claims are weightier in so far as more (or at least more central parts) of their self-chosen projects have come to revolve around the relevant place (Stilz 2013). This means that newcomers in a place, in so far as they quickly develop the requisite plans, might, in principle, have equally strong claims to stay as long-time residents. The claim to stay might protect a person not only from physical (or ‘direct’) displacement but also from ‘indirect’ displacement. A person who physically stays, might nonetheless experience indirect displacement ‘when incoming residents and/or businesses [. . .] [c]hange the feel, tastes, norms, and desires of an existing neighborhood, replacing the preferences or desires of existing residents’ (Versey et al. 2019: 4634). If the interest in non-interference with located activities protects against indirect displacement too, it may justify those policies that secure the elderly’s sense of belonging to and identification with their neighborhood.⁶ Although the interest in non-interference may ground a negative claim to stay, that is not enough to enable the elderly to age in place. According to the AFC initiative, an age-friendly city shall not merely (‘negatively’) refrain from interfering when the elderly make use of, say, a neighbourhood park in its current form. The city shall also (‘positively’) provide resources for making that park more accessible (e.g. by adding benches and toilets) to offset the effects of ageing. Such positive provision arguably sits at the core of the claim to age in place. Indeed, it is what enables the elderly to sustain their place-dependent activities above a certain level of functional ability. Because appeal to non-interference cannot ground the positive components of the claim to age in place, let us consider whether the elderly’s interest in having access to adequate options might do that job.

4.2 Access to ‘adequate’ options and positive claims to environment modification Recall that an autonomous person controls her own life. She shapes herself into a specific person with her own valued projects and identities. Those self-chosen features influence what counts as an adequate option set for her in the future. If you prefer a tranquil countryside life (Severinsen et al. 2016), having options in the city ⁶ For many, city life is an important source of identity and communal belonging—complementing (or even replacing) more traditional, nation-based identities (Bell and de-Shalit 2011).

220

Ageing in Place and Autonomy

might be insignificant. For someone like Sally, in contrast, who cares much about her urban activities, the option of ageing in the city with (roughly) sustained functional ability might well be part of any adequate set. If it is, then Sally’s interest in access to adequate options might ground a claim to age in place in that way. In other words, her interest in the second condition for autonomy might ground a claim to receive the resources she needs to sustain the relevant ability. The elderly’s interest in access to adequate options might thus straightforwardly justify the ‘positive’ claim to resource provision. Presumably, that may justify various elderly-friendly policies across domains (i)–(iii). That includes provision of universally designed outdoor and indoor spaces, modes of transport, and affordable housing. At this stage, one might question whether the elderly’s interest in autonomy is weighty enough to justify such (presumably costly) resource provisions. I cannot do more here than note this issue. For present purposes, I assume that the interest in autonomy is indeed weighty enough—as the AFC initiative implies—to ground the relevant claim, at least in principle. Whether the elderly’s autonomy-based claims justify the relevant elderly-friendly policies will of course further depend upon the existence and weight of competing autonomy-based claims held by younger inhabitants, an issue we shall return to below. Moreover, even if we find that the elderly's autonomy-based claims do prevail, their overall justification will of course still depend upon how they fare when we consider concerns other than autonomy. To keep things manageable, however, I presently restrict my analysis to what autonomy implies for the elderly’s claims. Bear in mind, though, that the present justification of the positive elderly-friendly policies thus becomes importantly contingent.

4.3 Mental ability Finally, consider the interest in having mental ability. Might that interest ground (aspects of) the claim to age in place? According to Cara Nine, ‘the ability to perform actions consistent with one’s commitments’ may come to rely upon having access to a particular dwelling place: one’s home (2018: 242). The idea is that by structuring their home environment in ways that facilitate their attainment of ends, ‘individuals can outsource their practical rationality to their environment’, making ‘certain actions easier or more difficult’ (Nine 2018: 249). When your home environment supports your goals and daily routines (say, eating and sleeping well, socializing, and exercising regularly), it may function as an agency-enabling ‘extended mind’ (2018). If Nine is correct, an elderly person’s interest in mental ability might ground a claim to age in her particular dwelling place. The same reasoning, I suggest, might support claims to age in other specific places—a particular neighbourhood, city, and so on—in so far as they are similarly crucial for cognitive ability.⁷ For the city’s elderly, however, ⁷ See Nine (2018) for some evidence that neighbourhoods and cities have such significance.

Ageing without Ageism?

221

the dwelling place is presumably the most important cognitive support, given how much time they generally spend there, in the United Kingdom, for example, 70–90 per cent (WHO 2017a: 40). I do not want to overstate the importance of this separate justification of elderlyfriendly policies, however. For one thing, the lapse in ‘cognitive efficiency’ experienced by a displaced person might be temporary. After ‘a period of adaptation’ to a new dwelling place, she might re-establish the ‘cognitive crutches that had helped [her] with [her] daily routines’ (Nine 2018: 250). Moreover, note that a person’s interest in mental ability might also support relocation measures. Imagine a person who, despite living at home, falls below an acceptable ability threshold. If she could best improve by moving to a nursing home, her interest in mental ability might support such relocation. Finally, in cases of relocation, resources could be provided to help people adapt (more) rapidly to their new homes. Those points suggest that people’s interest in mental ability might work best as a buttressing justification for elderly-friendly policies.⁸

5. Claims of the young We have seen that the elderly’s three-fold interest in autonomy, when taken together, may, in principle, ground a claim to age in place. We must now consider whether members of other age groups may also have autonomy-based claims in the city. If they do, the AFC initiative’s elderly-friendly policies might sit uncomfortably with its ostensible concern for promoting people’s autonomy ‘throughout their lives’ (WHO 2020: 9) and the declaration that autonomy is valuable ‘whatever one’s age’ (FRA 2018: 5). In this section, I shall focus on affordable and functionally adequate housing, which is arguably an essential part of people’s projects and pursuits in the city regardless of age. Above, we saw that the elderly’s claims depended upon the content of their chosen ends. The same holds for younger persons. Some might prefer to leave the city and settle in the countryside such that limited city housing options pose no threat to their autonomy. For others, continuing to live within reasonable distance of where they have their urban ties—say, their social, cultural, and family connections—might be crucial. If so, their interest in access to adequate options may ground a claim to affordable city housing. In most cases, the housing claims of younger inhabitants are presumably quite generic: they can be satisfied by having access to affordable housing somewhere in the city (not in any particular neighbourhood). Some young people, however, might have more particularized claims. Again, that depends upon the content of their projects.⁹ Imagine a young refugee who has just been granted residence in the city. ⁸ It is worth noting that the interest in mental ability will work also on narrow accounts of autonomy. ⁹ Because claims depend upon the content of people’s ends, one might worry about adaptive preference formation. Imagine that a city defaults on its duty to provide affordable housing. Poorer residents might

222

Ageing in Place and Autonomy

Although the city is still largely unfamiliar to her, having finally found a safe haven, she immediately starts to build her dreams and projects for the future around her new neighbourhood. Her interests in non-interference and access to adequate options might then ground claims to stay and to the housing provision she needs to continue her self-chosen activities in that particular place going forward. One might wonder whether a claim’s weight is somehow influenced by its degree of ‘particularization’. More specifically, one might wonder if the claim of an elderly inhabitant—which is typically more particularized (e.g. to age in her current home or neighbourhood)—is, for that reason, weightier than the more generic claim of a young inhabitant. For what it is worth, I doubt that a claim’s degree of ‘particularization’ will have such significance on its own. For both persons, it is their interest in having access to adequate options that grounds their relevant claims. Recall that we are concerned with promoting the autonomy of city inhabitants ‘throughout their lives’—that autonomy matters ‘whatever one’s age’, Imagine that what is at stake for the elderly person is whether the city gives her the option of remaining in her current flat (through subsidized retrofitting), or whether she must move to a nursing home in a different neighbourhood. For the young person, what is at stake is whether the city gives her the option of subsidized housing somewhere within reasonable distance of, say, her family or study place, or whether she must move far away. If the relevant options are, indeed, essential for the adequacy of those people’s respective sets, then it seems to me plausible to regard their claims as equally weighty, all else being equal. In practice, the city might satisfy people’s housing claims in various ways. Alternatives include: subsidizing people’s renting or buying on the private market (Yates 2012); passing ‘inclusionary’ zoning laws that require housing developers to include low-cost flats in their building projects (Padilla 1995); supporting ‘community land trusts’ which allow people to privately own low-cost homes (although restricting commercial resales) while owning land collectively (Saegert 2015); and building affordable public housing (Bloom 2008). Which policies to pursue will, of course, depend upon what empirically works, a topic that I cannot address here.1⁰ What seems clear enough, however, is this. Because different inhabitants presumably have place-dependent projects in different parts of the city, affordable options should be made available across its neighbourhoods.

6. Prioritization under scarcity Thus far, I have suggested that the AFC initiative’s appeal to autonomy may support claims to favourable urban development for old and young city dwellers alike. On the then reluctantly decide that continued city life is unsustainable and revise their plans accordingly. Their initial housing claims might then dissolve, which might seem problematic. 1⁰ Scholars disagree, for instance, about the effects of inclusionary zoning (see, e.g. Mukhija et al. 2010; cf. Stein 2018).

Ageing without Ageism?

223

plausible assumption of resource scarcity, however, we cannot satisfy all such claims. There will be conflicts between them—both between and within age groups. As an urban planner with the task of promoting city inhabitants’ autonomy, and with finite resources at your disposal, which claims should you prioritize?

6.1 Age effects, cohort effects, and lifetime assessment: Prioritizing today’s young The answer, of course, will depend upon the empirics of each case. In some cities, the age-variable may indeed track increasing autonomy deficits among inhabitants, in which case elderly-friendly policies should be prioritized. In other cities, the picture might well be mixed, or even reversed. If a person’s age is a quite imperfect indicator of people’s capacity to live autonomous lives in the city, it is worth asking whether other variables might allow us to say something further about how the urban planner should prioritize. That other variables are at least relevant seems clear enough. The income/wealth variable is a good example. After all, people might buy the relevant means for preserving their functional ability in the city on the private market, that is, requisite home modifications, private transportation services, et cetera. All else being equal, it is simply easier for a rich elderly inhabitant to sustain her place-dependent projects. That inequality is accentuated by the strong correlation between high income/wealth and better health/less need such that given levels of functional ability might effectively be preserved at lower costs for those who are already financially well off. So, even though there is obviously an age effect at play concerning people’s need for requisite modifications to the built environment, that effect is indirect. If being socio-economically disadvantaged is what effectively undermines an elderly inhabitant’s ability to age in place, the urban planner has reason to prioritize elderly inhabitants in the low income/wealth category. When we consider cohort effects, the age variable might actually work against prioritizing those who are currently old, even when they are also socio-economically disadvantaged relative to their peers. In many parts of the world, today’s urban planners may have reason to favour those disadvantaged individuals who are currently young. Why? Consider the ongoing global rise in individual socio-economic inequality, an inequality which cumulates over the lifespan and materializes in old age (OECD 2017). According to a recent OECD report, such inequalities are projected to increase in many of its member states, resulting in significantly higher inequalities between future retirees within the same cohort (2017: 38). The OECD projects, moreover, that the development of socio-economic inequality between cohorts will change direction. Reversing a pattern that has been stable since the 1910s, people born after the 1960s are projected, on average, to be poorer at the age of retirement than their parents. Due to the 2008 financial crisis, that tendency is expected to be especially pronounced for those who reached adulthood after 2000 (OECD 2017: 20).

224

Ageing in Place and Autonomy

Because socio-economic inequality increases both within and between cohorts (and the inter-cohort development disfavours today’s young), the urban planner has reason to prioritize those young people who are at significant risk of ending up with the most severe autonomy deficits in old age. All else being equal, their projected autonomy deficits (e.g. lack of adequate options) might well be much worse in absolute terms than those of current retirees. The latter, for example, typically own their (largely mortgage-free) homes—making them better off in terms of wealth and, hence, in terms of access to options. In contrast, the currently young risk never becoming homeowners at all. When suggesting that the urban planner prioritizes today’s young, we have measured individual autonomy deficits from a complete life (or ‘lifetimist’) viewpoint (see Gosseries: Chapter 6, this volume). Many believe that, when distributing benefits and burdens between persons, we should consider the morally relevant features over people’s whole lifetimes instead of focusing (only) on time-slice comparisons. The idea is to ask whether today’s young adults will suffer worse autonomy deficits overall than current retirees when comparing their whole lives. If the answer is yes, and if we care about reducing such gaps, we have reason to implement young-friendly policies. Doing so would prioritize the interests of those who are (otherwise) expected to experience the lowest degrees of autonomy during their lifetimes, regardless of their current age. Autonomy matters regardless not only of a person’s age but also of her cohort membership. It is therefore unwarranted for the urban planner to prioritize today’s elderly. Doing so would require that current retirees suffer from worse (complete-life measured) autonomy deficits than today’s young are expected to experience. If the OECD projections are accurate, however, the reverse is true. Today’s urban planners therefore have reason to devise young-friendly policies. Doing so would help to decrease inter-cohortal inequalities in people’s ability to live autonomous lives. Implementing elderly-friendly policies here and now would do the opposite.

6.2 Intergenerational housing: Benefiting old and young simultaneously Grand predictions like that of the OECD, of course, often come with significant uncertainty. For that reason, some might be queasy about prioritizing today’s young poor. Moreover, doing so runs counter to the stated policies of most age-friendly city initiatives, which do focus largely upon benefiting the elderly. Taking a precautionary approach, one might perhaps accept the importance of promoting the young poor’s autonomy, while insisting that we do so, whenever possible, by implementing all-age-friendly policies, those which also benefit the elderly. Some of the WHO initiative’s policies do qualify as all-age-friendly (see chapter 7, this volume). They include better street lighting and more parks, benches, and public transport options. The problem is that such policies might have very modest effects, if any, on the young’s autonomy. With the possible exception of more efficient/rapid

Ageing without Ageism?

225

public transportation, the young’s autonomy would be much better promoted by providing them with affordable housing. Luckily, all-age-friendly intergenerational housing policies seem to tick the right boxes. As an example, in Norway, Oslo recently launched a trial project where young students are offered low-rent accommodation in nursing homes in exchange for social companionship with their elderly residents (Schwebs 2021). Intergenerational housing may also be arranged in the elderly’s private homes, where spare rooms might be offered to young people in need of lowcost accommodation. Several such schemes are already in place in various countries, arranged by more or less profit-oriented organizations (see HomeShare International 2019). The city could either subsidize such private services or set up its own public arrangement. As for benefiting the elderly, intergenerational housing might, first and foremost, have beneficial synergy effects in policy domains beyond those concerning the built environment. The city might, for example, offer young tenants an especially low rent if they agree to perform some simple tasks for the elderly homeowner, such as grocery shopping, cleaning, and cooking, which would otherwise have been done by a formal carer. By thus reducing the elderly’s need for costly formal home care, the intergenerational housing arrangement contributes to elderly-friendly policy objectives in the AFC initiative’s domain (viii): community support and health services. In addition, note that the elderly’s social participation also increases through companionship with their young tenants, an important objective in domain (iv). Other important policies in that domain include: combating ageism; promoting a positive image of ageing; and enabling the exchange of skills, experience, and knowledge between generations (WHO 2017a: 79–80). By increasing contact between young and old and preventing age-based spatial segregation, intergenerational housing arrangements, if well-designed, might promote those policies as well. Again, the said policies do not directly concern the city’s built environment. It seems clear enough, however, that they do have positive effects upon the elderly’s ability to live autonomous lives. (Reduced out-of-pocket expenses for home care, for example, might effectively expand the elderly’s access to adequate options.) Intergenerational housing is therefore a particularly interesting all-age-friendly policy for cities concerned with promoting the autonomy of all its inhabitants. Given the attractive synergy effects, there is good reason to place intergenerational housing arrangements high on any city’s list of candidate policies.11

7. Conclusion All in all, the present analysis gives reason, I think, for cautious enthusiasm about the AFC initiative—at least for a somewhat reformed version of it. The initiative is 11 Some arrangements, like New York City’s ‘tax lien sale’ (Krinsky 2016), should obviously be taken off any such list.

226

Ageing in Place and Autonomy

correct that the elderly’s interest in autonomy (broadly construed) might justify various modifications to a city’s built environment. In its current form, however, the AFC initiative is too insensitive—even by its own lights—to the interests of younger inhabitants, whose autonomy might also depend upon certain urban developments. In fact, given the increased socio-economic inequality that the currently young are projected to experience in old age, today’s urban planners have reason to reverse the AFC initiative’s priorities. At least, they have reason to favour all-age-friendly policies, such as intergenerational housing arrangements, which bring significant benefits to old and young alike. Whether those recommendations will also hold in coming decades is, of course, empirically contingent. Future urban planners might well confront different patterns of existing and projected autonomy deficits, which, in turn, call for different priorities. At any rate, what seems clear enough is this. From the viewpoint of autonomy, age in itself is no ground for special treatment. Finally, it is worth giving the urban planner one further recommendation. Before choosing her policies, she would do well to also consider other values and normative principles. After all, personal autonomy—important as it is—is not all that matters.

Acknowledgements I thank Greg Bognar and Axel Gosseries for very helpful comments on this chapter.

References Bell, Daniel A., and Avner de-Shalit. 2011. The Spirit of Cities: Why the Identity of a City Matters in a Global Age. Princeton, NJ: Princeton University Press. Bloom, Nicholas Dagen. 2008. Public Housing that Worked: New York in the Twentieth Century. Philadelphia, PA: University of Pennsylvania Press. Buffel, Tine, Paul McGarry, Chris Phillipson, Liesbeth De Donder, Sarah Dury, Nico De Witte, An-Sofie Smetcoren, and Dominique Verté. 2014. ‘Developing Age-Friendly Cities: Case Studies from Brussels and Manchester and Implications for Policy and Practice’. Journal of Aging & Social Policy 26 (1–2): 52–72. FRA (European Agency for Fundamental Rights). 2018. Shifting Perceptions: Towards a Rights-Based Approach to Ageing. Luxembourg: Publications Office of the European Union: European Union Agency for Fundamental Rights. Greenfield, Emily A., Andrew E. Scharlach, Amanda J. Lehning, Joan K. Davitt, and Carrie L. Graham. 2013. ‘A Tale of Two Community Initiatives for Promoting Aging in Place: Similarities and Differences in the National Implementation of NORC Programs and Villages’. The Gerontologist 53 (6): 928–938. https://doi.org/10.1093/geront/gnt035. HomeShare International. 2019. Nurturing a Global Network: Homeshare in Action 2014– 2019. Oxford: Homeshare International.

Ageing without Ageism?

227

Krinsky, John. 2016. ‘Passing the Buck on Affordable Housing’. Gotham Gazette, 12 May. https://www.gothamgazette.com/opinion/6327-passing-the-buck-on-affordablehousing. Lefkowitz, David. 2015. ‘Autonomy, Residence, and Return’. Critical Review of International Social and Political Philosophy 18 (5): 529–546. https://doi.org/10.1080/ 13698230.2014.927117. Moore, Margaret. 2015. A Political Theory of Territory. A Political Theory of Territory. Oxford: Oxford University Press. Moulaert, Thibauld, and Suzanne Garon. 2016. ‘Introduction: Toward a Better Understanding of AFCC’. In Age-Friendly Cities and Communities in International Comparison: Political Lessons, Scientific Avenues, and Democratic Issues, ed. Thibauld Moulaert and Suzanne Garon. International Perspectives on Aging. Cham: Springer, pp. 1–16. https://doi.org/10.1007/978-3-319-24031-2_1. Mukhija, Vinit, Lara Regus, Sara Slovin, and Ashok Das. 2010. ‘Can Inclusionary Zoning Be an Effective and Efficient Housing Policy? Evidence from Los Angeles and Orange Counties’. Journal of Urban Affairs 32 (2): 229–252. Nine, Cara. 2018. ‘The Wrong of Displacement: The Home as Extended Mind’. Journal of Political Philosophy 26 (2): 240–257. https://doi.org/10.1111/jopp.12133. OECD (Organisation for Economic Co-operation and Development). 2017. ‘Preventing Ageing Unequally’. Paris: OECD Publishing. https://doi.org/10.1787/9789264 279087-en. Oshana, Marina. 2006. Personal Autonomy in Society. Aldershot, UK: Ashgate Publishing. Padilla, Laura M. 1995. ‘Reflections on Inclusionary Housing: A Renewed Look at Its Viability’. Hofstra Law Review 23 (3): 539-626. Raz, Joseph. 1986. The Morality of Freedom. Oxford: Clarendon Press. Saegert, Susan. 2015. ‘Interrupting Inequality: Crisis and Opportunity in Low-Income Housing Policy’. Metropolitics, 26 January. https://metropolitics.org/InterruptingInequality-Crisis-and-Opportunity-in-Low-Income-Housing-Policy.html. Scharlach, Andrew. 2016. ‘Age-Friendly Cities: For Whom? By Whom? For What Purpose?’ In Age-Friendly Cities and Communities in International Comparison: Political Lessons, Scientific Avenues, and Democratic Issues, ed. Thibauld Moulaert and Suzanne Garon. International Perspectives on Aging. Cham: Springer International Publishing, pp. 305–329. https://doi.org/10.1007/978-3-319-24031-2_17. Scharlach, Andrew, and Amanda Lehning. 2016. Creating Aging-Friendly Communities. Oxford/New York: Oxford University Press. Schwebs, Ine. 2021. ‘Søstrene Mari og Helene bor på eldrehjem–Vi er aldri alene’. Universitas, 17 March. https://www.universitas.no/alternative-studenthjem/sostrenemari-og-helene-bor-pa-eldrehjem-vi-er-aldri-alene/133517. Severinsen, Christina, Mary Breheny, and Christine Stephens. 2016. ‘Ageing in Unsuitable Places’. Housing Studies 31 (6): 714–728. https://doi.org/10.1080/02673037.2015. 1122175.

228

Ageing in Place and Autonomy

Sneddon, Andrew. 2013. Autonomy. Bloomsbury Ethics. London: Bloomsbury Academic. Stein, Samuel. 2018. ‘Progress for Whom, toward What? Progressive Politics and New York City’s Mandatory Inclusionary Housing’. Journal of Urban Affairs 40 (6): 770–781. https://doi.org/10.1080/07352166.2017.1403854. Stilz, Anna. 2013. ‘Occupancy Rights and the Wrong of Removal’. Philosophy & Public Affairs 41 (4): 324–356. https://doi.org/10.1111/papa.12018. Stilz, Anna. 2019. Territorial Sovereignty: A Philosophical Exploration. Oxford: Oxford University Press. Taylor, James Stacey. 2005. ‘Autonomy, Inducements and Organ Sales’. In Philosophical Reflections on Medical Ethics, ed. Nafsika Athanassoulis. London: Palgrave Macmillan UK, pp. 135–159. https://doi.org/10.1057/9780230273931_7. Versey, H. Shellae, Serene Murad, Paul Willems, and Mubarak Sanni. 2019. ‘Beyond Housing: Perceptions of Indirect Displacement, Displacement Risk, and Aging Precarity as Challenges to Aging in Place in Gentrifying Cities’. International Journal of Environmental Research and Public Health 16 (23): 1-21. https://doi.org/10.3390/ ijerph16234633. Waldron, Jeremy. 2004. ‘Settlement, Return, and the Supersession Thesis’. Theoretical Inquiries in Law 5 (2): 237-268. https://doi.org/10.2202/1565-3404.1093. WHO (World Health Organization). 2007. Global Age-Friendly Cities: A Guide. Geneva: World Health Organization Press. https://apps.who.int/iris/handle/10665/43755. WHO. 2016. Creating Age-Friendly Environments in Europe: A Tool for Local PolicyMakers and Planners. Copenhagen: Regional Office for Europe of the World Health Organization. WHO. 2017a. Age-Friendly Environments in Europe: A Handbook of Domains for Policy Action. Copenhagen: Regional Office for Europe of the World Health Organization. WHO. 2017b. Global Strategy and Action Plan on Ageing and Health. Geneva: World Health Organization. WHO. 2020. ‘Decade of Healthy Aging: Baseline Report’. Geneva: World Health Organization. Yates, Judith. 2012. ‘Housing Subsidies’. In The Sage Handbook of Housing Studies, ed. David F. Clapham, William A. V. Clark, and Kenneth Gibb. London: Sage, pp. 397–418.

16 An Age-Based Delayed Housing Wealth Tax Daniel Halliday

1. Introduction: The economic generation wars? Most of the world’s societies, particularly those with developed economies, are now subject to ageing populations. With declining birth rates and increased life expectancies, populations are gradually becoming older. Accompanying this trend is a sense that today’s young people are faced with poorer economic prospects relative to those that were enjoyed by the retired or ‘baby-boomer’ generation. Attempts to justify this view typically make reference to a period of prosperity during the decades when baby boomers were of working age, which has been brought to an end by the onset of economic stagnation in the final decades of the twentieth century. There is a growing feeling, in some quarters, that society has entered a phase in which the elderly, having enjoyed favourable conditions throughout life, are being increasingly supported by younger people, who will not have it so good but whose taxes continue to enhance the lives of the old (particularly through the provision of health care). The narrative of a ‘selfish’ older generation benefiting at the expense of the young is now the subject of several book-length treatments.1 The above makes for a very general concern about intergenerational wealth inequality between two birth cohorts. These are those born in the mid-twentieth century (baby boomers) and those born towards the end of that century and perhaps the start of the twenty-first (millennials). It is notable that the concern is not wholly about intergenerational inequality as such. Rather, two narrow dimensions of inequality have some prominence. These are (i) inequality of housing wealth in particular, with baby boomers having benefited from a housing boom that has left younger people struggling or unable to buy a home; and (ii) that younger birth cohorts continue to pay taxes which fund the provision of state-provided services consumed disproportionately by the elderly. Both of these factors, to the extent that they occur, look unjust. They motivate a search for proposals that simultaneously aim to make housing

1 On the United Kingdom, see Willetts (2020). On the United States, see Sternberg (2019). The problem of young people accessing home ownership in the United Kingdom is the focus of Timperley (2020). See also Ryan-Collins et al. (2017: 181–182). Daniel Halliday, An Age-Based Delayed Housing Wealth Tax. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Daniel Halliday (2023). DOI: 10.1093/oso/9780192894090.003.0017

230

An Age-Based Delayed Housing Wealth Tax

more accessible to younger people whilst also shifting the tax burden somewhat onto older homeowners. In what follows, I will defend a proposal of this sort. One immediate question is whether we are facing a special case of inequality between just one pair of birth cohorts (the millennials and the baby boomers) or just the latest iteration of what might be a recurring, or even constant, inequality between old and young. The presence of birth-cohort inequality can easily be mistaken for a more permissible kind of inequality between age groups. Age-group inequality can still involve younger people subsidizing benefits consumed by older people. But cross-subsidizing of one age group by another is not obviously troubling, particularly if members of a younger age group can expect to consume similar benefits once they become old, subsidized by a later generation of young people. The relative permissibility of age-group inequality is traceable to the fact that a society with large inequalities between age groups could still be subject to total equality between people in different birth cohorts once their entire lives are compared against each other.2 In effect, then, birth-cohort inequality can be associated with the sort of one-off events, or historical phases, that work against the possibility that successive generations enjoy equality once their complete lives are taken into account. While the popular narrative frames the baby-boomer generation as benefiting from the sacrifices of millennials, it was being argued some decades ago that the baby boomers were being burdened by policies that benefited persons born between 1920 and 1945.3 That being said, the prosperity enjoyed by the baby-boomer cohort may still have been substantial enough to count as a historical anomaly, coinciding with a period of economic growth that had no precedent and has little likelihood of coming back.⁴ I won’t try to settle this question here. Whether we are, indeed, dealing with a one-off instance of severe birth-cohort inequality involving boomers and millennials or merely a recurring inequality between age groups may depend on which goods it is whose distribution is given significance: at a very general level, there may, indeed, be a recurring phenomenon, subject to variations across jurisdictions, whereby members of a recent birth cohort absorb some costs so that elderly members of a cohort can receive certain benefits.⁵ At this general level, I am unsure whether it is right to single out the relationship between millennials and baby boomers as particularly special. What I take to be somewhat distinctive, however, is the tendency for millennials to face high obstacles to home ownership while paying relatively high income taxes, where these taxes subsidize benefits consumed by the prior cohort, many of whom 2 Here, I follow Bidadanure (2021), who provides a fuller defence of these claims. In addition, see the entries in this volume by Valente (chapter 17) and Pestieau & Ponthiere (chapter 18). 3 Thomson (1994) raised the concern that the baby boomer generation was in fact subsidizing the cohort of those born between 1920 and 1945, based on an analysis of New Zealand welfare policy. This was not especially about access to housing wealth. ⁴ Here, I allude to the narrative of the ‘U-shaped curve’ introduced by Piketty (2018). For another detailed account of how the standard of living rose and fell in accordance with economic growth, focusing on the United States, see Gordon (2016). An informative discussion of the way in which the babyboomer generation coincided with important changes to housing markets is Ryan-Collins et al. (2017: esp. Chapter 4). ⁵ To get a sense of the variation across countries with respect to inequalities in the prospects of successive birth cohorts, see Chauvel (2013).

Ageing without Ageism?

231

have been able to accumulate housing wealth and who may have worked during an era when taxes were more progressive than they are now. So described, this inequality may be one between two specific birth cohorts rather than a recurring inequality between age groups. Let me offer a few remarks about how taxes and housing have normative significance. It is philosophically uncontroversial that the distribution of the tax burden can be unfair: paying tax is clearly burdensome in the first place, and the distribution of burdens matters from the point of view of justice. If younger people carry a relatively high tax burden overall, while being materially worse off than older people, then taxation is thereby open to the objection that it is regressive in the familiar sense. The relationship between housing and justice is less evident than that between tax and justice. While housing is clearly a good that is distributed in some manner, the harder question is why the more specific good of home ownership is important and why justice might imply a distinction between homeownership and any alternative in which people access housing without owning it. There is relatively little philosophical work explaining why home ownership is more important than (mere) access to housing. One claim is that the housing boom reflects a problematic shift away from regarding homeownership as a means to personal security and shelter and towards treating homes as simply a financial asset (Ryan-Collins et al. 2017: Chapters 4–5; Timperley 2020: Chapter 6). Apart from that, the importance of home ownership receives some indirect defence in political philosophy by way of arguments drawing on the ways in which renting a home puts one at a disadvantage. Renters are exposed to domination by landlords or, at least, the precariousness associated with the relatively short lifespan of most rental contracts. Renting can have aggregate macroeconomic effects that may undermine justice. These include tendencies towards more transient (and thus less cohesive) communities as renters tend to move home more often than owner-occupiers. Some of these problems are exacerbated by periods of high growth in the market value of homes: the process of gentrification often breaks up established communities by displacing renters who become priced out (Huber and Wolkenstein 2018; Putnam 2021). It remains an open question as to whether such considerations count in favour of policies that can make it easier to actually own a home as opposed to better protections for renting. Some philosophical work has recently defended the latter idea (Wells 2019). Many people in practice still strive to become owners. This means that housing booms have negative consequences even if one thinks the case for homeownership is not philosophically strong. These include the problem of increased private debt due to the larger mortgages being taken on by younger people (Timperley 2020: Chapter 6). The above considerations could be discussed at greater length, and there remains room for the view that many defects of rental markets are merely contingent and would not exist under ideal conditions (Meyer 2018). Nevertheless, I suspect that a just society is one in which home ownership is relatively accessible to members of the community who want it. At any rate, I am going to assume that current inequalities of housing wealth are unjust in so far as they lead younger people to bear the brunt of large mortgages or the disadvantages of renting, while members of earlier cohorts

232

An Age-Based Delayed Housing Wealth Tax

enjoy the appreciation of their housing wealth. Under this description, inequalities of housing wealth are relatively distinctive when compared with other dimensions of wealth inequality that might have occurred even if housing wealth was equally distributed. Wealth inequality as a general phenomenon might raise issues of justice. But it is possible that we would be less troubled by wealth inequalities between birth cohorts if the prospects for home ownership had remained constant from one birth cohort to the next. My goal is to develop a proposal for addressing birth-cohort inequality that targets both the distribution of housing wealth and the distribution of tax burdens. More specifically, I defend what I call the delayed housing wealth tax. The core idea is that, past a certain age (roughly coinciding with a typical retirement age), homeowners have their housing wealth taxed according to both the value of their home and the number of years of post-retirement life for which they continue to retain ownership of their home. This tax is designed so as to incentivize retired homeowners to sell their homes and ‘downsize’ to a home of lower value, thereby reducing, or even removing, their liability to the tax. This serves to increase the supply of larger homes for younger people seeking to accommodate a family and/or live in areas with employment opportunities. Homeowners who enjoy especially high longevity, and accordingly consume a larger quantity of state services than those who die younger, are taxed more as a result. Because the tax is charged to the estate of retirees only after death, its economic feasibility is relatively strong, and, indeed, homeowners can choose not to be deprived of any wealth while actually alive. Section 2 provides some foundations for this proposal and gives some reasons to reject alternatives that might otherwise seem promising. Section 3 develops and defends the idea of a delayed housing wealth tax itself. Section 4 then addresses some objections.

2. Inequality between birth cohorts versus inequality within birth cohorts The claim that millennials ‘have it worse’ than baby boomers is open to the objection that it privileges a narrow set of goods and opportunities. It may be fairly straightforward to establish that, at least on average, members of the baby-boomer cohort enjoyed superior prospects for wage growth and access to home ownership than members of the millennial cohort. But this is not to say that they necessarily had better lives overall. Being born in 1945 very often meant being exposed to various ills from which those born since 1985 have been largely spared. Baby boomers faced the possibility of conscription in a war (such as Vietnam) and the criminalization of sexual lifestyles now both legal and socially accepted. Millennials are much less exposed to these hazards. Baby boomers also had generally poorer access than millennials to a host of material goods and services that became affordable towards the end of the twentieth century, such as air travel and computer technology, not to mention

Ageing without Ageism?

233

substantial improvements in health care.⁶ Such factors are typically excluded from narratives that focus on economic prospects relating to labour market opportunities and the cost of housing. Justice, however, is about the benefits and burdens of participation in society broadly construed: whether a society is overall just should take into account these other variations in freedom and well-being that have changed over the course of the past century. Accordingly, a narrow focus on economic prospects relating to wages and housing might be arbitrary. Similar complications apply to whether one birth cohort might be subsidized by the next. Young people certainly do subsidize benefits consumed by older people, so far as taxpayer-funded benefits are concerned. But, again, it may be arbitrary (from the point of view of distributive justice) to focus on these ‘upward’ subsidies while ignoring other kinds of ‘downward’ transfers that do not involve taxes or the state services that they fund but that are nonetheless real. Members of the older generation routinely provide help to members of the younger generation. This occurs not just through wealth transfers but also through unpaid care work. Care for young grandchildren is usually crucial to the continued labour market participation, and hence wealth accumulation, of parents of pre-school children. In addition, there is the familiar point that preceding generations leave benefits for subsequent generations in a more collective sense, for example, by way of the total accumulation of knowledge and culture in society. These points should temper the popular narrative somewhat. What still remains is the fact that the distribution of housing wealth is skewed in such a way that younger people either cannot access homeownership or must do so by taking on historically large debt, whereas large amounts of housing wealth have been accumulated by some members of the currently retired cohort. Recognizing that the facts are more complicated with respect to any more general assessment of which cohort has ‘had it best’ may rule out especially proposals that are especially one-sided in favouring millennials over baby boomers. A proposal will go some way towards addressing inequalities with respect to the distribution of tax burdens and housing wealth but in ways that avoid presupposing that birth cohort inequality applies to the quality of life all things considered. Inequality between two birth cohorts can occur alongside substantial inequality within either or both of the birth cohorts concerned. Even if baby boomers are, on average, better off than millennials, certainly not all baby boomers have been able to accumulate substantial housing wealth. There is substantial inequality among young people, but still some earn enough for the path to homeownership to remain open. This may also be connected with inequality in preceding birth cohorts since wealthy parents and grandparents can transfer wealth to children and grandchildren (Wolff 2020). It is important that proposals do not end up addressing inequality between cohorts while also exacerbating or compounding inequality within cohorts.

⁶ One might still argue in the other direction: perhaps, earlier birth cohorts enjoyed more privacy prior to the emergence of surveillance technology in the twenty-first century.

234

An Age-Based Delayed Housing Wealth Tax

It is instructive to register some proposals that can be rejected on grounds that they exhibit this defect. One approach would be to develop policies aimed at keeping people in the labour market for longer, for example, by increasing the retirement age. This would be one way of increasing the extent to which older birth cohorts absorb the cost of the state institutions that look after them by ensuring that older people pay income tax for longer. Also in favour of such proposals is evidence that continued labour market participation can enhance health and well-being later in life, though this may be dependent on the type of work and on participation being part-time rather than fulltime (Wester and Wolff 2010). While there is scope for reforming retirement age in ways that might promote justice (Halliday and Parr 2022), there are reasons to resist any blanket proposal that aims to keep people working into old age. It is typically ‘blue-collar’ work that is physically demanding and more burdensome to perform at an advanced age. So, there may be an element of unintended class discrimination in efforts to make older people stay in work. Moreover, the more physically demanding jobs are often those at the lower end of the income distribution, making it more likely that they are held by people who do not, in fact, have substantial housing wealth. Accordingly, any solution driven by getting people back into work would likely compound inequalities within the boomer birth cohort. Another approach would be to privatize services consumed disproportionately by the elderly and funded by the income taxes of younger people. Considered in isolation, self-subsidy may seem fairer than cross-subsidy. Again, though, the problem with this approach is that it will compound inequalities already present within the older generation. Not all old people can afford to purchase private services, and, again, this is particularly so if they don’t own a home that they might sell or ‘reverse mortgage’ so as to free up the necessary funds. There is also the fact that increased age often means a new dependence on certain services, for example, a need for public transport when declining eyesight makes it difficult to continue driving one’s car. While I don’t want to rule out the possibility that some moderate privatization proposals might be defensible, this is not an approach that is very attractive overall. To simultaneously achieve sensitivity to both inequality between birth cohorts and inequality within birth cohorts, we need a proposal that actually targets the distribution of housing wealth more directly without burdening baby boomers who are badly off.

3. Proposal: A delayed housing wealth tax As I have said, the popular sense that younger birth cohorts ‘have it worse’ owes much of its moral force to the fact that house prices, particularly in urban areas, have grown out of proportion to wages in recent decades. A natural approach, then, is to adjust the share of the tax burden so that the housing wealth of older people can be included in the tax base used to generate benefits disproportionately consumed by members

Ageing without Ageism?

235

of that birth cohort. One way of doing this is to adopt what I call the delayed housing wealth tax. To a first approximation, here is how the delayed housing wealth tax would work: owners of homes will be taxed some proportion of their home’s value for each year between retirement age and death. There are longstanding theoretical and practical problems with taxing wealth (Rakowksi 1999; Fleischer 2017). For example, homes are like most hard assets in having their value ‘unrealized’ until they are sold. To address this, the proposed tax would not be levied until death. In other words, the idea is that the tax office ‘bills’ estates according to how long a homeowner retained a valuable home during retirement. (Obviously, homeowners who are married or cohabiting will typically not die at the same time. The delayed housing wealth tax would be applied only after the death of the surviving spouse.) To the extent that this tax is not levied (because homeowners sell up and downsize so as to avoid it), it helps to stimulate an increase in the supply of housing. To the extent that this tax is levied (because homeowners retain their home in spite of it), it provides a means of funding the benefits that retired homeowners can be expected to consume during their advanced years. The delayed housing wealth tax accommodates the considerations about intragenerational inequality highlighted in section 2. First, it imposes no burden on retired people who do not own a home and could be designed so that little or no burden is imposed on people owning homes of relatively low value. Second, the selection of some common retirement age rather than actual retirement age limits class discrimination. Were the process to begin at actual retirement age, those able to do so would have an incentive to work longer, meaning those less able to do so might be unfairly disadvantaged. Accordingly, then, the tax would take effect at some universal age (perhaps 70), even for homeowners still working at this age. A key feature of this proposal is the incentive it creates for retired homeowners to ‘downsize’ in ways that may free up housing stock for younger people.⁷ Since the tax liability is incurred only at the point at which both of the retired couple have died, the tax could be avoided if a couple were to sell and buy a property of lower value. In this scenario, they would be permitted to keep the proceeds from the sale of the valuable home they have sold, less what has been charged according to years spent in the home between reaching retirement age and selling the home. If done on a large scale, downsizing would have the effect of making urban housing more accessible to younger people and may even cause a reduction in prices (or a slowing of inflation). It could also see older people moving away from expensive urban areas where employment opportunities are concentrated to areas where relative scarcity of high-paying jobs means lower housing values.⁸ From a macroeconomic perspective, this might again be a good thing: As well as improving affordability for younger home buyers in

⁷ Our discussion here connects with that of Kim Angell (in chapter 15) on whether social housing might exhibit a similar downsizing requirement. ⁸ See Angell (this volume: Chapter 15 ).

236

An Age-Based Delayed Housing Wealth Tax

urban areas, the movement of ‘cash-rich’ retirees to areas with lower property values can improve economic outcomes in those areas due to increased local consumption. Questions remain about how to further refine a delayed housing wealth tax. The gradual surrendering of value could be either linear or progressive in ways linked to longevity: if homeowners live into their nineties, they might be required to begin surrendering a larger portion of their remaining housing wealth per year. In so far as long-lived individuals tend to be wealthier, one might regard the tax as ‘doubly progressive’ as the more valuable homes will be the ones whose owners live longer. Another possibility is that the tax is progressive simply with respect to a home’s value rather than (or perhaps together with) the degree of longevity enjoyed. I will not take a more precise stance on these questions here or on the strength of any correlations between housing wealth and longevity. My selection of a universal age as the tax’s point of commencement requires some defence. One might think that the pressure to downsize should begin when homeowners conclude their careers. Since homes are often purchased for their geographic proximity to employment opportunities, it may seem reasonable to introduce incentives to downsize once homeowners have exited the labour market. On the other hand, homes are often purchased for their size, so that children can be housed, along with proximity to schools and other resources. So, one might think that the age at which children reach adulthood might be selected instead. In other words, the question is whether pressure should be put on retirees or on ‘empty nesters’ in so far as these are distinct groups. A compromise view introduces the tax at an age when both of these events are likely to have occurred, such as the age of 70.⁹ Designed in this way, the proposal thus gives weight, in effect, to both the criteria of retirement and ‘empty nesting’ as triggers for the commencement of the tax. Some thought needs to be given to retired couples who separate or retirees who remained single throughout life. Policies that treat monogamous couples or ‘households’ as an appropriate unit of regulation risk having problematic implications regarding their treatment of individuals not fitting this description. Policies around the distribution of parental leave, for example, often either discriminate against single parents or give cohabiting or married parents a perverse incentive to separate (Gheaus and Robeyns 2011). The case of a couple who separates may not raise any problems of differential treatment in the event that a jointly owned home is sold by a couple who then purchase individual homes of lesser value. This would just be a case of downsizing, albeit perhaps for motivations other than tax avoidance. A harder case emerges if a couple separates such that there is no sale of a valuable home, as might happen if one spouse moves out and rents a home somewhere else. (If the proposal is to impose reduced liability in this case, it might thereby incentivize separation.) Another hard case might come in the form of a couple where there is a large enough age differential that one spouse reaches the age of 70 some time before the other. ⁹ It is possible that an adult child will remain in or return to their parents’ home, perhaps to facilitate saving to enable a home purchase. One way of modifying a delayed housing wealth tax would be for homeowners to claim a ‘freeze’ for any year in which they were accommodating children. I lack space to discuss what reasons there might be both for and against such a modification.

Ageing without Ageism?

237

We need a proposal that has clear implications for how these different cases will be handled differently. I suggest that the right approach is to simply tax individuals rather than households. It is simple enough to tax a home’s value for each owner of the home who is aged above 70, running separate calculations for each spouse and then billing the estate when the second spouse dies. This will mean that spouses who co-own a home will pay more tax than single persons, holding fixed longevity and home value. In the case of spouses who reach 70 at different times, they are essentially treated the same as if they reached 70 at the same time since the tax is essentially applied to each year of each owner’s life in which they are still owning the home past this age.1⁰ In this way, there is no commitment to treating a monogamous couple as a single unit of regulation. Instead, there are separate tax calculations for each individual, with these both being individually tied to the home’s value and the longevity of the person in question. These simply work alongside each other in the case of retired cohabiting homeowners. An effect of this is that single, retired persons do not typically end up paying the same amount of tax as couples. In practice, someone who has never entered into a monogamous relationship is likely to own a smaller and therefore less valuable home. It is independently plausible, then, that they are subject to less of a tax burden than the combined burden imposed on a couple collectively. For couples that separate and cease to cohabit but do not downsize by selling the home in which they had been living together, the tax remains applicable only to the spouse who remains an occupier of the home.11

4. Objections One objection concerns the impact that our proposal has on whatever offspring have their inheritance reduced by the proposal. The objection here may be that a delayed housing wealth tax has some potential to be regressive: offspring who have at least one parent with high longevity will have their inheritance most greatly reduced by the tax, though, in practice, these children may well have incurred a larger burden than those whose parents die younger by way of more years spent providing care as their parents age. Thus, the tax is regressive in the sense that many of these individuals will suffer reduced inheritance in spite of being worse off in terms of having delivered a greater amount of care. These are hard cases. One possibility is that rich people tend to live longer, meaning that offspring who have their inheritance reduced are among the wealthier families 1⁰ In practice, a couple where one spouse reaches 70 some years later than the other may mean that the tax applies over a greater number of years (though not life years). This would mean more years of value appreciation and hence more tax to be paid. 11 One might wonder whether, as in the parental leave case, this feature creates an incentive for couples to separate. The spouse who moves out of the home will typically either incur rental payments or come to own another home on which the tax would apply. These costs offset the benefit of avoiding the tax, meaning that the incentive is unlikely to be strong.

238

An Age-Based Delayed Housing Wealth Tax

in the first place. The problem of informal (i.e. unpaid) caregiving is already subject to some discussion in political philosophy and there is an independent case for regarding it as unjust that some people find themselves caring for elderly relatives on an unpaid basis. For one thing, the burdens of caring for elderly family members are of a somewhat distinctive sort (Philp 2016). There are also longstanding tendencies for the provision of care work, whether paid or unpaid, to be highly gendered (Moos 2021). Any further proposal for providing relief to caregivers in these contexts could work as a modifier of the delayed housing wealth tax. Where the provision of care by an offspring can be demonstrated, an entitlement to inheritance might be restored, up to a point, perhaps commensurable with the opportunity cost (in labour market terms) of having provided care to an elderly parent (Hacker 2023). Another objection might be that moving home during old age is rather disruptive and that there is something to be said for not pressurizing retired couples to move house during a phase of life when this might be difficult. While some may find this objection intuitive, it is not especially compelling. First, it is very easy for retired couples to simply hold on to their home if they want to. The only ‘cost’ of doing this is that there will be less wealth around after they have died that can be bequeathed. Homeowners must simply choose between retaining housing wealth (when this is substantial enough to incur the tax) and being able to pass that wealth on. It is unclear why there should be an entitlement to have it both ways and retain wealth for as long as possible while also retaining the same entitlement to transfer that might have been exercised earlier on (Halliday 2013). At any rate, the situation of retired homeowners should be weighed against that faced by younger citizens, particularly ones with families, who incur the cost of being unable to access home ownership. Since the delayed housing wealth tax aims to induce downsizing in ways that provide relief to at least some of those currently struggling to access home ownership, there is a strong case for regarding any burden to homeowners as outweighed by the benefit to younger people of an enhanced housing supply. A third objection is that all of this might be better pursued through an increase in inheritance taxation. While there may be reasons of justice to adopt an inheritance tax alongside our proposal, a delayed housing wealth tax nevertheless has some features that an inheritance tax cannot easily capture. Among these is the fact that an inheritance tax applies to wealth other than housing. Accordingly, an inheritance tax will not be as effective in incentivizing retired homeowners to sell their home so as to make housing available for younger people. Indeed, downsizing is really just a means of converting part of one’s housing wealth into cash wealth. This will not help with the avoidance of inheritance tax except to the extent that the cash wealth is spent down, which may not always happen and is often neither easy nor rational. In sum, then, the two principal advantages of the delayed housing wealth tax over inheritance (or wealth transfer) taxes are that (i) it retains the presumption that there is something important about the distribution of housing wealth and (ii) it is designed to create incentives to induce retired homeowners to sell their home and make housing stock in urban areas more available to younger people. These points continue to

Ageing without Ageism?

239

hold even if an inheritance tax is made age-sensitive, that is, taxes estates at a higher rate according to the longevity of the decedent. I have discussed this possibility at greater length elsewhere (Halliday 2018). While an age-sensitive tax provides a means of giving moral weight to the fact that long-lived persons may consume a greater amount of state-provided benefits and might otherwise have much to recommend it, it still lacks the downsizing incentive that is an important feature of the delayed housing wealth tax. That being said, I have no objection to inheritance taxes or associated wealth transfer taxes per se, and they might perform a valuable function alongside any tax on housing wealth. Suffice it to say that the case for or against inheritance taxes depends on the wider view one takes about factors like inequality and wealth transfers with respect to the overall wealth distribution. A final objection, or perhaps just an observation, is that the delayed housing wealth tax does not really serve to redistribute wealth directly from the baby-boomer cohort to the millennial cohort. Any ‘redistribution’ is indirect, by way of baby boomers downsizing, so as to improve the supply (and hence affordability) of housing for people of working age. Apart from that, there is no actual redistribution even if there is scope for the tax burden to shift somewhat onto older people in cases where downsizing is foregone in favour of retaining a valuable home post-retirement. While there is a sense in which the tax is de facto levied on members of the millennial cohort (by depleting their inheritance), this is so in ways that burden only those from wealthier families, whilst all members of the cohort stand to benefit from the effects of greater downsizing on the supply of housing. What makes this more of an observation than an objection is that it confirms that our proposal avoids being especially radical. This is in keeping with our earlier qualifications about whether the popular narratives around inter-cohort inequality are accurate when they suggest that the baby-boomer generation really had things much better all things considered, something that may need to be established if more aggressively redistributive proposals are to prove defensible.

5. Conclusion I have argued that while narratives about intergenerational justice sometimes present an exaggerated picture of how much older birth cohorts have benefited at the expense of younger ones, there remains the problem of what to do about inter-cohort inequality with respect to accessing home ownership. I have defended a proposal that seeks to address this kind of inequality by imposing an age-based tax on retired homeowners, whose liability is calculated according to the value of a home and the number of years of retirement spent living in it. A chief theoretical advantage of this proposal is its ability to create an incentive for retired homeowners to sell their housing wealth, while also imposing a higher tax burden when homeowners enjoy longevity (and thereby consume more state-provided services). The proposal also shows sensitivity to inequality within generations by not targeting retired persons who do not

240

An Age-Based Delayed Housing Wealth Tax

possess housing wealth. While there is likely more that could be said about the details, an age-sensitive tax on housing wealth has some distinctive appeal and avoids some objections that could be made to its rivals.

Acknowledgements Work on this chapter was supported by an Australian Research Council Discovery Grant entitled Sharing the Wealth: Tax and Justice in the Slow Growth Era, grant ID# DP210100924.

References Bidadanure, Juliana. 2021. ‘Justice between Co-Existing Generations’. In The Oxford Handbook of Intergenerational Ethics, ed. S. Gardiner. Oxford: Oxford University Press. Chauvel, Louis. 2013. ‘Welfare Regimes, Cohorts, and the Middle Classes’. In Income Inequality: Economic Disparities and the Middle Class in Affluent Countries, ed. Janet. Gornick and Markus. Jantti. Redwood City, CA: Stanford University Press, pp. 115–141. Fleischer, Miranda. 2017. ‘Not So Fast: The Hidden Difficulties of Taxing Wealth’. Nomos 58: 261–308. Gheaus, Anca, and Robeyns, Ingrid. 2011. ‘Equality Promoting Parental Leave’. Journal of Social Philosophy 42 (2): 173–191. Gordon, Roger. 2016. The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War. Princeton, NJ: Princeton University Press. Hacker, Daphna. 2023. ‘Inheritance Law and the Challenge of Securing Care in Old Age: A Three-Pronged Solution’. In Inheritance and the Right to Bequeath: Legal and Philosophical Perspectives, ed. Hans-Christoph. Schmidt-am-busch, Thomas. Gutmann, and Daniel. Halliday. Abingdon: Routledge: 144–165. Halliday, Daniel. 2013. ‘Is Inheritance Morally Distinctive?’ Law & Philosophy 32 (5): 619–644. Halliday, Daniel. 2018. The Inheritance of Wealth: Justice, Equality and the Right to Bequeath. Oxford: Oxford University Press. Halliday, Daniel., and Parr, Tom. 2022. ‘Aging, Justice and Work: Alternatives to Mandatory Retirement’. In The Ethics of Aging, ed. Christopher Wareham. Cambridge: Cambridge University Press: 228–242. Huber, Jakob., and Wolkenstein, Fabio. 2018. ‘Gentrification and Occupancy Rights’. Politics, Philosophy & Economics 17 (4): 378–397. Meyer, Marco. 2018. ‘The Right to Credit’. Journal of Political Philosophy 26 (3): 304–326. Moos, Katherine. 2021. ‘Care Work’. In The Routledge Handbook of Feminist Economics, ed. Gunseli. Berik and Ebru. Kongar. London: Routledge, pp. 90–98.

Ageing without Ageism?

241

Philp, Mark. 2016. ‘Justice, Realism and Family Care for the Aged’. Social Philosophy & Policy 33 (1–2): 413–433. Piketty, Thomas. 2018. Capital in the Twenty-First Century, trans. A. Goldhammer. Cambridge, MA: Harvard University Press. Putnam, Daniel. 2021. ‘Gentrification and Domination’. Journal of Political Philosophy 29 (2): 167–187. Rakowski, Eric. 1999. ‘Can Wealth Taxes Be Justified?’ Tax Law Review 53: 263–376. Ryan-Collins, Joshua., Lloyd, Toby, and Macfarlane, Laurie. 2017. Rethinking the Economics of Land and Housing. London: Bloomsbury Press. Sternberg, Joseph. 2019. The Theft of a Decade: How the Baby Boomers Stole the Millennials’ Economic Future. New York: Public Policy Press. Thomson, David. 1994. Selfish Generations: The Aging of New Zealand’s Welfare State. Winwick: White Horse Press. Timperley, Chloe. 2020. Generation Rent: How Greed Killed the Housing Dream. Kingston Upon Thames: Canbury Press. Wells, Katy. 2019. ‘The Right to Housing’. Political Studies 67 (2): 406–421. Wester, Gry., and Wolff, Jonathan. 2010. ‘The Social Gradient in Health: How Fair Retirement Could Make a Difference’. Public Health Ethics 3 (3): 272–281. Willetts, David. 2020. The Pinch: How the Baby Boomers Took Their Children’s Future—and Why They Should Give It Back. London: Atlantic Books. Wolff, Jonathan. 2020. ‘Family Fortunes’. Law, Ethics & Philosophy 8: 73–85.

17 Two Types of Age-Sensitive Taxation Manuel Sá Valente

1. Introduction: Clement vs Obama People’s income-related fiscal obligations fluctuate with age because annual income tends to rise until middle age and to decline afterwards.1 In a hump-shaped income distribution, middle-aged people pay higher yearly income taxes than the young and the elderly. This chapter discusses two strategies to make income taxation rates sensitive to age. The first strategy is a form of cumulative income taxation, which takes into account all earlier fiscal years beyond the last one. It comes in two versions, a ‘total’ and an ‘average’ one. The former adjusts yearly tax rates to the total income a person earned until then, irrespectively of their age or the number of years they have worked. The latter adjusts yearly tax rates to the average annual income earned so far. Consider, for instance, Canadian politician Tony Clement’s 2004 proposal to exempt from income tax anyone who had not yet earned a total of $250,000 in life.2 This proposal belongs to the total version. It favours the young because total lifetime income increases as we grow older. Depending on how income evolves with age, this would not necessarily hold if we focused on average yearly income over people’s lifetime instead. The second type of age-sensitive tax that is relevant to us here is age-differentiated taxation. It adjusts the rate applicable to the previous year’s income to the taxpayer’s age. Just like the previous proposal, fiscal systems do not tend to include this proposal.3 Yet, they could. In 2008, for instance, US presidential candidate Barack Obama proposed to ‘exempt anybody aged 65 and older, and making no more than $50,000 per household, from paying income taxes’.⁴ While cumulative income taxation is indirectly age-sensitive, age-differentiated taxation is explicitly so. What (if anything) should egalitarians think about these two types of age-sensitive taxes? The broadly egalitarian rule I adopt is ‘maximin’. By this, I mean that we should prefer policies that make the worst off people (over their entire

1 See, e.g. Atkinson (2015: 41). This observation is robust across countries and levels of development (Lee et al. 2014). 2 See Clement (2004). 3 Contrary to fiscal obligations, fiscal entitlements are often explicitly age-based, for example, statutory retirement benefits. ⁴ See Solomon (2008).

Manuel Sá Valente, Two Types of Age-Sensitive Taxation. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Manuel Sá Valente (2023). DOI: 10.1093/oso/9780192894090.003.0018

Ageing without Ageism?

243

lives) better off than they would otherwise be under another set of rules, even at the cost of increasing inequalities.⁵ The chapter proceeds as follows. Section 1 presents the main reasons supporting cumulative income taxation. Section 2 looks at how cumulative income taxes affect fiscal obligations across life. Section 3 introduces a proposal for what the tax system should aim at when taxing income across life. I argue that it should aim at fiscal changes that benefit the young and the elderly but not the middle-aged. Section 4 questions whether cumulative income taxes can deliver this without age differentiation. Section 5 concludes that, while cumulative income taxation can benefit the young, age-differentiated taxes are necessary to benefit the elderly.

2. Age-independent justifications for cumulative taxation Like ordinary taxation, cumulative income taxation also taxes the income from the previous year. However, in defining the tax rate applicable to it, it considers earlier income years instead of only the last one. This strategy has been defended on essentially three grounds. (a) Addressing income volatility. Vickrey (1939, 1947) showed that progressive noncumulative taxation has less impact on those with stable incomes than on taxpayers with fluctuating ones.⁶ Hence, progressive income taxation based only on the previous year’s income disproportionately burdens those incurring high-income volatility. If low-income earners tend to have a more volatile income, maximin egalitarians would have a first reason to support cumulative income taxation.⁷ (b) Capturing lifetime circumstances. A second reason is that cumulative income taxes convey a better picture of the taxpayer’s lifetime circumstances.⁸ Considering income through a greater time lapse gathers more information about one’s economic life. In this regard, Levell et al. (2015: 5) provide a helpful analogy: to identify the better off through a snapshot of their yearly income is like deciding who the winning football team is based on who scored in the last five minutes. Such examples suggest that maximin egalitarians should get closer to the lifetime view. The lifetime view claims that the distribution of some good, say income, cannot be assessed unless we know about the distribution of that good over the lifetime of those involved.⁹ One is closer to identifying the lowest lifetime incomes if tax rates take into account all earlier income years beyond the last one.1⁰ ⁵ See, e.g. Daniels (1988: 501) and Van Parijs (2003). ⁶ In Chapter 18 of this volume, Ponthiere and Pestieau discuss another proposal from Vickrey, which consists of taxing bequests based on the age gap between the donator and the receiver. ⁷ See Batchelder (2003) for evidence of a significant and negative correlation between income volatility and income. ⁸ See, e.g. Bamford (2014: 2) and Levell et al. (2015). ⁹ See, e.g. Gosseries (2014: 66–67) and Chapter 6, this volume. 1⁰ Egalitarians who do not embrace the lifetime view can still accept the first and third reasons, suggesting that one does not need to accept the lifetime view to endorse cumulative income taxation.

244

Two Types of Age-Sensitive Taxation

(c) Reducing tax avoidance. Vickrey (1939, 1947) offers a third reason against annual non-cumulative taxation. He suggests that cumulative income taxes prevent people from moving their income across accounting periods to reduce their tax liability. Such time optimization strategies may reduce tax yields, with negative consequences for the least well off. Assuming that high-income earners are more likely to engage in such strategies than low-income earners, maximin egalitarians may be sensitive to this concern for a shrinking tax yield. Hence, cumulative income taxes can benefit the lifetime worst-off by (i) making sure they are not harmed by income volatility, (ii) relying on more appropriate criteria to identify the worst off, and (iii) preventing the better off from escaping their fiscal obligations. The two first reasons have a connection with the age dimension. As for (i), it is likely that young workers experience more year-to-year earnings volatility than prime-age workers.11 On (ii), the extent to which lifetime income is adequate to identify the worst-off seems to increase with age. The young have a shorter past and their future is filled with uncertainty, such as of premature death. Here again, the football analogy is helpful. Watching a game from the beginning (i.e. cumulative income) does not tell us who the winner is if the game just started, but it does at minute 90′ . The more that is known and the less predicted, the closer one is to knowing who the winner is. Despite their age dimension, these justifications of cumulative taxes do not directly appeal to any conception of justice between age groups. They appeal to fairness between taxpayers of different income patterns or circumstances, which does not explicitly involve a concern with age.

3. Consequences for the net income distribution across age groups Even though none of the three reasons explicitly involve concerns of justice between age groups, moving to cumulative income taxation would affect the net income distribution across age groups. Consider first the total version. Tax rates are calculated based on total lifetime income levels. It adds each annual income to what the taxpayer has earned so far. As long as people earn income each year they live, their total lifetime income will increase with age and so will the tax rate applicable to the last year. The elderly would be the wealthiest for accounting purposes in almost any income distribution. Even if yearly income were to decline from the age of 25 onwards, total cumulated income would still rise as people age. As long as annual income remains above zero, the total strategy taxes the young less at the expense of higher tax rates for older age groups. Alternatively, there is the average version.12 Like the total strategy, it adds each income earned so far. However, unlike the latter, it divides that amount by the 11 See, e.g. Venn (2011: 18–19). 12 For a defence of the average proposal, see Vickrey (1939, 1947) and, more recently, Bamford (2014).

Ageing without Ageism?

245

taxpayer’s age, years, or hours worked.13 The fact that the denominator tends to increase with age counterbalances the fact that the numerator (total lifetime income) is also rising with age. Hence, the average strategy delivers a gentler tax rate across life. While more subtle, the average proposal still changes the fiscal obligations of age groups. The primary source of differential fiscal treatment is that the average cumulative proposal taxes rising incomes less than standard, non-cumulative tax policy. This is key because income increases until midlife before decreasing afterwards.1⁴ The average lifetime income is thus lower than the annual income during the first half of a worker’s life. It is more beneficial for young people with increasing earnings to have tax rates that apply to their average lifetime income rather than to their annual income. Therefore, the average proposal taxes the young (and the young middle-aged) less than non-cumulative taxes when the income distribution is hump-shaped. How does the average proposal affect our fiscal obligations as we reach old age? One could say that the proposal is worse for the elderly because their income declines with age. However, that depends on which reference point defines the ‘worse off ’. If the point is what the elderly would otherwise pay under non-cumulative yearly taxation, the claim above can be false. If annual income at old age declines below average lifetime income, cumulative income taxes will indeed be higher than noncumulative income taxes. However, that is not the case if the elderly’s yearly income remains above their lifetime average. Average cumulative taxation would then cost the elderly less than non-cumulative taxation. All age groups may pay less in absolute terms with average taxes than they would with non-cumulative taxation, depending on how gross income evolves. For instance, average lifetime income is always lower than annual income if income is constantly rising. Of course, egalitarians do not necessarily want to reduce taxes for everyone. For that would reduce the fiscal transfers to the worst off.1⁵ Keeping the tax revenue constant, we know that taxing one group less entails higher relative taxes for others. Hence, the question becomes: which age group benefits in relative terms from the average proposal? In a hump-shaped income distribution, the average proposal burdens the young less than the elderly. It benefits those on the left side of the hump more than those whose income has already peaked. The oldest persons to benefit more from average cumulative taxes are the middle-aged, who stand on the top of the hump. Afterwards, they will benefit less.1⁶ The average proposal still favours the young relative to the elderly. 13 See Bamford’s (2014) argument for considering ‘hours worked’. As opposed to years lived, a benefit of considering hours worked is that it is neutral concerning the age at which people enter the labour market. 1⁴ See, e.g. Lee et al. (2014) for a more precise description of the hump-shaped curve of labour earnings. Interestingly, they find that labour earnings peak later in high-income than in low-income countries. 1⁵ Reducing the tax revenue affects those who rely the most on fiscal assistance (such as the elderly). In what follows, I assume that cumulative taxation does not alter our tax expenditure. 1⁶ See how Vickrey (1939: 386–387, 1947: 179–180) deals with the problem of declining incomes. One solution is to overtax when income increases above the lifetime average and credit when income declines below the latter. I discuss this solution in section 4.

246

Two Types of Age-Sensitive Taxation

Moving from non-cumulative to cumulative taxes changes our fiscal obligations across life. Both the total and the average versions benefit the young relative to older persons. Yet, there are two differences. First, while they are both more favourable to the young than non-cumulative income taxation, what drives this favourable treatment differs in both cases. In the total version, it has to do with the fact that cumulated income almost invariably increases with age, regardless of whether the distribution is hump-shaped, decreasing, or increasing. In the average version, the favourable treatment has to do with the dynamics of rising and declining wages across people’s lives. There is also a second difference: the average version does not disadvantage older persons as much as the total version. However, it also means that it does not benefit the young as much as the total proposal. What should maximin egalitarians think about these fiscal changes?

4. The old poor, the short-lived, and the hump-shaped tax rate Here, we must investigate what the tax system should aim at when taxing income across life. From a maximin perspective, we should prefer tax bases, rates, and yields that benefit those with the lowest lifetime incomes.1⁷ We should give strong priority to benefiting those who have the least amount of good over a complete life (in our case, the good is income). This assumes we accept the lifetime view—the view that we can only tell who has the least of x once we know about the distribution of x across the entire life of those involved.1⁸ At first, it seems that lifetime egalitarians are indifferent to how income evolves across life. If we go through the same income distribution as we age, there might be inequality at different ages, but there will be equality over entire lives. For example, consider the hump-shaped distribution. The elderly may have no complaint against a higher income during middle age if they have also benefited from it in the past. The young might have no similar objection if everyone has begun their life as a low-income earner. Many income distributions inside our lives can be consistent with treating people equally over their entire lives.1⁹ Still, lifetime egalitarians have reasons to care about the distribution of income within lives. To see why, it is essential to distinguish between two sources of disadvantage in life. Some people are resource-poor: they earn little in life because their annual income is low. On top, they can be longevity-poor: their lifetime income is 1⁷ See, e.g. Rawls (1971: xiii); Daniels (1988: 501); and Van Parijs (2003). For Rawls, the worst off are those who have the least ‘primary goods’, which he understands to be goods that ‘persons need as free and equal citizens [. . .] over a complete life’ (Rawls 1971: xiii). Rawls includes income in the index of primary goods (1971: 79). I assume that justice requires ‘maximinizing’ lifetime income independently of shares of other goods, including leisure, which means that one can also include the unemployed among the lowest lifetime incomes. 1⁸ The lifetime view does not necessarily involve only retrospective assessments, for example, going back in time to assess the history of one’s income. It is also open to prospective adjustments based on future expectations such as probabilities and expected lifetime income. 1⁹ On age and complete-life neutrality, see Gosseries (2014: 66–70) and Gosseries as well as Berndt Rasmussen, Chapter 6 and 1 in this volume.

Ageing without Ageism? Standard income and age-based tax rates

247

Proposed income and age-based tax rates

Figure 17.1 Standard and proposed distribution of income (full line) and tax rates (dotted line) Source: author.

low because they die early. Both dimensions matter. An income profile benefits both of these groups if it includes two components. First, it benefits the short-lived if it enhances income early in life.2⁰ Reducing income taxes early in life contributes to this. Second, the profile benefits the resource-poor if it smooths income during old age. Lowering taxes for seniors attenuates the extent to which income declines late in life. In hump-shaped income distributions, this recommends hump-shaped tax rates that benefit the young and the elderly.21 The rationale behind the second component is as follows. When thinking about distributing resources within life, perhaps behind a reasonably thick veil of ignorance, it is tempting to select a stable profile. If we had to decide which income profile is best while not having any reason to privilege specific periods over others, the reasonable choice would be to distribute income roughly equally across life.22 This ‘income-smoothing’ strategy entails redistribution from when one is wealthier to when one is poorer. It mainly benefits those whose annual income is low. The greatest challenge for those who lack resources is the risk of becoming poor or unable to maintain their living standards if they live long enough with a declining income. The resource-poor need a redistributive scheme that protects them against the risk of a long life ending in poverty.23 On top of this, they must also keep in mind the risk that their lives might be shorter than they expect. This brings us to the first component of our proposal, which enhances income early in life.2⁴ The fact that some of us live longer than others raises complicated questions. The main difficulty is that we often do not know ex ante who will end up being shortlived. We only know who the short-lived are once they die, after which compensation is no longer possible (assuming that post-mortem compensation is impossible). In 2⁰ See Fleurbaey et al. (2014). 21 To be sure, one can meet the second component by increasing the progressivity across income brackets of a non-cumulative tax system as that also contributes to greater income equality across life. That strategy will not be enough to enhance income early in life in accordance with our first component. The discussion that will follow concerns only our two types of age-sensitive taxation, and, in that regard, it assumes that progressivity across income brackets would remain constant. 22 See Daniels (1988). He recognizes that treating people differently at different ages can make one’s life better (than treating all ages equally). In this regard, lowering taxes on all young workers might bring such lifetime gains (e.g. Weinzierl 2011). 23 This is a standard rationale of old-age pension schemes: offering insurance against the risk of old-age poverty for long-lived individuals (Barr and Diamond 2006: 16). 2⁴ See Lazenby (2011) for the role that the assumption that ‘equal life length’ plays in Norman Daniels’s argument to distribute resources roughly equally across life as opposed to concentrating resources at a young age.

248

Two Types of Age-Sensitive Taxation

other words, the short-lived can neither be identified ex ante nor be compensated ex post.2⁵ Statistically, we know that some groups tend to live longer than others—for instance, the resource-poor are more likely to be short-lived.2⁶ While we can predict that some groups will live shorter lives than others, premature death remains largely unpredictable within these groups. All of these include unidentifiable shorter-lived people who, other things being equal, will come to have less of goods (e.g. income) over their entire lives. Premature death might be inevitable, but societies can mitigate its lifetime impact. If we can only compensate people before they die, compensating those who die young is possible if we concentrate goods (like income) early in the lifecycle. For instance, enhancing income early in life (instead of at old age) promotes income equality between the short- and the long-lived. There is a tension between the claims of the ex ante unidentifiable short-lived and those of the elderly (long-lived) resource-poor. The more resources we shift into old age to protect the poor’s long lives, the greater the inequality between the short- and the long-lived.2⁷ The higher the old-age income, the more lifetime income those who die prematurely lose. Aversion to a short life would recommend income to peak early in life and to decrease afterwards. However, that would not necessarily protect the resource-poor against the risk of a long life ending in poverty. The profile I propose attaches value to both concerns. It supposes that when thinking about how people would distribute resources across life, again behind a relatively thick veil of ignorance, they would attach importance to both risks: an early death and a poor end of life.2⁸ Any profile that enhances income early in life and smoothens income afterwards protects people against both risks. The maximin egalitarian income profile is at odds with the hump-shaped income distribution where the young and the elderly are poorer than the middle-aged, in two specific senses. First, the young’s tendency to be poorer than the middle-aged harms the short-lived the most. Second, the declining income after middle age harms the poorest among the elderly, who risk falling into poverty if they live a long time. All else equal, egalitarians should move towards hump-shaped tax rates that benefit the young and the old further but not the middle-aged. Hump-shaped income profiles should be dealt with through hump-shaped tax rates. Can cumulative income taxation (either total or average) deliver this without resorting to age differentiation?

5. Can we escape explicit age differentiation in tax rates? One way of achieving hump-shaped tax rates is through explicit age-based differentiation in the income tax rate, involving lower rates for the young and the elderly 2⁵ See Fleurbaey et al. (2014). 2⁶ See, e.g. Chetty et al. (2016). 2⁷ See, e.g. Gosseries (2022). 2⁸ There is a question about how much importance they would attach to each risk and whether people would have bought insurance against a short life. I agree with Ponthiere (2020) that they would.

Ageing without Ageism?

249

and higher ones for the middle-aged. However, these policies generally involve explicit age-based differential treatment. While such policies might be acceptable, they are objectionable in at least one way. It is because they invite the charge of ‘ageism’, making them less attractive to policymakers and to the general public compared to policies that do not explicitly rely on the age concept. Can cumulative income taxation achieve hump-shaped tax rates without explicit age-based differential treatment? Consider first the total proposal. We have seen that it benefits the young. The problem is that it disadvantages the elderly, including the poorest among them. To prevent it from harming the elderly poor, one cannot apply tax rates to total lifetime income once people reach old age. It is possible to do this without explicit tax differentiation. What I have in mind is a threshold of total lifetime income (as in the ‘Clement Plan’). Below the threshold, annual tax rates increase with total lifetime income levels. Above it, ordinary non-cumulative taxation could apply. This proposal does not involve explicit differential treatment based on age. Yet, it excludes the elderly from having tax rates that apply to total lifetime income. When their total lifetime income exceeds the threshold, the elderly pay tax rates based on their annual income, as they do now. Hence, this strategy will not reduce their fiscal obligations relative to noncumulative taxation. To increase the elderly’s income further, age-differentiated tax rates in their favour would be necessary. Consider now the average proposal. As seen earlier, the proposal favours the young and the middle-aged. Of the three age groups we have considered, the middle-aged is the only group that maximin egalitarians have no reason to favour, other things being equal. How to ensure that the average proposal benefits the elderly more than the middle-aged? I see three possible strategies. One strategy is to prevent middle-aged people from averaging their income. Excluding the middle-aged from the average proposal ensures that they do not benefit from it. As it stands, I do not think that this strategy must resort to explicit age differentiation. For instance, one could prevent people from averaging their income above a threshold of total lifetime income. That would be a hybrid of the average and the total strategy. Despite not benefiting the middle-aged, this strategy would, again, not reduce taxes for the elderly compared to non-cumulative taxation. A second strategy is to increase taxes when annual income rises above the lifetime average to credit taxpayers when their income falls below the average.2⁹ While it seems not to resort to age differentiation, the proposal has two problems. First, it increases taxes for the young. Their annual income also tends to rise above the lifetime average, even if they are poorer than the elderly. Since our concern with the short-lived prohibits net transfers from the poor young to the poor elderly, these transfers must increase with age (explicitly or implicitly) to exclude the young from paying them. Second, this strategy benefits the middle-aged and elderly alike when their income declines. Yet, as seen earlier, we should welcome a decreasing income at middle age, not at old age. Explicit age differentiation seems necessary to benefit incomes only when they decrease at old age. 2⁹ See n. 18.

250

Two Types of Age-Sensitive Taxation

Finally, there could be differentiated tax rates that are higher for the middle-aged and lower for the elderly. This would promote our aim. Yet, it resorts to explicit age differentiation. Accepting explicit age differentiation would make our aim easier to achieve since, so far, cumulative income taxes seem unable to achieve hump-shaped tax rates without it.3⁰ Can we escape explicit age differentiation in tax rates if we want to? The answer depends on the age of the taxpayer that justice requires benefiting the most. Moving to cumulative taxes without age differentiation is unlikely to benefit the elderly (at least, without also benefiting the middle-aged). Early in life, cumulative taxes can reduce tax burdens without fiscal benefits for the middle-aged. Using a threshold of total lifetime income allows us to do so without age differentiation. There is still a question about how much to reduce taxes early in life. Answering that question will have implications for whether tax rates adjust to total or to average lifetime income until the threshold. The former permits greater fiscal reductions than the latter. Hence, the choice between the total and the average proposal (until the lifetime income threshold) will depend on how far justice requires reducing the fiscal obligations of the young. The lifetime income threshold benefits both low-income earners working at their maximum capacity and those working few years at high incomes.31 Yet, it benefits high-income earners for a shorter time than low-income workers. While it might seem unfair to benefit young high-income earners, one must not forget that they remain exposed to the risk that their lives or careers are short-lived. Using a total lifetime income threshold benefits the short-lived while showing sensitivity to resource poverty as fiscal reductions will last longer the poorer one is. Nonetheless, that is not to say that the lifetime income threshold suffices to address the needs of the worst off. They need the most assistance not only at the beginning of life but also during their old age. Cumulative income taxation remains incomplete in that regard. In general, it does not seem to be necessarily better than non-cumulative taxation at reducing tax rates for the elderly. Therefore, it makes sense that Barack Obama proposes age-differentiated tax rates to benefit the elderly, whereas Tony Clement suggests total cumulative income taxation to favour the young.

6. Conclusion What should maximin egalitarians think about our two types of ‘age-sensitive’ tax regimes, that is, cumulative and age-differentiated income taxes? Generally, egalitarians have reasons to endorse a more age-sensitive scheme than the ordinary one. 3⁰ Personally, I do not consider all forms of explicit age differentiation to be wrongful. Still, I think it is better to be able to avoid it. It increases the range of possible policies that societies can use to advance a particular goal, which is especially valuable if people have strong preferences against certain ways of implementing social objectives (as they might regarding explicit age differentiation). 31 See, e.g. Bamford (2014: 16), who poses this point as an objection against the total lifetime income proposal.

Ageing without Ageism?

251

The best scheme for egalitarian purposes depends on the age of the taxpayer that justice requires benefiting the most: while cumulative income taxation can benefit the young, age-differentiated taxes remain necessary to benefit the elderly. To conclude, let me highlight four points that this discussion has neglected. One is that other normative views exist both within and outside egalitarianism. For instance, not all egalitarians embrace the lifetime view. If, instead, our concern is the worst off at any given time, the income profile will have to be smoother. The reason is that every better-off time implies, by comparison, a worse-off time that could be improved. Such views still demand hump-shaped taxes in a hump-shaped income distribution. Hence, they can welcome age-sensitive taxes as well. The second point is that egalitarians may prefer tax bases other than income, such as consumption, wealth, or endowments. To be sure, my claim does not rely on income being a suitable tax base. For instance, one could defend an age-differentiated or a total lifetime consumption tax. Arguably, the point I made about age-sensitive taxation can apply to any tax base. A third point left aside is the possibility of using tax expenditure to achieve the same income profile. In the end, what egalitarians should care about is the net effects of taxation. An alternative to changing fiscal obligations (what one pays) is altering fiscal entitlements (what one receives). Tax expenditure favours the young without explicit age differentiation if it prioritizes the lowest total lifetime incomes. For instance, the young could benefit from a one-off (or decreasing regular) grant bestowed at the age of majority or from a right to extra retirement income at a young age.32 To benefit the elderly, tax expenditure might have to rely on explicit age thresholds as it happens with retirement income. This focus on tax expenditure can also accentuate or attenuate our reasons to reduce the elderly’s fiscal obligations, depending on the level of old-age pension benefits they receive. The final point is that one might prefer income profiles that contradict the ‘fair’ distribution between age groups. For instance, income smoothing can sacrifice social mobility by preventing the worst off from ‘climbing the ladder’ as they age.33 There is a tendency to see something positive in ‘ending life on a high’, in starting from the bottom and ending on the top. Egalitarians still need to explain what makes ‘high ends’ attractive. Taking high ends seriously can be a reason to enhance income at old age, or earlier, depending on when life ends.

Acknowledgements I am especially grateful to Greg Bognar and Axel Gosseries for their help and to Kevin Hartmann, Niko Väänänen, and Vincent Vandenberghe, the ARC project ‘Sustainable, Adequate and Safe Pensions’ (2018–2023), and the participants of the CRÉ, 32 See Valente (2022) on basic income and Ponthiere (2020) on retirement schemes. 33 On fairness and mobility and the asymmetry between downward and upward mobility, see Swift (2004: 8).

252

Two Types of Age-Sensitive Taxation

Parr & Chaire Hoover Annual Workshop 2021 for their comments and suggestions. I acknowledge financial support from the convention ARC No. 18-23-088 (‘SAS’ Pensions research project, UCLouvain). I remain fully responsible for any mistakes.

References Atkinson, Anthony. 2015. Inequality: What Can Be Done?. Cambridge, MA: Harvard University Press. Bamford, Douglas. 2014. Rethinking Taxation: An Introduction to Hourly Averaging. London: Searching Finance Limited. Barr, Nicholas, and Peter Diamond. 2006. ‘The Economics of Pensions’. Oxford Review of Economic Policy 22 (1): 15–39. Batchelder, Lily L. 2003. ‘Taxing the Poor: Income Averaging Reconsidered’. Harvard Journal on Legislation 40: 395–452. Chetty, Raj, Michael Stepner, Sarah Abraham et al. 2016. ‘The Association between Income and Life Expectancy in the United States, 2001–2014’. Journal of the American Medical Association 315 (16): 1750–1766. Clement, Tony. 2004. ‘Canada at a Time of Tectonic Shift—Opportunity Knocks for Conservatives’. Interview by Ian MacDonald. Policy Options, March: 5–10. Daniels, Norman. 1988. Am I My Parents’ Keeper?: An Essay on Justice between the Young and the Old. Oxford: Oxford University Press. Fleurbaey, Marc, Marie-Louise Leroux, and Gregory Ponthiere. 2014. ‘Compensating the Dead? Yes We Can!’ Journal of Mathematical Economics 51: 28–41. Gosseries, Axel. 2014. ‘What Makes Age Discrimination Special: A Philosophical Look at the ECJ Case Law’. Netherlands Journal of Legal Philosophy 43: 59–80. Gosseries, Axel. 2022. ‘Ageing, Unequal Longevities and Intergenerational Justice’. In Cambridge Handbook of the Ethics of Ageing, ed. Christopher Wareham. Cambridge: Cambridge University Press, pp. 212–227. Lazenby, Hugh. 2011. ‘Is Age Special? Justice, Complete Lives and the Prudential Lifespan Account’. Journal of Applied Philosophy 28 (4): 327–340. Lee, Ronald, Sang-Hyop Lee, and Andrew Mason. 2014. ‘Charting the Economic Life Cycle’. Population and Development Review 34: 208–237. Levell, Peter, Barra Roantree, and Jonathan Shaw. 2015. ‘Redistribution from a Lifetime Perspective’. Institute for Fiscal Studies Working Paper W15/27: 1–90. Ponthiere, Gregory. 2020. ‘A Theory of Reverse Retirement’. Journal of Public Economic Theory 22 (5): 1618–1659. Rawls, John. 1971. A Theory of Justice. Cambridge, MA: Harvard University Press. Solomon, Deborah. 2008. ‘Obama Courts Seniors with Tax Break’. Wall Street Journal, 28 June. https://www.wsj.com/articles/SB121461191851712527. Swift, Adam. 2004. ‘Would Perfect Mobility Be Perfect?’ European Sociological Review 20 (1): 1–11.

Ageing without Ageism?

253

Valente, Manuel Sá. 2022. ‘Basic Income and Unequal Longevity’. Basic Income Studies 17 (1): 1–14. Van Parijs, Philippe. 2003. ‘Difference Principles’. In The Cambridge Companion to Rawls, ed. Samuel Freeman. Cambridge: Cambridge University Press, pp. 200–240. Venn, Danielle. 2011. ‘Earnings Volatility and Its Consequences for Households’. OECD Social, Employment and Migration Working Papers 125: 1–36. Vickrey, William. 1939. ‘Averaging of Income for Income-Tax Purposes’. Journal of Political Economy 47 (3): 379–397. Vickrey, William. 1947. Agenda for Progressive Taxation. New York: The Ronald Press. Weinzierl, Matthew. 2011. ‘The Surprising Power of Age-Dependent Taxes’. Review of Economic Studies 78 (4): 1490–1518.

18 An Age-Differentiated Tax on Bequests Pierre Pestieau and Gregory Ponthiere

1. Introduction Whereas the taxation of inheritance or bequests is a widespread practice around the world, existing fiscal systems involve tax rates on bequests that do not—at least explicitly—depend on the age of the deceased.1 If a person dies and leaves a bequest to their descendant, this bequest will be, in a given country, taxed at the same rate independently from whether the donator died at age 40 or 95. Would it be socially desirable to make the taxation of bequest depend (explicitly) on the age of the deceased? If yes, on what grounds could one justify an agedifferentiated tax on bequests? In which direction should the age of the deceased affect the rate at which bequests are taxed? An age-differentiated tax on bequests is far from an uncontroversial proposal. This may explain why no country in the world applies an age-differentiated tax on inheritance. The goal of this chapter is twofold. First, we defend an age-differentiated tax on bequests, with a tax rate increasing with the age of the deceased: the older the deceased is, the higher the tax on their inheritance should be. Second, although there exist several distinct arguments supporting such an age-differentiated tax on bequests, we argue that the most persuasive argument is a compensation for short longevity. It states that, in a world of imperfect information about durations of life, an age-differentiated tax on bequests is a non-substitutable way of improving the situation of individuals who turn out to be short-lived. This chapter thus proposes to defend a particular kind of age-differentiated bequest tax (i.e. increasing with the age of the deceased) on a particular ground, that is, the compensation of unlucky prematurely dead individuals in a world of imperfect information. This chapter proceeds in two stages. In a first stage, we review some theoretical arguments supporting, on various grounds, an age-differentiated tax on bequests. Some arguments support a tax rate on inheritance that is decreasing with the age of the deceased; others support a tax rate that is increasing with the age of the deceased. 1 This does not mean, of course, that the tax rates that are applied do not account, indirectly, for the age of the deceased. For instance, if wealth accumulates (respectively, decumulates) with age, and if bequest taxation is progressive, the tax rate on bequest is implicitly increasing (respectively, decreasing) with the age of the deceased. Pierre Pestieau and Gregory Ponthiere, An Age-Differentiated Tax on Bequests. In: Ageing without Ageism?. Edited by Greg Bognar & Axel Gosseries, Oxford University Press. © Pierre Pestieau and Gregory Ponthiere (2023). DOI: 10.1093/oso/9780192894090.003.0019

Ageing without Ageism?

255

Then, in a second stage, we propose a critical assessment of those arguments, and we argue that the most persuasive one is based on compensation for premature death, which supports a taxation of inheritance that is increasing with the age of the deceased.

2. Strategic testamentary dispositions and tax revenues Within the public economics literature, one of the first arguments for taxing bequests differently depending on the age of the deceased is the one formulated by Vickrey (1945 [1994]). To be precise, the proposal made by Vickrey is not a tax based on the age of the deceased but on the age gap between the donator and the receiver of the bequest. However, assuming that the age of the receiver is constant and that the donator is the deceased, this proposal amounts to imposing a tax rate on bequests that increases with the age of the deceased. Vickrey’s argument is motivated by the goal of minimizing strategic testamentary dispositions, which can lead to (legal) fiscal avoidance. That argument relies on the idea that a uniform, age-invariant, tax on bequests has the undesirable feature of influencing the form of testamentary dispositions. More precisely, such a uniform bequest tax pushes the donator to transmit their wealth with as few steps as possible (e.g. directly to grandchildren or even great-grandchildren) in order to avoid taxation of the bequest. There is thus a danger that donators adopt transmission strategies in order to reduce the amount of tax they have to pay, leading to tax avoidance and losses of fiscal revenues for the state.2 Here is an example. Suppose that an elderly person would like to transmit their wealth b to their family. This elderly person has, for simplicity, one child and one grandchild. Under a uniform tax rate τ on bequests, the transmission to the child will make the child receive (1 – τ)b.3 Then, when the child will transmit the wealth to their own child, the amount transmitted will now be (1 – τ)2 b. But if the parent transmits the wealth directly to the grandchild, the amount transmitted is equal to (1 – τ)b. As a consequence, if the elderly parent wants to transmit their wealth to their descendants, it is more profitable to transmit it directly to the youngest descendant by minimizing the number of intermediate steps. As the example illustrates, the uniform tax on bequests has the consequence of influencing the form of testamentary dispositions. Rich elderly persons will systematically opt for transmission with as few steps as possible. In order to avoid such an influence on the tax, Vickrey proposed that the tax rate on bequests should be increasing with the age gap between the donator and the receiver. If the tax is properly designed, this makes the burden of the tax invariant to the number of steps in 2 This possibility of fiscal arbitrage arises because the tax system does not usually take into account the frequency of transfers (i.e. their relation with past/future transfers) but only their levels. 3 For simplicity, we assume that the interest rate equals 0.

256

An Age-Differentiated Tax on Bequests

the succession.⁴ According to Vickrey (1945 [1994]: 136), such a tax would imply that the same burden is imposed on the transfer of a given sum of money from one generation to another regardless of the number of steps in which it is done. To see how this proposal could make the tax on bequest neutral for testamentary dispositions, let us suppose that, in our previous illustration, there is a ℓ-year age gap between each generation (i.e. the donator is aged 2ℓ, their child is aged ℓ, and the grandchild is a newborn) and denote by τℓ the tax rate on bequest when the age gap between the donator and the receiver is equal to ℓ. If the parent adopts the strategy of transmitting the wealth directly to their grandchild, the amount transmitted is (1 – τ2ℓ )b, whereas, in case of a two-step transmission, it is equal to (1 – τℓ )2 b. Hence, there is indifference between the two transmission strategies if, and only if, τ2ℓ = τℓ (2 – τℓ ), which implies τ2ℓ > τℓ , that is, a tax on bequest that is increasing with the age gap between the donator and the receiver. Thus, imposing a tax on bequest that is increasing in the age gap between the donator and the receiver is a simple way to avoid strategic testamentary dispositions leading to losses of fiscal revenues for the state. That argument can be easily translated into an argument supporting a tax on bequests that is increasing with the absolute age of the donator, provided the age gap between the donator and the receiver increases with the age of the donator.⁵ Vickrey’s argument faces two main criticisms. First, from the perspective of avoiding strategic testamentary dispositions, taxing bequests based on the degree of parenthood is a better policy response than bequest taxation based on the age of the deceased. If, for instance, a large age gap between a donator and a receiver (their son) is due to the late birth of the latter, there is little support for applying a higher tax rate on inheritance. Thus, Vickrey’s argument can only be regarded as providing a second-best argument for age-differentiated taxation of bequests, the age gap between the donator and the receiver being taken as a proxy for the degree of parenthood between them. Second, Vickrey’s argument relies on the goal of raising fiscal revenues. However, as argued by Rawls (1971: section 43), the primary function of inheritance taxation is not to raise fiscal revenues but, rather, to correct for inequalities in economic power, inequalities that often affect the quality of political democracy. If one adopts Rawls’s view, inheritance taxation rules should be designed to achieve more social justice. From that perspective, a major problem is that Vickrey’s taxation rule may, in some cases, conflict with basic intuitions about a fair taxation of inheritance.⁶ Take the example of two rich old individuals, John and Jim, with the same characteristics and the same families, except that the grandchild of Jim was victim of a fraud and fell into poverty. Because of that circumstance, Jim prefers to transmit his wealth directly to ⁴ This argument assumes that the age gap between the donator and the receiver can be taken as a proxy for the degree of parenthood between them (see below). ⁵ This is the case when the donator must wait longer in order to be able to transmit their wealth to a more distant descendant. ⁶ Note that it is not necessarily the case. Minimizing fiscal avoidance contributes also to the reduction of the concentration of wealth in some dynasties, and, as such, can be defended on the ground of equalizing opportunities across families (see Halliday 2018).

Ageing without Ageism?

257

his grandchild, unlike John, who transmits his wealth to his son. Based on Vickrey’s argument, one should tax the bequest left by Jim at a higher rate because of a larger age gap. However, it is not clear why Jim should be penalized with respect to John. If the government cares equally about their interests, it is fair to tax the two bequests at the same rate. Moreover, Jim’s transfer would minimize inequalities among descendants by compensating the unlucky grandchild. Vickrey’s tax proposal is here in conflict with basic intuitions about fairness.

3. Utilitarianism, age at death, and accidental bequests Consider now inheritance taxation in an economy where there is inequality of longevity: some individuals die prematurely, while others enjoy a long life.⁷ Assuming the absence of annuity markets, the possibility of premature death gives rise to what can be called ‘accidental bequests’, that is, bequests that would not have been transmitted provided a premature death had not occurred.⁸ In the case of a long life, the inheritance includes only an unconditional component, which the person had planned to give to their descendants, whatever the time of their future death. However, for the prematurely dead, bequests include two components: an unconditional component, which is independent of the age at which the person dies, and an accidental component, which depends on the age at which the person dies.⁹ The presence of the accidental component is explained as follows: in the case of a premature death, the inheritance includes, in addition to the unconditional component of bequest, the deceased’s ‘lost saving’, that is, the resources that the deceased had saved for their old age and that they would have consumed provided their early death did not take place. A utilitarian government, whose goal is to maximize the sum of individual utilities in the population, leads to the equalization of the marginal utility of everyone, which has strong implications for the taxation of bequests. Let us assume that an individual’s joy-of-giving function is increasing and concave.1⁰ Utilitarianism prevents some individuals from giving more than others.11 This has an unambiguous implication ⁷ This section relies on the study of optimal inheritance taxation in Fleurbaey et al (2022). By emphasizing heterogeneity in longevity, that paper differs from Cremer et al. (2012); Fahri and Werning (2013); and Piketty and Saez (2013), which focus on heterogeneity in wages and parental altruism. ⁸ Otherwise, if there existed an annuity market, the savings of the dead would be redistributed among the surviving old of the same cohort so that there would be no accidental bequests going to descendants. ⁹ Note that, since individuals anticipate what is given to their children under different longevity outcomes, the donator has the intention to give the unconditional component in case of a long life and both the unconditional and the accidental components in case of a short life. Thus, the ‘accidental’ component of the bequests is not unintentional. We use the term ‘accidental’ to highlight that this component of bequests would not have been transmitted provided a premature death had not taken place, unlike the unconditional component. 1⁰ The joy-of-giving function is the component of the individual’s utility function that captures the utility derived by the person from the act of giving some resources to a close relative. We assume an individual’s total utility to be additive in its consumption component and in its joy-of-giving component. 11 The intuition goes as follows. Utilitarianism equalizes the marginal utility of giving for all donators. If a person were to give more to their descendant, the associated marginal utility of giving would be, for that donator, smaller, which contradicts utilitarian optimality. At the utilitarian optimum, all marginal utilities of giving are equalized, implying that all individuals give the same amount, whether these are short-lived or long-lived.

258

An Age-Differentiated Tax on Bequests

for bequest taxation: accidental bequests should be taxed at a 100 per cent rate and redistributed equally among all young adults. The intuition behind that result is that, given the concavity of the joy-of-giving function, social welfare is maximized when all individuals who die give the same amount. From that perspective, one should confiscate all accidental bequests, which are given only by the prematurely dead, but leave unconditional bequests untaxed. However, confiscating accidental bequests can be criticized on the grounds of penalizing the short-lived, who are, under general conditions, the most disadvantaged. Prematurely dead individuals are disadvantaged since they have, due to their shorter life, fewer opportunities to transform resources into well-being. Under general conditions, short-lived individuals have a lower lifetime well-being in comparison to the long-lived. A confiscatory tax on accidental bequests, by preventing the prematurely dead from giving more to their descendants (in comparison to long-lived persons), worsens even more the situation of the most disadvantaged individuals. This definitely goes against the intuition of justice. Of course, in the real economy, it is difficult for a government to impose distinct (explicit) tax rates on the different components of bequests—accidental and unconditional. We are actually in a second-best world, where the government has access to only a limited number of fiscal instruments. However, since the age of the deceased is observable, the government can use the age of the deceased as a proxy variable: the accidental component of bequest is larger when a person dies early and vice versa. It is thus possible for a government to achieve a differentiated taxation of the two components of bequests by differentiating the tax rate on total bequests based on the age of the deceased. When solving the optimal (second-best) problem faced by the utilitarian government, Fleurbaey et al (2022) show that it is optimal to tax total bequests at a rate that decreases with the age of the deceased. The argument goes as follows. We know from above that a utilitarian government would like, ideally, to tax accidental bequests at a 100 per cent rate. This cannot be done directly in a second-best world where only total bequest can be taxed, but it is nonetheless possible, by taxing bequests at a rate that decreases with the age of the deceased, to tax more the bequest left by those who die early and for whom the accidental component of the bequest is the largest. Age-differentiated bequest taxation is thus only an indirect way to tax accidental and unconditional bequests differently, in the absence of tax instruments for the distinct components of bequests. Quite interestingly, the utilitarian argument considered here leads to justifying age-differentiated taxation of bequests but in a direction opposite to the one considered in section 2, which is Vickrey’s argument. The reason why the two arguments differ in conclusion lies in the fact that Vickrey considers an economy where individuals can follow strategies to avoid paying inheritance tax, strategies that are not available in this model. Another fundamental difference is that we consider here inequalities in the duration of life, giving rise to accidental bequests, something that is not considered in Vickrey (1945 [1994]).

Ageing without Ageism?

259

Although appealing at first glance, the second-best utilitarian argument for a tax on bequests that is decreasing with the age of the deceased can be criticized on the same grounds as the 100 per cent tax rate on accidental bequests in the first-best setting: such an age-differentiated tax on bequests penalizes the short-lived with respect to the long-lived, which is counter-intuitive. Given that prematurely dead persons are disadvantaged with respect to the long-lived, imposing a higher tax rate on bequests left by the former amounts to imposing on them a kind of ‘double penalty’. Fairness would recommend allowing the disadvantaged to transmit more—and not less—to their descendants. A bequest tax that is decreasing with the age of the deceased has implications contrary to our basic intuitions about social justice.

4. Compensating the unluckily short-lived The ‘double penalty’ imposed to the short-lived under utilitarianism is particularly counter-intuitive since it goes against the intuition of compensating the unluckily short-lived. Given that a large part of longevity inequalities are due to circumstances, there is strong ethical support for compensating the unluckily short-lived. This support for compensating the short-lived takes its roots in the Principle of Compensation (Fleurbaey and Maniquet 2004; Fleurbaey 2008). This principle states that, when well-being inequalities are due to circumstances over which individuals have no control, the government should intervene to abolish those inequalities.12 The utilitarian criterion goes against the Principle of Compensation since it leads to penalizing short-lived persons even more. This counter-intuitive feature of utilitarianism motivated the search for alternative social welfare criteria, which do more justice to the compensation of the short-lived. One of those criteria is the ex post egalitarian criterion proposed by Fleurbaey et al. (2014), which gives priority to the worst-off individual defined in ex post terms. Under unequal lifetimes, that criterion aims to maximize the realized lifetime well-being of prematurely dead individuals, who are usually the worst off ex post. Ex ante (i.e. before individual longevities are revealed), no one—neither individuals, nor the state—knows who will end up being short-lived or long-lived. But, using life tables, the state can anticipate that some individuals will turn out to die prematurely and design policies in such a way as to favour the interests of individuals who will be short-lived, even if those persons cannot be identified ex ante. When applied to the issue of bequest taxation, the ex post egalitarian criterion leads to conclusions that are opposite to the ones derived under utilitarianism (Fleurbaey et al. 2022). Provided that individuals care about how their lost savings are 12 While Dworkin (2000) highlighted the distinction between circumstances and choices, hypothetical insurance markets à la Dworkin lead generally to allocations violating the Principle of Compensation. Indeed, hypothetical insurance devices only lead to allocations that are fair ex ante, whereas the Principle of Compensation is about fairness ex post, that is, once the states of the world are known. See Fleurbaey (2008: 172–173).

260

An Age-Differentiated Tax on Bequests

redistributed in case of premature death, accidental bequests can be regarded as allowing for the compensation of the unlucky short-lived. Short-lived persons cannot, by definition, consume what they saved for their old days, but if they prefer those lost savings to be transmitted to their children, then, allowing them to transmit these to their children contributes to making them better off. Thus, accidental bequests can be regarded as an indirect way to compensate unlucky short-lived persons.13 As a consequence, if the goal of the state is to compensate the prematurely dead, accidental bequests should neither be confiscated nor taxed. If a government could impose different tax rates on the different components of bequests (accidental and unconditional), then it would not tax but subsidize accidental bequests. That result, shown by Fleurbaey et al. (2022), goes against the confiscatory tax that would prevail under utilitarianism (see section 3). The logic goes as follows. If the state aims to compensate prematurely dead individuals, this entails subsidizing the accidental component of bequests since this component is transmitted only by prematurely dead persons and not by long-lived individuals. Subsidizing accidental bequests amounts to allowing the prematurely dead individual to transmit more resources to their descendants and thus contributes to the compensation for premature death. In real-world economies, it is not possible to impose distinct (explicit) tax rates on the different components of bequests, that is, accidental and unconditional. However, it is possible, as in section 3, to use the age of the deceased as a proxy variable since it is reasonable to assume that the share of the accidental bequest in the total bequest is likely to decrease with the age of the deceased. Hence, since the accidental component of bequests becomes less and less sizeable with the age of the deceased, one way to compensate the short-lived is to tax bequests less when the deceased is younger and more when they are older. There is thus a second-best ex post egalitarian argument for taxing bequests at a rate increasing with the age of the deceased. That conclusion is the exact opposite of what prevailed under utilitarianism, which recommends taxing bequests at a rate that decreases with the age of the deceased. The intuition is that utilitarianism leads to the equalization of the marginal utilities of alternative uses of resources. This leads to preventing some individuals from giving more than others. In contrast, the ex post egalitarian criterion aims to equalize the levels of utility between short-lived and long-lived individuals. From that perspective, it is optimal to allow prematurely dead individuals to give more, net of tax, to their descendants. This motivates taxing bequests at a rate that increases with the age of the deceased. Although attractive, that—second-best—argument supporting a tax on bequests that increases with the age of the deceased nonetheless faces one major limitation. 13 One may wonder at which precise time the compensation for premature death takes place, in the same way as some philosophers, such as Bradley (2009), try to localize the harm due to a premature death. The answer is that the compensation takes place over the entire life of the prematurely dead and is thus not located at a particular point within that life. More technically, Fleurbaey et al (2022) measure ex post lifetime well-being as a numerical representation of preferences over degenerate lotteries of life. Subsidizing accidental bequests contributes to increase ex post lifetime well-being of the prematurely dead when measured in that manner.

Ageing without Ageism?

261

The argument states that allowing prematurely dead individuals to give more (net of taxes) is a way of providing some compensation for the harm of having a shorter life. But this is true only if individuals care about what they give to others in case of premature death.1⁴ If, alternatively, individuals did not care about how lost savings are distributed in case of premature death, then this argument would not hold anymore. Age-differentiated taxation of bequest would be ineffective at compensating such short-lived persons and the only avenue left would need to rely on modifying consumption/leisure profiles.1⁵

5. Compensating unluckily deprived widows and orphans There exists another equity argument that also supports age-differentiated taxation of bequests. That argument does not rely on the idea of compensating a person for their own premature death. It considers, instead, the compensation of widows and orphans who are the collateral victims of the premature death of a person. This argument is very different from the previous one since it involves a demand for justice not between the deceased but among the surviving persons. Age-differentiated taxation of bequests is regarded here as a way to reduce inequalities when some persons face the premature death of a close relative. In order to do justice to the idea of compensating orphans/widows, let us adopt, as in section 4, the ex post egalitarian criterion, which gives priority to the worst off ex post. In the context under study, this social criterion supports a taxation of bequests that increases with the age of the deceased. Several variants of the argument can be proposed. The first variant of that argument is based on the simple idea that families are economic and social units in which some members benefit from resources provided by other family members. In that context, the death of a parent has a strong impact on the resources enjoyed by the rest of the family. Quite importantly, the impact of the premature death of the parent on the resources of the children (and, hence, on their well-being and their development) is likely to vary with the age of the children and, hence, with the age of the parent, assuming a constant intergenic interval. If the parent dies at a young age, this implies, in purely material terms, a bigger loss for (younger) children in the sense that this penalizes their future development more strongly.1⁶ In light of this, it appears intuitive, from an ex post egalitarian perspective applied to the families of the deceased, to tax bequests at a lower rate when the deceased was younger and at a higher rate when the deceased was older.1⁷ 1⁴ The reason why they care about giving to their descendants is irrelevant for that argument: as long as it matters for the deceased to give, allowing them to give more in case of a premature death brings some compensation for an early death. 1⁵ See Fleurbaey et al (2014); Leroux and Ponthiere (2018). 1⁶ This argument is, like the previous one, a second-best argument, that is, under a limited set of policy instruments. 1⁷ We assume here implicitly that the degree of material dependence between family members is decreasing with the age of the deceased.

262

An Age-Differentiated Tax on Bequests

There is another variant of that argument, which considers the emotional impact of a death, that is, its impact in well-being terms, independently from material resources. As shown by Blanchflower and Oswald (2004), becoming a widow(er) has a major negative impact on subjective well-being.1⁸ Note that, if we adopt a life-cycle perspective, the well-being loss due to the death of someone is likely to be larger when the deceased is younger. The underlying intuition is that, in case of interest for joint survival, individuals are like ‘durable goods’, which are enjoyed during longer time periods in case of a late death and during shorter time periods in case of an early death.1⁹ Thus, in terms of compensation, even if all persons in widowhood face a strong deprivation, this deprivation is likely to be quantitatively larger when the deceased is younger because the opportunity cost of the non-lived coexistence time is then larger. In light of this, an ex post egalitarian view applied to the surviving family members would, again, recommend taxing bequests at rates that increase with the age of the deceased, as under the first variant. While appealing at first glance, that argument faces two main criticisms.2⁰ First, the argument supports, strictly speaking, a taxation of bequests not based on the age of the deceased but on the age of the descendants of the deceased. True, in many cases, the younger the deceased is, the younger their descendants are. However, in case of differential fertility timing, this would not be true any longer, and this argument would not support a differentiation of bequest taxation based on the age of the deceased. A second, related point, concerns the tension between compensating the deceased and compensating their descendants. While those arguments seem to go in the same direction, conflicts may arise in case of differences in parent–child age gap. Take two individuals: the first one, who had a son at age 20, died at age 40, whereas the second one, who had a son at age 40, died at age 50. The compensation for a shorter life would lead to imposing a lower tax rate on the bequest left by the first individual, whereas the compensation for losing a parent would lead to imposing a higher tax rate on the bequest left by the first individual. Thus, the two compensation motives can lead to contradictory policy recommendations.

6. Discussion Up to now, this chapter has examined four distinct arguments supporting an agedifferentiated taxation of bequests. Given that those arguments can conflict regarding the direction and extent of age-differentiation of bequest taxation, it is necessary to compare and weight these to determine which one is the most persuasive. 1⁸ Blanchflower and Oswald (2004) show that $100,000 per year would be required to compensate someone for the death of a spouse. 1⁹ This argument may not hold for all ages of life. For instance, Broome (2017) discusses some reasons why the death of a young adult may be seen as worse than the death of a child. 2⁰ These apply to both variants of the argument.

Ageing without Ageism?

263

Regarding the direction of discrimination with age, among the four arguments considered, only one—the utilitarian argument—recommended a tax that decreases with the age of the deceased. But, as we have seen, that argument is quite questionable. Such a direction for age-differentiated inheritance taxation leads to penalizing the prematurely dead, which is counter-intuitive since the prematurely dead is the main victim of a harm for which they are not responsible, and, as such, they should be compensated rather than penalized.21 The utilitarian argument for age-differentiated bequest taxation being questionable, we will leave it aside in the rest of this section and focus on the three other arguments. Among the three justifications for an inheritance tax increasing with the age of the deceased, we believe that the most persuasive argument is the one based on the compensation of the prematurely dead. Our rationale is based on (i) the hierarchy of priorities among goals and (ii) the (un)availability of alternative policy instruments allowing for achievement of those goals. The three arguments under comparison rely on distinct goals: raising fiscal revenues, compensating the prematurely dead, and compensating the heavily dependent in case of death of a relative. We believe that, among those goals, the second and the third one should have priority over the first one. The reason is that the harm due to a premature death—either for the deceased themselves or for their family—is such a source of misfortune that the compensation for those losses should have priority over the goal of raising fiscal revenues per se. It is more difficult to rank the two compensation goals: compensating the deceased themselves or their surviving family. Both the deceased and their surviving family are victims of a serious harm, for which these are not responsible.22 In both cases, the Principle of Compensation recommends a state intervention aimed at compensating those victims. One cannot state that some compensation should have priority over another one. Both harms matter, and the associated demands of justice are equally valuable. In order to weigh our different compensation-based arguments supporting an inheritance tax increasing in the age of the deceased, we thus need to consider the second aspect of the comparison: the possible substitution by means of other policy instruments. From that perspective, there exists a major difference between the two arguments. The relevance of age-differentiated bequest taxation for the compensation of the prematurely dead is reinforced by the presence of informational constraints specific to the compensation of the prematurely dead and which make that policy instrument hardly substitutable for that particular purpose. The informational constraint is the following: short-lived persons can hardly be identified before durations of life are

21 Note that, in the case of a suicide, the issue of the responsibility of the deceased for their premature death could be raised. This goes outside the scope of this chapter, which is about ordinary premature deaths. 22 Again, we focus here on ordinary premature deaths and exclude suicides and murders.

264

An Age-Differentiated Tax on Bequests

revealed, and there is little that can be done for them after the short-lived have been identified since they are dead. In light of those serious limitations, the only possible strategy to improve the situation of the prematurely dead is to design policies that favour all young individuals since the young necessarily include the unlucky individuals who will turn out to die prematurely. True, this can be done by modifying consumption profiles, as in Fleurbaey et al. (2014), or leisure profiles, as in Leroux and Ponthiere (2018). But since the capacity of the prematurely dead to convert resources into well-being is de facto limited by their short life, allowing them to give more to their loved ones is a major channel through which we can improve their lives. Thus, excluding age-differentiated bequest taxation would remove an important policy instrument from the perspective of compensating the prematurely dead. Therefore, under those informational constraints, age-differentiated bequest taxation is a non-substitutable way of improving the situation of the unlucky individuals who turn out to be short-lived. But the same is not true when considering the compensation of orphans/widows. Unlike the prematurely dead, those individuals can be identified when alive and can receive some compensation once the harm has occurred. It is thus possible to help them without using age-differentiated bequest taxation. For instance, if orphans are suffering from lack of funding for higher education, special student scholarships can be introduced to correct for unequal endowments. Similarly, if orphans are suffering from the absence of their parent, some particular services/allowances could be introduced to help them. But none of those instruments could be used for the compensation of the short-lived themselves. In sum, informational constraints are weaker for the compensation of orphans/widows than for the compensation of the prematurely dead, who can only be identified when dead. Differences in informational constraints make age-differentiated bequest taxation hardly substitutable for the compensation of the prematurely dead but easily substitutable for the compensation of orphans/widows. As a consequence, age-differentiated inheritance taxation is best supported by the argument based on the compensation of the prematurely dead.

7. Conclusion This chapter considered whether it is justified to tax bequest differently based on the age of the deceased. We examined four arguments supporting age-differentiated taxation of bequests: three arguments supporting a tax on inheritance that increases with the age of the deceased and one argument supporting a tax that decreases with the age of the deceased. We argued that, from an ethical perspective, there is strong support for a tax that increases with the age of the deceased, whatever we consider the compensation of

Ageing without Ageism?

265

the short-lived themselves or the compensation of the family in widowhood. Among those two arguments, we argued also that the former is more persuasive because age-differentiated bequest taxation is, from the perspective of achieving compensation of the prematurely dead, a non-substitutable instrument (due to stronger informational constraints), unlike the situation that prevails for the compensation of the family of the short-lived.

Acknowledgements The authors would like to thank Greg Bognar, Guido Erreygers, Eric Fabri, Joset Ferret-Mas, Axel Gosseries, Daniel Halliday, Danielle Schwartz, Shlomi Segall, and Manuel Valente, as well as participants of the Fair Inheritance Workshop No. 1 for their comments and suggestions.

References Blanchflower, David, and Andrew Oswald. 2004. ‘Well-Being over Time in Britain and the USA’. Journal of Public Economics 88: 1359–1386. Bradley, Ben 2009. Well-Being and Death. New-York: Oxford University Press. Broome, John 2017. ‘The Badness of Early Death’. In Saving Lives from the Badness of Death, ed. Espen Gamlund and Carl Tollef Solberg. New-York: Oxford University Press, 105–115. Cremer, Helmuth, Firouz Gahvari, and Pierre Pestieau. 2012. ‘Accidental Bequests: A Curse for the Rich and a Boon for the Poor’. Scandinavian Journal of Economics 114: 1437–1459. Dworkin, Ronald 2000. Sovereign Virtue. The Theory and Practice of Equality. New York: Harvard University Press. Fahri, Emmanuel, and Ian Werning, 2013. ‘Estate Taxation with Altruism Heterogeneity’. American Economic Review 103: 489–495. Fleurbaey, Marc 2008. Fairness, Responsibility and Welfare. New York: Oxford University Press. Fleurbaey, Marc, and François Maniquet. 2004. ‘Compensation and Responsibility’. In Handbook of Social Choice and Welfare, vol. 2, ed. Kenneth Arrow, Amartya Sen, and Kotaro Suzumura. Amsterdam: North-Holland. Fleurbaey, Marc, Marie-Louise Leroux, and Gregory Ponthiere. 2014. ‘Compensating the Dead’. Journal of Mathematical Economics 51: 28–41. Fleurbaey, Marc, Marie-Louise Leroux, Pierre Pestieau, Gregeory Ponthiere, and Stephane Zuber. 2022. ‘Premature Deaths, Accidental Bequests and Fairness’. Scandinavian Journal of Economics 124: 709-743.

266

An Age-Differentiated Tax on Bequests

Halliday, Daniel 2018. The Inheritance of Wealth: Justice, Equality and the Right to Bequeath. New York: Oxford University Press. Leroux, Marie-Louise, and Gregory Ponthiere. 2018. ‘Working Time Regulation, Unequal Lifetimes and Fairness’. Social Choice and Welfare 51: 437–464. Piketty, Thomas, and Emmanuel Saez. 2013. ‘A Theory of Optimal Inheritance Taxation’. Econometrica 81: 1851–1886. Rawls, John 1971. A Theory of Justice. New York: Harvard University Press. Vickrey, William 1945 [1994]. ‘An Integrated Successions Tax’. In Public Economics. Selected Papers by William Vickrey, ed. Richard Arnott, Kenneth Arrow, Anthony Atkinson, and Jacques Drèze. Cambridge: Cambridge University Press, 368–374.

Index ability 5, 7, 9, 15, 29–30, 33–4, 46, 117, 120, 133, 136, 186, 190 nn.13, 119, 215–21, 223–4 ableism 3, 29, 31, 33–4, 38 actuarial fairness 200 nn.7 age discrimination 1–3, 13–14, 18–23 see also discrimination -friendly city 7–8, 214–16, 219, 224–6 -group inequality 17, 75–6, 230–1, 246 limits 82, 86–92, 115, 118–9, 124, 159, 164, 166, 173, 176, 178, 181, 195 as a proxy 7, 15–16, 21, 25, 44, 91, 120, 199, 202, 211, 256, 258, 260 is special 1–2, 13–19, 21–5, 82, 85, 87–8, 91–2 see also specialness of age ageism 1, 3, 6, 22–4, 31, 34, 159, 172, 178, 225, 249 forms of 19–20 Alzheimer´s disease 34, 37, 44 nn.4, 51, 71–2, 74, 79 analogical arguments 129–31 Arneson, R. 18 nn.9, 21, 83 nn.3 Atkinson, A. 64 nn.18, 242 nn.1 autonomy bodily 172, 174 capacity for 78–80 in decision-making 7, 190 personal 46, 48 nn.8, 180–1, 214–26 reduction 42, 46, 179, 190 spatial 7, 214–26 threshold of 191–3 youth 216, 224–5 baby boomers 229–30, 232–4, 239 badness of disability 3, 28–38 old age 30–5, 36 nn.11 suffering 72, 77–80 basic income 251 nn.32 bequests 8, 243 nn.6, 254–64 see also inheritance accidental 257–60 unconditional 257–60 bias cognitive 143, 190 in conceptions of the good life 73 see also neutrality in estimation 134 nn.18, 205 human 45 implicit 20–4 mid-life 35

Bidadanure, J. 1 nn.1, 15 nn.4, 16–17, 22 nn.13, 75–6, 85 nn.8, 88 nn.10, 144, 230 nn.2 birth cohort 8, 14, 18, 24, 88, 96–7, 148–50, 223–4, 229–35, 239, 257 nn.8 Bognar, G. 6, 15, 17 nn.6, 18, 35, 37, 49 nn.11, 159, 161 nn.2, 166 nn.8 Brexit 128, 131 Broome, J. 83 nn.5, 262 nn.19 care child- 7, 41, 117, 175, 179 elderly- 8, 41–2, 71–2, 74–5, 79, 149, 225, 237–8 -freeness 5, 117–18, 122–3 health see also health care homes 44, 51 work 233, 238 childhood 3, 5, 35, 43, 46, 94, 101–4, 107, 124–5, 166, 201 as mere deficiency 5, 115–21 as a special good 5, 115–18, 121–3 citizen´s assembly 150–3 coercion 42–3, 190 nn.14, 192 Cohen, GA. 83 nn.2 compensation interpersonal 58 intrapersonal 58, 72, 73, 77, 84 post mortem 247 premature death 9, 90, 247–8, 254–5, 259–65 responsibility-sensitive 83, 254, 261–2 widows/orphans 261–5 competence age-based 3, 9, 15, 41–51, 116, 118–20, 122–3, 189–92 -based reasons 41–7, 50–1, 118–20, 138 experience-based 15 political 9, 120, 123, 135, 138 threshold of 41–51, 116, 118–20, 123 complete-life equality see entire-lives equality neutrality 17, 24–5, 90 nn.14, 246 nn.19 view see lifetime view compulsory driving exams 88 education 3, 7, 41, 90, 184–5, 188–9, 194–5 military service 88 retirement 14, 18, 25, 89–90, 128 see also retirement

268

Index

consent 98–9, 184 constitution 146 nn.5, 151 contractualism 72 nn.3 cost-benefit analysis 4, 53–5, 58, 60–1, 63–4, 66–8 COVID-19 6, 53–4, 66, 91, 145, 159, 163–9 see also pandemic cumulative income 8, 242–51 see also income currency of justice 83, 85 resources 72 nn.4, 83 well-being 58, 72 nn.4, 83 years of life 168 DALYs 15 see also QALYs Daniels, N. 1 nn.1, 17 nn.6, 73, 94, 96, 243 nn.5, 246 nn.17, 247 nn.22 & nn.24 democracy 1, 105, 146 nn.5, 194, 256 and disenfranchisement 128, 136–9 see disenfranchisement and voting non-instrumental value of 119–20 and vote-buying 129–30 see vote-buying demography 7, 99–100, 106, 153, 199–202, 205–6, 208–11 dignity 72, 75–6, 78–80, 88 nn.10, 215 disability 1–3, 7, 9, 15, 28–39, 82, 119–20, 122, 163–5, 175, 201–2, 211 the bad-difference view of 28–39 the mere-difference view of 29–39 discount rates 147–50 discrimination 9, 13–25, 30, 88 nn.10, 159, 163–5, 168, 178, 234–6, 263 age 1–3, 13–14, 18–23 see also age class 234–5 direct/indirect 19–20, 165 disability 2–3 see also disability ethnicity-race 2, 13–14, 17–18 see also race and ethnicity sex-gender 2, 13–14, 17–18 see also sex and gender statistical 19–21 wrongful see wrongful discrimination disenfranchisement old-age 5, 128, 131, 135, 139, 144 youth 5, 87, 90, 115, 118–19, 121–2 distributive justice 3–4, 72 nn.4, 186, 217 nn.4, 233 domination 4, 75, 88 nn.10, 117, 231 duty age-based 43, 89, 124–5 of assistance 79 intergenerational 5, 83, 116, 123 of paternalism 43 Dworkin, G. 42, 45 Dworkin, R. 72 nn.4, 124 nn.5, 259 nn.12 education 1, 5, 16, 82, 180–1, 264 categories 200–2, 205 nn.17, 206–11 compulsory 3, 7, 41, 90, 184–6, 193–6 early enrollment in 121 nn.4

efficiency in 186–9 and employment 17, 20, 23, 200 paternalism in 189–193 see also paternalism resource account 7, 184–6, 188, 193–5 systems 15–17, 47, 176, 184–9 value of 46–7, 119 efficacy 99, 137–8 efficiency 16–17, 89–91, 96 cognitive 221 economic 137 nn.24 educational 7, 90, 186–9, 195 lifespan 17 nn.6, 71–2, 74, 78 means-end see means-end efficiency political 144–5 as a sequence 17, 23–5, 90 see also sequence efficiency urban 214, 224 egalitarianism 3, 4, 71–2, 82–3, 96, 98, 180, 191 age-group justice 73–5, 79 nn.14, 80 complete-lives see entire-lives equality ex ante/ex post 200–1, 208, 210–1, 259–62 maximin/leximin 8, 72, 83, 242–6, 248–9, 250–1 political 146, 191 relational 4, 17 nn.7, 72, 75–7, 79–80 responsibility sensitive 83–4 snapshot 85 see also time-specific election 44 nn.4, 87, 128, 137, 151–2 eligibility adoption 86 criteria 163–4, 166–9 disability benefits 211 ICU treatment 159, 166, 168 pension benefits 7, 88, 211 voting 5, 138, 153 entire-lives equality 4, 9, 18, 58, 60, 73–5, 82–92, 94, 96, 120, 165–6, 168, 224, 230, 242–3, 246, 248, 250, 260 nn.13 entitlement 84, 86, 100, 125, 139, 184–5, 195, 238, 242 nn.3, 251 environment modification 219, 223, 226 natural 149 school 188 social 3, 7, 28–39, 187, 214–16, 220 socioeconomic 88 urban 214–26 work 188 equality capabilities 83 opportunity 13, 83, 87, 145, 165–6, 256 nn.6, 258 outcome 99, 200 over complete lives see entire-lives equality primary goods 83, 246 nn.17 equity 162–3, 261 ethnicity 137, 164 see also race

Index exclusion 120, 165 exploitation 88 nn.10, 181 fair innings 18, 65 nn.19, 168 Feinberg, J. 41–4, 48, 50 feminism 172 fertility 6, 9, 49, 100, 172–3, 177–81, 199, 262 first-come, first-served 166–7 Fleurbaey, M. 84 nn.7, 89 nn.13, 200, 201 nn.9, 208, 247 nn.20, 248 nn.25, 257 nn.7, 258–61, 264 freedom 5–6, 35, 43, 83, 125, 233 ability-based 5 choice 46, 189–90 see also paternalism of movement 6 reproductive 7, 172–3 symbolic 134 free-riding 130 free time 117, 122–4 future generations 6, 116, 143, 146, 150–2 futures assembly 150–4 gender and age 1–2, 9, 14–18, 23, 82, 86–7, 91 discrimination 13, 163, 172, 201 division of care work 238 equality 151 and health 200–1, 205 nn.17, 210–11 gentrification 231 González-Ricoy, I. 136–7, 143, 146 Goodin, R. 47 nn.6, 128 nn.1, 138, 151 Gosseries, A. 4, 15–18, 58, 75–6, 82, 83 nn.4, 88–91, 125 nn.6, 128 nn.4, 131 nn.14, 137, 139 nn.25, 143, 146, 165 nn.7, 200, 224, 243 nn.9, 246 nn.19, 248 nn.27 Gross Domestic Product (GDP) 145, 201 nn.11, 206, 210 Halliday, D. 8, 89 nn.12, 229, 234, 238–9, 256 nn.6 harm account of discrimination 24–5 in driving 129, 131–5 expressive 119–20 intrapersonal distribution of 73, 77 to oneself 42–3, 47, 50 in paternalism 42, 45, 50 in premature death 260 nn.13, 261, 263–4 in short-termism 144, 146 in stereotypes 33–4 in voting 135–8 health care 1–2, 5–6, 71, 72 nn.4, 82, 89, 97, 125, 145, 174, 179, 225, 229, 233 emergency 159–60, 164, 166, 168–9 expected 201–2, 205 nn.17, 210 function of 147 index 203–11 inequality 199–211

269

infant 174–5 lifestyle 30 old age 74, 134 resources 15 risks 174 home ownership 8, 224–5, 229 nn.1, 230–3, 235–9 housing intergenerational 8, 215, 224–6 public 215, 222 stock 235, 238 wealth tax 8, 229–40 hypothetical impartiality 22 see also impartiality hypothetical insurance 4, 72 nn.4, 259 nn.12 impartiality 74, 76 hypothetical 22 incentives 6, 130, 139, 143–4, 151–2, 154, 188, 192, 194, 232, 235–6, 237 nn.11, 238–9 income basic 251 nn.32 replacement 4–5, 95, 97–103, 105–9, 199 volatility 243–4 indexation 199 nn.1, 200 index of ill-health 203, 205, 208–9 primary goods 246 nn.17 inequality age-group 17, 31, 75–6, 230–1, 246 birth-cohort 8, 223, 229–34 complete-lives 17, 84–5, 165 endowments 264 health 7, 211 longevity- 7, 199, 210, 248, 257–9 snapshot 84–5, 88 nn.10 see also time-specific socio-economic 7, 88, 201–2, 208, 211, 223–4, 226, 233–5, 239–40 inflation 7, 185, 235 inheritance 5, 8–9, 123, 175, 237–9, 254–8, 263–4 insurance biological 180 hypothetical see hypothetical insurance sickness see sickness insurance social see social insurance intensive care unit (ICU) 86, 89, 159, 162–4, 166–8 intergenerational bargaining 152 nn.9 housing 8, 215, 224–6 inequality 8, 229 justice 5, 123–4, 128 nn.2, 137, 147, 239 transfers 102 nn.7 isolationism 84 legitimacy 116, 120, 165, 194, 209 challenge of 145–6 long-term 94 perceived 96–9, 153 leisure 55, 246 nn.17, 261, 264

270

Index

leximin 72, 83, 87 see also egalitarianism liberal egalitarianism 191, 218 nn.5 liberty 41–2, 46, 116 life expectancy and age 6, 15, 47–50, 62, 64–5, 90–1, 150, 154, 159, 160–6, 168–9 as a criteria 6, 159–66, 168–9 differential 88, 148, 200, 208 nn.20 healthy 199 and income 59, 64 quality-adjusted see QALYs rising 199–200, 229 life satisfaction 95, 103–5, 107–9 life-saving treatment 18 lifetime income 65, 242–6, 248–51 view 58, 73 nn.6, 224, 243, 246, 251 see also whole lives and entire-lives egalitarianism well-being 58–62, 64–6, 68, 72, 74, 117, 122, 258–9, 260 nn.13 Lippert-Rasmussen, K. 3, 13 nn.1, 16, 19, 24, 28, 75 nn.10, 88 nn.10, 89 nn.12 longevity -adjustment 87, 236–7, 239 differential 7, 9, 85, 89–90, 91, 199, 210–1, 232, 246–7, 254, 257, 259 ex ante/ex post 199–202, 205 nn.17, 210, 259 -income bundles 60 long-termism 6, 9 marginalization 4, 75–7 market annuity 257 housing 230 nn.4, 231 hypothetical insurance 259 nn.12 see also insurance labour 2, 17, 21, 23, 97, 193, 233–4, 236, 238, 245 nn.13 -oriented welfare states 97 private 222–3 stock 84 voting 129–30 see also vote-buying McKerlie, D. 1 nn.1, 18, 71 nn.2, 73–6, 83 nn.5, 85 nn.8, 94, 120 means-end efficiency 15–16, 18–19, 21, 25 rationality 217 means-test 97–9, 105 mere-difference view 3, 29–39 Mill, JS. 41–2, 44–6, 48 nn.8, 190 millennials 229–30, 232–3, 239 mobility physical 30, 33, 134, 194, 204 social 251 monetary equivalent 54–5, 58, 60, 63–4, 67–8 monogamy 236–7

mortality and retirement age 202 risk 61, 168 and sex-gender 208 nn.20 motherhood 7, 172, 174–7, 179 Nagel, T. 73, 77 nationality 164 neutrality complete-life 17, 24–5, 90 nn.14, 246 nn.19 state 218 nn.5 non-domination 4, 88 nn.10, 117 non-ideal theory 123–4, 186 non-identity problem 176 nn.2 non-interference 216, 218–19, 222 obligations 77, 84, 86, 131, 143 fiscal 8, 242–6, 249–251 OECD 8, 32, 102–4, 200, 205 nn.16, 223–4 old-age dependency 200 income 248 pension benefits 100, 103, 211, 247 nn.23, 251 poverty 8, 104, 247–8 risks 100, 102 nn.6, 103 support 1–2, 5, 100, 103, 211 opportunity costs 119, 124, 238, 262 oppression 28 nn.2, 29, 179 pandemic 6, 66 nn.21, 86, 143, 145, 159, 163–4, 167–9 see also COVID-19 paradox of redistribution 98, 105 parental leave 95, 97, 100, 236, 237 nn.11 parenthood 95, 178–9, 256 Parfit, D. 49 nn.9, 56, 72 nn.3, 176 nn.2 passage of time 15, 17, 82 paternalism 2–3, 7, 9, 41–51, 189–95 best-judge view 44–7 development view 44, 46–7, 50 in education 3, 7, 189–195 hard 41–3, 45–6, 48 nn.8, 49, 51 soft 41–6, 49–51 pension see also retirement adequacy 99 eligibility 7, 88, 211 old-age 71, 95, 97, 100, 103, 211, 247 nn.23, 251 see also old age as pay-as-you-go 200 reforms 210 Ponthiere, G. 8–9, 200, 210, 230 nn.2, 243 nn.6, 248 nn.28, 251 nn.32, 254, 261 nn.15, 264 population ageing 1–2, 98, 100, 128 nn.2, 137, 151, 199, 229 positive-sum solutions 95–6, 102, 106 poverty 5, 95, 98, 103–5, 107–9, 193, 250, 256 in old age 8, 104, 247–8

Index pregnancy 6, 172–5, 177–9 premature death 8, 9, 33, 64, 85, 88–9, 120, 200 nn4, 244, 246–50, 254–5, 257–65 prioritarianism 4, 53–4, 56–61, 64–8, 72 nn.3, 74–5 procreation 179, 181 productivity 15, 19, 89 prudence 41–2, 46–8, 51, 72, 76, 80, 83, 96, 115, 117, 122–3, 150, 190 prudential lifespan account 72–4, 76, 80, 96 QALYs 15, 161–3, 166, 169 see also DALYs quality of life 49, 60, 62, 161–2, 164, 166, 169, 175–7, 233 quantity of life 160–2, 167–9 quotas 144, 146 race 1–2, 9, 13–18, 29, 79, 82, 91, 164, 180, 193 see also ethnicity rationality 17 nn.6, 118–19, 121, 136, 139, 144, 152, 190–1, 216–17, 220, 238 rationing 6, 159–60, 163, 169 Rawls, J. 17 nn.6, 246 nn.17, 256 Raz, J. 216 nn.3, 217 redistribution 94–100, 102 nn.7, 105–6, 146, 239, 247, 257–60 the paradox of 98, 105 referenda 128, 131, 148–9, 151 relational egalitarianism 4, 17 nn.7, 72, 75–7, 79–80 religion 20–1, 163–4, 218 rent 222, 225, 231, 236, 237 nn.11 reproduction 172–81 assisted 3, 6, 172–81 respect 7, 22–3, 76–7, 79–80, 120, 139 nn.25, 151, 174, 180–1, 190–1, 215 retirement age 7, 199–211, 232, 234–5 bonus 152 early 209, 210 nn.22 income benefits 100, 242 nn.3, 251 mandatory 14, 18, 25, 89–90, 128 right to 201, 209–10 rights 46, 86, 116, 124–5, 128–33, 135, 137–9, 146 nn.5, 210 risk reduction 4, 53–4, 58–68, 145 saving lives 18, 61–2, 64–5, 161–2, 167–9 savings 236 nn.9, 257, 259–61 scarcity extreme 160, 164, 166–7 health 15, 53, 66, 96, 160, 164, 166–7, 169 jobs 235 urban planning 214, 222–3 segregation 23–4, 76, 225

271

self-determination 174 self-interest 97, 130, 137–8 sequence efficiency 90 individual 17, 24, 25 collective 17, 23, 25 sexism 16–17, 79, 164 short-termism 5–6, 45, 143–54 sickness insurance 95, 100–1 see also social insurance social freezing 3, 7, 178–81 social insurance 4–5, 94–109 income replacement 4–5, 95, 97–103, 105–9, 199 see also income solidarity 96, 153 specialness of age 1–2, 13–19, 21–5, 82, 85, 87–8, 91–2 sufficiency 44, 75–7, 118 surrogacy 144, 154 survival expectancy 147, 159 probability 59–66, 159, 164, 167 short-term/long-term 160, 164–5 targeted programmes 41, 43, 94–5, 98, 105, 209, 239 target variables 15–16, 19 taxation age-sensitive 8–9, 239–40, 242–251, 254, 256, 258–9, 261 cumulative 8, 242–51 housing wealth 8, 229–40 inheritance 8–9, 237–9, 254–8, 263–4 progressivity 231, 236, 243, 247 nn.21, 254 nn.1 revenue 245, 255–6, 263 thought experiments 34 nn.9, 73 see also veil of ignorance time preferences 143, 147, 149 time-specific (or time-slice) priority 74–7, 224 sufficiency 75–7 view 58, 224 triage 6, 159–69 unconditional bequests 257–8, 260 unemployment 18, 95, 99–101, 149, 193, 200, 246 nn.17 universalism 5, 94–5, 97–9, 105 urban planning 214–16, 223–4, 226 utilitarianism 4, 17, 53–68, 71, 145, 168, 200 nn.7, 257–60, 263 utility aggregate 71 expected 149 marginal 64, 257 sequence efficiency 17, 25

272

Index

utility (Continued) total 16–17, 25, 56, 60, 64, 168, 257 nn.10 versus egalitarianism 260 vaccination 53, 66–8, 91 Vanhuysse, P. 94, 100–1 Van Parijs, Ph. 83 nn.1, 128 nn.2, 131 nn.14, 137, 144, 243 nn.5, 246 nn.17 veil of ignorance 17 nn.6, 247–8 Vickrey, W. 243–5, 255–8 voluntariness 3, 41–6, 48–51, 185, 188 nn.9 vote-buying 129–30 voting age-adjusted weights 90–1, 144, 146, 150 age-based rights in 5–6, 14, 86, 88, 90, 115–16, 118–25, 128–31, 135–9, 144, 153, 185 nn.2, 195 behaviour 144, 148–51, 153–4 and cognitive impairment 44 nn.4 and driving see voting-driving analogy and future generations 152 longevity-adjusted 87 storable 87 voting-driving analogy 5–6, 128–39 cognition 135–6 contributory efficacy 137–8 harm 136–7

moral cost 138–9 war 88, 145, 160, 229, 232 wealth 8, 64 nn.16, 85, 145, 148, 163, 185, 216, 223–4, 229–40, 244, 247, 251, 254–7 welfare state 3–5, 71, 80, 94–100, 105–6 whole lives 4, 17–18, 31 nn.5, 35, 105, 121, 146, 165, 168, 224 see also lifetime and entire-lives equality willingness to accept 54–5, 98 to pay 54–5, 64, 173 work capacity to 7, 9, 71, 199–200 pro-work 101–5 working age 99–105, 108, 199–200, 229, 239 workplace 172, 179 World Health Organization 1 nn.3 & 4, 32, 34, 214–15, 221, 224–5 wrongful discrimination 2, 14, 250 nn.30 harm accounts of 24–5 intention-focused accounts of 21–2 social meaning accounts of 22–4 youth assemblies 6, 144, 150–4 autonomy 216, 224–5 quotas 144, 146 see also quotas