4. Autonomy-Based Reasons for Limitarianism

Danielle Zwarthoed

1. Introduction

On June 15, 2013, Ethan Couch, a Texan teenager, stole two cases of beer from a Walmart store, took the wheel of his father’s pickup, drove at 110 km per hour on a rural road where the speed limit was 64 km per hour, got off the road, slammed into three cars and killed four people. Tests revealed his blood alcohol level was three times the legal limit. He had also consumed cannabis and Valium. The prosecutors sought a twenty years prison sentence. Couch’s attorney argued that his client had affluenza: being raised in a very wealthy family, which never set limits to him, he could not be held fully responsible for his actions. A psychologist testified that Couch was not a responsible agent. The judge accepted the argument and decided Couch needed rehabilitation rather than jail. He granted him a ten-year probation period. The decision triggered a lot of criticisms. Critics pointed out that wealth should not be taken into account in sentence decisions and that such decisions place the wealthy above the law (Eckenroth 2015).

Critics are right to worry about a criminal justice system that applies double standards and punish the same crimes differently depending on whether the culprit is wealthy or poor. However, there might be a grain of truth to the affluenza defence. This paper does not submit that affluenza is a mental disorder, nor that it is a valid legal defence. Defending these claims is beyond the area of expertise of a philosopher. Nor does the paper aim to investigate the significance of the affluenza defence for the philosophical theory of criminal law and punishment. Its aim is to look at whether the affluenza defence might teach liberal political philosophers something about distributive justice. More precisely, if extreme wealth undermines the capacity for individual responsibility (at least in a personal or moral sense, if not in the legal one), then we might hypothesise a negative correlation between high levels of wealth and individual autonomy.1 Individual autonomy (broadly construed) is a goal most liberal theories aim to secure and promote through a just distribution of advantages. It is thus worth examining whether these theories should not be wary of extreme wealth.

This paper attempts to expand this line of thought and develops an autonomy-based argument for limitarianism. Whilst sufficientarianism affirms it is of primary moral importance that everyone gets enough (Casal 2007; Gosseries 2011), limitarianism affirms it is of primary moral importance that no one gets too much (Robeyns 2017). Ingrid Robeyns provides two instrumental arguments in support of limitarianism. According to Robeyns, in our world, limitarianism is instrumental to the achievement of two valuable goals, democratic equality and meeting the poor’s urgent needs (Robeyns 2017, sec. 3 and 4). Robeyns’ argument may be classified as an other-regarding justification of limitarianism. This paper pursues a different argumentative strategy. It does not address the question of whether a person’s having too much prevents others from receiving their fair share of democratic power and material wealth. The question this article focuses on is whether a person’s having too much prevents this very person from accessing a specific good. Hence this is a self-regarding justification of limitarianism.2 Although there is a wealth of empirical literature on the negative impacts of excessive wealth on well-being,3 the good this paper focuses on is not well-being. We could indeed adopt a normative political theory that acknowledges the value of both well-being and autonomy and make trade-offs if needed. Such pluralist self-regarding case for limitarianism might perhaps be stronger. However, political philosophers who are sceptical of well-being-based political morality may find it more palatable to appeal to the sole value of autonomy. They would point out that, in pluralistic societies, there are several competing conceptions of well-being. They could worry that a state that purports to promote a controversial conception of well-being might fail to demonstrate adequate respect for its citizens’ own views about the good life. I shall thus solely focus on the possible benefits of limitarianism for rich people’s autonomy. Even in pluralistic societies, liberal democracies have the mission to promote autonomy because it secures adequate democratic participation as well as the individual capacity to reflect on, revise or exit conceptions of well-being and the good life.

This article thus propounds and discusses arguments to support the two following theses: (1) Above a certain wealth ceiling, a person’s having more material resources does not always increase her autonomy; (2) Above such wealth ceiling, material possession might even be detrimental to the development and the exercise of rich people’s autonomy, or at least some rich people’s autonomy. The aim of such discussion is threefold. Firstly, it purports to support the plausibility of empirical conjectures regarding the detrimental impact of excessive wealth on autonomy. The article gathers together and reinterprets different strands of empirical research in psychology and in sociology for this purpose. Then, it draws the normative implications of autonomy-based political philosophy, should these conjectures be true. The paper argues that a possible implication would be the implementation of a limitarian distribution of material resources through a 100% wealth and income tax beyond a certain wealth ceiling. Finally, bearing in mind that several liberal-minded scholars and policy-makers seem willing to grant that paternalistic measures (such as in-kind transfers) aimed at poor people are legitimate, thereby assuming they are not fully autonomous, the paper aims to restore an epistemic balance between our often critical assessment of the autonomy of the poor and our obliviousness to the lack of autonomy of the rich.

The paper unfolds as follows. Section 2 expounds a multi-dimensional account of autonomy. Section 3 examines hypotheses regarding the ways excessive wealth might undermine the development and exercise of autonomy. Section 4 suggests that, if these hypotheses are true, and even if they are not true for all rich people, a limitarian distribution of wealth might be a tool to secure the rich’s autonomy. It also discusses the level of wealth that should be limited. Section 5 addresses the following issue: if limiting wealth facilitates the development and exercise of autonomy, does it imply coercive measures to prevent people from becoming too rich are justified? Section 6 briefly addresses the problem of incentives.

2. A Multi-Dimensional Account of Autonomy

This section expounds an account of the conditions of autonomy, drawing on literature in political philosophy and on relational autonomy. In the philosophical literature, the word “autonomy” is sometimes used to refer to only some dimensions of autonomy.4 Although addressing the immense philosophical literature on autonomy is beyond the scope of this paper, a clarification of the definition, conceptions, and conditions of autonomy will be useful for its purpose. Autonomy involves multiple dimensions whose interpretations are context-dependent (Mackenzie 2014). In this paper, the account of autonomy aims to fulfil the purpose of identifying the morally desirable distribution of wealth. Autonomy will refer here both to a personal capacity and to the set of conditions that permit and facilitate the development and exercise of this capacity. These conditions may be grouped in two general dimensions of autonomy: internal and external conditions.

Internal conditions refer to the conditions of self-governing agency and involve at least three subsets of conditions. First, the agent must be equipped to a sufficient degree with the mental abilities, capacities, and skills needed to select the appropriate means to achieve a goal, to plan actions and to achieve these plans. These include, among others, the ability to find information and to check the truth or the probability of a claim, the ability to design a strategy, and the skills needed to overcome weakness of the will and procrastination. Since these abilities, skills, and capacities do not necessarily serve autonomously chosen goals, another subset of conditions is necessary: authenticity conditions. Authenticity involves being able to reflect critically upon one’s first-order goals, to revise them so that they cohere with one’s reflectively constituted higher-order commitments and conception of oneself, as well as to revise these commitments and conception of oneself too.5 Feminist critiques have pointed out that agents are unlikely to meet authenticity conditions if they are subject to oppressive socialisation and norms. The third subset of conditions aims thus to enable the agent to regard herself as capable and authorised to define her commitments and act in accordance with them. These conditions include self-trust, self-respect as well as being recognised and treated by others as an autonomous agent.6

For most persons, the exercise and development of a self-governing agency require favourable external conditions. These include independence conditions: interferences of others such as manipulation, indoctrination, pressures and unjustified coercion should be mitigated and eliminated if possible.⁷ External conditions also include securing adequate levels of basic political and social liberties.⁸ These liberties would include freedom of conscience, freedom of expression, freedom of association, freedom of movement, freedom of political participation. For the exercise of autonomy to be meaningful, agents must have an adequate set of options to choose from (Raz 1986, 372–375). An adequate set of options must include a sufficiently varied range of options to enable the agent to make important as well as more trivial choices (Raz 1986, 374–375). For an agent to be able to fully exercise her autonomy, the options offered should not be such that the agent faces a tragic dilemma. In addition, access to these options must be genuine. The agent should have been sufficiently exposed to them and should be able to give them serious consideration.

The account of autonomy provided here is a relational account in two senses. Firstly, its analysis of the factors that impede or help the development and exercise of autonomy emphasises the role of social relations. This is rather obvious for the third subset of internal conditions (self-respect and being treated as an autonomous agent) and for independence conditions. But favourable social and interpersonal relations are also crucial for the adequate development of mental and critical abilities. Secondly, this account considers social relations to be one (but not the only) indicator of the degree of autonomy enjoyed by agents. In other words, autonomy could be at least partly assessed by focusing on social conditions and relations, without having to look into the actual psychological state of the agent.9 As we shall see in the following section, the level of wealth could be one of these indicators of autonomy.

3. How Excessive Wealth Might Undermine Autonomy

The development and exercise of at least a basic degree of autonomy can be jeopardised by a distribution of wealth failing to address material poverty and inequalities. But this consideration focuses on just one facet of the relationship between autonomy and material wealth: the beneficial effects of material wealth on autonomy. If it is likely that lack of material resources damages autonomy, is it equally true that the wealthier one is, the more autonomous? Shouldn’t we also investigate the potential detrimental effects of money on autonomy? To address this question, this section suggests and expounds five mechanisms which suggest that, first, above a certain wealth ceiling, a person’s having more material resources does not always increase her autonomy and, second, material possession might even be detrimental to the development and the exercise of rich people’s autonomy, or at least of some rich people.

The identification of these mechanisms appeals to the analysis of autonomy itself, but also to empirical conjectures. Since this investigation is undertaken from a philosophical armchair, I must lay out the status of such conjectures within the normative reflection pursued here. For this purpose, it is helpful to scrutinise the status of the opposite conjecture, that is, the conjecture that having more wealth (or having a lot of it) is always, or generally, beneficial to the development and exercise of autonomy. Consider the following quote from Rawls’s Theory of Justice as a way of illustration:

“[primary goods] are things which it is supposed a rational man wants whatever else he wants. Regardless of what an individual’s rational plans are in detail, it is assumed that there are various things which he would prefer more of rather than less. With more of these goods, men can generally be assured of greater success in carrying out their intentions and in advancing their ends, whatever these ends may be. The primary social goods, to give them in broad categories, are rights, liberties, and opportunities, and income and wealth.” (Rawls 1999, 79)

According to Rawls, a rational human being should prefer more money and material resources than less. Rawls does not appeal to empirical evidence to sustain this claim. As he makes it clear in the Preface to the 1999 edition of his Theory of Justice (Rawls 1999, Xiii), this claim is not to be understood as a descriptive claim stating a fact about actual human psychology. The “rational man” does not refer to observable individuals. It refers to a conception of the person that embodies certain political ideal democratic societies endorse, the ideal of free and equal persons. These persons are to be regarded as having an interest in developing and exercising their autonomy, or, in Rawls’s terms, their two moral powers, which include rational capacities. The purpose of this conception of the person is not to describe real-world individuals but to derive, justify and systematise principles of justice and the demands these place on institutions. If from the perspective of Rawlsian rational and reasonable persons, it is always better to have more material resources than less, then the only legitimate reason to limit people’s share of resources is the fact that these resources are scarce and subject to competing claims.

These assertions can be challenged. There are reasons not to move too quickly from the claim that free and equal persons have an interest in developing and exercising their autonomy to the claim that free and equal persons should (as a matter of conceptual necessity) prefer having more than less material resources, and from that to the claim that there are no reasons other than scarcity to limit people’s fair share. The analysis of autonomy does not necessarily lead to the implication that more wealth equates to a higher degree of autonomy. Furthermore, some empirical conjectures suggest excessive wealth might be damaging to autonomy in a number of ways.

The following subsections expound and discuss five “mechanisms” through which excessive material wealth might fail to contribute to the development and exercise of autonomy and even undermine it. These mechanisms are to some extent conjectural because the empirical literature on the negative effects of wealth tends to couch these effects in terms other than the conceptions of autonomy prevailing in the philosophical literature. Several studies have investigated relationships between wealth and happiness (e.g. Blanchflower and Oswald 2004; Brickman et al. 1978; Csikszentmihalyi 1999; Cummins 2000; Diener et al. 1985; Diener and Biswas-Diener 2002; Diener and Oishi 2000; Easterlin 1973, 1995, 2001; Frey and Stutzer 2002; Myers 2000).10 But happiness is not autonomy. A happy person can be heteronomous and an autonomous person can be unhappy. With these limitations regarding existing the effects of wealth on autonomy in mind, let us look at the five hypothetical mechanisms through which wealth might undermine autonomy.

3.1. Extreme Wealth Might Hinder the Development of Deliberative Capacities

The first mechanism can be summarised as follows: lack of material constraints might hinder the development of deliberative capacities, which is a condition for the exercise of autonomy. The argument in support of this first mechanism draws on the “ecological” account of rational practical deliberation (Morton, 2011). According to the ecological account, the guiding norms of practical deliberation respond to the interactions between the psychological capacities of the agent and her environment.11 The agent’s environment includes, among other features, material constraints and availability of resources. This means that scarcity, as well as an abundance of material resources, affect the guiding norms of individual practical deliberation. Hume’s discussion of the circumstances of justice provides interesting insights into the effects of wealth on deliberative capacities (Hume 1751 sect III, part I). Hume discusses the scenario of abundance. In such a scenario, justice would be a useless norm, because agents cannot fail to meet its demands. But justice is not the only norm that affects practical deliberation. These reflections on justice can be extended to other areas of practical deliberation (Morton 2011, 568). For example, a guiding norm like long-term planning is unnecessary to an agent who has access to abundant material resources, since this agent cannot fail to acquire such resources for lack of planning (Morton 2011, 570). It seems plausible that successful long-term planning and other deliberative capacities involved in the acquisition of scarce resources require training. Therefore, a state of abundance might sometimes contribute to hinder the formation of deliberative capacities, which constitute one of the dimensions of autonomy. Of course, insofar as non-material resources such as time remain scarce, rich agents still need to deliberate about what ends to pursue with the limited time allotted to them. But they do not have to produce an additional reflection on the wisest use of material resources. In other words, moderate material constraints12 might contribute to the development of deliberative capacities. This mechanism could be the rationale underlying the affluenza defence I mentioned in the introduction in this way (Dart 2014; Eckenroth 2015, 456–457). The affluenza defence suggests that some extremely wealthy agents cannot be held fully responsible for their actions because their wealth has prevented them from forming the deliberative capacities one needs to be appropriately considered an autonomous and responsible agent.

3.2. Extreme Wealth Might Be Conducive to Problematic Adaptive Preferences Formation

One could object to the first mechanism that many rich people seem to have excellent deliberative capacities. For example, Warren Buffett is famous for his wise investment decisions. In addition, insofar as deliberative capacities develop during childhood and young adulthood, the mechanism described above suggests wealth hinders autonomy only at earlier life stages.13 But there is a second mechanism through which wealth might undermine autonomy. This second mechanism can be summarised as follows: excessive wealth, as well as excessive poverty, may induce the formation of problematic, autonomy-deficient, adaptive preferences. In political and social philosophy, adaptive preferences are most often discussed in relation to material poverty and deprivation (e.g. Nussbaum 2000; Sen 1985). Scholars point out that the content of preferences adjusts to conditions of material deprivation and oppression. They argue that preferences formed in response to such conditions should not be considered reliable and authoritative judgment about the agent’s well-being. The questionable status of adaptive preferences is the basis for a critique of subjective welfarism, which assumes that preference satisfaction equates with well-being.

There are several accounts of adaptive preferences (e.g. Khader 2011). Some accounts consider adaptive preferences problematic because they are contrary to well-being. Such accounts endorse an objective theory of well-being. They stipulate that adaptive preferences should not be considered authoritative judgments about well-being because of their content. This implies the satisfaction of adaptive preferences does not make the agent objectively better off (although it may make her subjectively better off). Conditions of deprivation induce agents to downgrade their expectations, to the point that the content of their preferences becomes contrary to their objective well-being. A predictable objection to this account is that it risks failing to respect value pluralism and justifying inappropriate paternalism. But since this article is not concerned with well-being but with autonomy, I move on to discuss accounts that consider adaptive preferences problematic because their formation involves a deficit of autonomy. These accounts stipulate that adaptive preferences are problematic not because of their content, but because the history of their formation involves a deficit in autonomy. For instance, adaptive preferences occur when a change in the set of options induces the agent to unconsciously reverse her preference ordering (Elster 1982). Preference change does not result from deliberate and intentional revision of the agent’s desires, but from a drive, a psychological mechanism of which the agent is not fully aware. What is at stake here is authenticity: preference adaptation is deficient in autonomy when preference reversal does not follow from a conscious revision of higher-order commitments and conceptions of oneself.14

What is interesting is that, although it is commonplace in the adaptive preferences literature, the connection between material poverty and adaptive preferences is contingent (in objective good accounts as well as in autonomy-based ones). Other kinds of circumstances may facilitate adaptive preferences formation. I would like to submit that extreme wealth can also be conducive to problematic preference adaptation. How could that be possible? We can observe that people adapt to affluence, that is, to the standards of living associated with high incomes. Rich people are lastingly or permanently exposed to affluence. The preferences of people who are lastingly or permanently exposed to affluence are formed in response to such exposure to wealthy lifestyles. Now, such adjustment is not sufficient to diagnose adaptive preferences, since non-autonomous adaptive preferences must also involve a complete reversal of preference orderings (Elster 1982, 229). Yet it might be the case that exposure to affluence provokes such reversal. For example, rich people, who are lastingly exposed to affluence, might be induced to evaluate and rank certain lifestyles differently from the way they would have evaluated them had they not been exposed to affluence. They might be led to downgrading options they would have otherwise positively appreciated without having reflected on their reasons for downgrading such options.

The idea that the rich suffer from adaptive preferences may seem counterintuitive. The rich seem to have access to more options than the non-rich. If so, isn’t she more autonomous? In the two following subsections, I shall explain why it might be the case that the rich have actually fewer options than the non-rich, despite their purchasing power.

3.3. Extreme Wealth Might Erode Our Capacity to Revise Our Conception of the Good Because it Habituates Us to Expensive Lifestyles

Even if exposure to a great number of options favoured autonomy, there are reasons to believe the rich do not have access to more options than other people. It is now time to challenge the idea that the rich have access to more options than the non-rich. Of course, many options cost money. At a first glance, it thus seems the rich must have access to more options than the non-rich: they have access to both expensive and inexpensive options, whilst the non-rich only have access to inexpensive options. But this reasoning does not take into account psychological obstacles to the enjoyment of certain options. A psychological mechanism, which constitutes the third mechanism through which wealth might undermine autonomy, may hinder wealthy people’s access to some of the options theoretically available to them through habituation to comfort and expensive lifestyle. As we have already noted, wealthy people are regularly exposed to and get used to lifestyles that are inaccessible to most people. With some exceptions, most wealthy people habituate to expensive lifestyles and a high level of comfort. They are more likely to develop expensive preferences and habits. By saying that a person has expensive preferences, I mean that she needs a comparatively high level of material resources and money to reach a given level of satisfaction. When most human beings might be sufficiently satisfied with a yearly net income of, say, 20 000 €, the person who has expensive preferences might need thirty times as much to reach the same level of satisfaction.

Insofar as she is habituated to affluent lifestyles, the rich are much more likely to need more material resources than the non-rich to reach the same satisfaction level. We can intuitively grasp how easier—psychologically speaking - it is to switch from a frugal lifestyle to a costly one than the opposite. For example, it seems most people do not have issues with transitioning from the standard of living that is typical of students to the one full time paid workers can afford (although they might miss other aspects of student life, such as having a lot of time devoted to learning things for their own sake). By contrast, the decrease in income caused by events such as divorce or the loss of a job seems to cause major decreases in well-being (although income is certainly not the only factor in those examples, it has its importance). Some empirical research provides indirect support to the hypothesis that the more wealthy people are, the more money they are likely to need to reach a given level of well-being. For example, Frey and Stutzer show that the rich’s appreciation of a “sufficient income” is higher than the non-rich’s (Frey and Stutzer 2002).

Expensive preferences do not only result from the fact that one becomes accustomed to a high level of comfort and luxury. They also result from prevailing consumption norms in one’s social group of reference. Robert Frank illustrates this phenomenon with the following example: suppose a person wants to replace her old 90$ outdoor grill. Nowadays, most people in her social group buy upmarket grills, which may cost up to 5000$. This person starts wondering whether she should not replace her 90$ grill with at least a 1000$ model. The fact that other people in the social circle of this person spend so much money on luxury grills changes the conventional definition of what an acceptable grill is in a given social group (Frank 1999, 10–11). Since rich people tend to frequent other rich people, it becomes inconceivable for them not to follow expensive consumption patterns.

To summarise, wealthy people are more likely to develop expensive preferences because they get used to luxuries available to them and because their consumption patterns tend to replicate those of other rich people. We can thus argue that expensive preferences impede autonomy because they increase the psychological costs of revising one’s conception of the good life. The argument is the following: a person who is rich becomes habituated to the standard of living associated with wealth. She thus develops expensive preferences and habits. Since expensive preferences render us less capable of being satisfied with little material resources, it is more difficult to transition from expensive preferences to non-expensive ones than the reverse. Now, each possible conception of the good life is only compatible with a limited set of standards of living. For example, many interesting careers are not likely to make someone very rich: farmer, teacher, artist, nurse, priest, police officer, baker, carpenter, musician or reporter, to name only a few. Choosing or transitioning to such careers is likely to be psychologically difficult for someone who has expensive preferences. Such psychological costs and obstacles can thus prevent an agent from considering many potentially worthwhile lifeplans. In that sense, expensive preferences created by wealth erode our capacity to revise our life plans and to act upon our authentic judgments. Less comfortable options are theoretically available to the rich, but in practice, they are rarely seriously considered. Since being able to revise our lifeplans and to act upon our authentic judgments is a crucial dimension of autonomy, extreme wealth may, through the mechanism I have just described, erode autonomy.

3.4. Extreme Wealth Might Erode Our Capacity to Revise Our Conception of the Good Because it Could Trigger a Fear of a Drop in Status

Another psychological obstacle that might render some options unavailable to extremely wealthy people is related to over-attachment to social status. The key idea of this fourth mechanism is that being wealthy induces a fear of a drop in status, which undermines authenticity and restricts the range of options the agent has genuine access to. The identification of the mechanism starts with the observation that we, human beings, have a tendency to desire to keep up with the people who have the same social status as us (or a slightly higher social status). Social status is generally related to wealth. If we want to keep up with those who have a given social status, we are likely to want to keep up with those who have a certain level of income. The higher our income and wealth are, the higher the social status we want to keep up with. Empirical evidence shows that an increase in income leads to an increase in social aspirations (Diener 2000). The problem is that, if wealthy people want to keep up with a high social status, their life choices have to align with this goal. Their choices must not conflict with the need to keep up with other rich people. Therefore, in order to retain their social status, they will be prompted to eliminate certain possibly valuable options from their options set, including choices of career and marriage partners. This means that the only options most agents seriously consider are the options that involve them being at least as wealthy as they currently are (or as wealthy as their parents currently are). Although rich people can conceivably flourish with less money and status than they currently have, their significant life choices (including career choices and the choice of a partner) are likely to be driven by a “fear of a drop in status”.15 They do not worry about not having enough material resources to pursue the conception of the good they genuinely value (this worry is perfectly compatible with the exercise of autonomy). They worry about keeping their status, and this prevents them from giving serious and genuine consideration to alternative palatable career or marriage options. If we frame the problem in terms of adequacy of options, this means the higher one’s wealth-related social status is, the fewer options one has, since there are only a few social positions that still deserve consideration, given the fear of dropping. The higher one’s wealth-related social status is, the less scope one has to revise one’s goals so that they cohere with one’s authentic self.

This mechanism belongs to a class of mechanisms through which having more choices may actually make people less free, due to the expectations and pressures from others that accompany these new choices. Gerald Dworkin gives the example of the choice of determining the sex of one’s children. He suggests having this choice may not make future parents freer, because of “the social pressures that are likely to be exerted on parents to produce one sex rather than another (the grandparents who always wanted a little girl or the community that needs more soldiers)” (Dworkin 1988, 68). On the one hand, parents have more freedom of choice. On the other hand, they have less autonomy since this new choice provides others with a reason to put them under pressure (a violation of the independence condition of autonomy) and may threaten their capacity to live in accordance with the values they cherish the most (a violation of the authenticity condition). Analogously, when we mitigate the pressure to conform, the rich seem to have more options with respect to career choices than the poor. But once we take into account such pressure, it might be that careers such as electrician, baker, nurse or primary school teacher are in fact inaccessible to the rich.

At this point, the reader might wonder why the choice to keep up with a high social status should be considered less autonomous than the choice to pursue a career or to marry a partner that does not fit well with the expectations associated to a high social status. The reader might think rich people can deliberate carefully about options involving a drop in status and consciously settle upon prestigious career and marriage choices. Insofar as a choice is autonomous in virtue of its history rather than in virtue of its content, the choice to keep up with a high status can conceivably be considered as autonomous. However, sociological studies suggest that the education and socialisation of the progeny of rich people are designed in such a way as to ensure rich families will retain their status across generations (e.g. Pinçon and Pinçon-Charlot 2009, 101–111). Spatial segregation and endogamy act as safeguards against individual choices that could threaten the sheer existence and interests of the very rich (Pinçon and Pinçon-Charlot 2009, 52–68). To describe the phenomenon in Pierre Bourdieu’s terms, rich people are characterised by a habitus, that is, a set of stable dispositions, which includes beliefs, desires, values, and behavioural patterns. Habitus is the embodiment of the fact that one belongs to a given social class (Bourdieu 1979, 112–113; 1984 133–136). Such habitus is not freely and deliberately acquired. Habitus is acquired through class socialisation and conditioning. Habitus contributes itself to the reproduction of the conditions of class socialisation. Rich bourgeois people and working-class people each have their own habitus. Habitus involves a hierarchical classification of lifestyles: the habitus of the rich is ranked higher than the habitus of the poor (for example, their artistic tastes will be deemed more refined than those of working-class people). Such unchosen classification turns into a virtue in the fierce competition for wealth and power by inducing agents to select options that fit well with their social group (Bourdieu 1979, 195). Therefore, study and career paths, spatial segregation, endogamy, consumption habits, and aesthetic tastes are part of the strategies the rich use to retain their social position and ensure social reproduction over time. Such strategies are all the more effective as they are unconsciously pursued (Bourdieu 1979, 285).16 If a person becomes fully aware that she marries another person in order to retain her social status, and not because she is genuinely in love with her partner, she will probably start wondering whether she should not consider other marriage options, whether such options are not valuable too, whether the goal she pursues by marrying a rich partner is genuinely her own, and so on. The rich, whose interests qua members of the social group or class of rich people, are the most likely to be threatened by a change of social status involved by an atypical marital or career choice (since there are fewer social status options enabling them to keep up with their current social position). Although the fear of a drop in status is present in every social group (except the lowest ones, who do not have anything to lose), it is likely to be stronger in the upper layers of society. Therefore, other things being equal, and insofar as a fear of a drop in status hampers the capacity to elaborate and revise an autonomous conception of the good life, the rich might be less autonomous than other social groups.

The idea that psychological obstacles (broadly understood) can restrict the number of options available to the agent may raise the following worry: having strong commitments, such as religious or ethical commitments, seems also to lead the agent to disregard a range of potentially valuable options.¹⁷ But the line of reasoning developed above does not imply autonomy-based normative theories should be wary of strong commitments (this would be odd). From an autonomy-based perspective, the crucial problem with the third and the fourth mechanism is not the mere fact that wealth limits the rich’s option. The crucial problem pertains to the process through which the rich, or some rich, are induced to fail to consider a range of options. This process has nothing to do with rational deliberation and genuine commitments. It is triggered by unthinking dispositions such as habituation, fear and social habitus. Likewise, the sheer fact that agents cease to consider a range of options other might deem valuable as a result of their religious or ethical commitments is not problematic from an autonomy-based perspective. It would be problematic only if these agents came to embrace such commitments in the wrong kind of way, for instance as a result of anxiety.18

3.5. Extreme Wealth Might be Incompatible with Transparency with One’s Own Values

The fifth mechanism can be summarised as follows: insofar as, in a world of finite resources, an extremely wealthy lifestyle is incompatible with social and environmental justice, and insofar as we, humans, tend to shun the belief that our own behaviours and values are harmful to others, extreme wealth is not conducive to transparency with one’s reasons for action, which is a condition of autonomy. This last mechanism is related to the logic exhibited by Festinger’s theory of cognitive dissonance (Festinger 1962). In a nutshell, cognitive dissonance refers both to the inconsistencies between one’s beliefs, or between one’s values and one’s actions, and to the discomfort these inconsistencies generate. Evidence shows humans shun such inconsistencies. We are motivated to resolve them, either by revising our beliefs or by changing our behaviours. Recently, research on cognitive dissonance has focused on the hypothesis that the main motivation for overcoming dissonance is to sustain the belief that one is a good person (Monin 2008). Cognitive dissonance theory may explain why, in consumerist societies, there is be a general tendency to ignore or minimise information about environmental problems requiring major consumption behaviour changes (Kollmuss and Agyeman 2002, 254). When we face a conflict between our ethical values (environmental preservation or social justice) and our desires (living a comfortable and luxurious life), we unconsciously solve the inconsistency by selectively perceiving the information that confirms the path of behaviour we want to adopt, or ignoring or minimising the information that contradicts it.

As to the rich, their path of behaviour and even the sheer existence of extremely wealthy people are likely to be inconsistent with a wide range of conceptions of social as well as environmental justice. In a world of finite resources, the appropriation of a significant quantity of resources by a small minority of people will threaten the access of other people to their fair share and even their capacity to meet their basic needs. Wealthy people do not only save and consume resources over which others may have legitimate claims, they also tend to adopt practices inconsistent with the stability of just institutions, such as fiscal evasion and political pressure to reduce taxation on income and wealth. Furthermore, luxury consumption habits such as frequent travelling are probably not compatible with the long-term preservation of the capacity of ecosystems to meet human needs. Some would probably object that rich people also invest their capital, that such investments are required to improve the fate of less advantaged social groups (through job creation, for instance) and that rich people need to be incentivised to make such contribution to the social product. If this logic is true, the fact that some people are extremely wealthy is not incompatible with social justice. Yet there are a couple of reasons to doubt the validity of such logic. First, recent economic history suggests that, although some inequalities might be needed to incentivise people to invest their financial as well as their human capital, those inequalities need not be as extreme as they currently are: economic history suggests top income shares were substantially lower in the postwar decades than they are now (e.g. Atkinson et al. 2011). The point here is not to argue for a specific conception of social justice, but to suggest that a broad range of widespread and plausible conceptions of social justice is unlikely to deem the existence of extremely rich people acceptable. If this is the case, the extremely rich person who, like most of us, wants to sustain the belief that she is a good person might face an inconsistency between her path of behaviour and plausible conceptions of social justice. In order to avoid cognitive dissonance and solve this inconsistency, she could either change her behaviour, give away her money (but this is unlikely to happen), or change her beliefs about social justice (this is more likely to happen). She might thus be induced to ignore, minimise or reinterpret true information on the harmful impacts of her behaviours and actions. She might assess conceptions of social justice as well as empirical research in economics and social sciences not on the basis of their true merits, but on the basis of their consistency with the existence of her social class. She might also come to believe that cynicism is de bon ton. She might wholeheartedly embrace the belief that wealth has beneficial trickle-down effects on the poor, or that charity is more efficient than tax and transfers, not because these beliefs are valid (although they might be so), but because they are consistent with her sustaining her path of behaviour, values, and lifestyle.

Cognitive dissonance is bad for autonomy because autonomy involves transparency with one’s own values and reasons for actions. One way of regaining autonomy is to minimise the importance of sustaining a given path of behaviour or to revise our desires so that satisfying them does not contradict our values. Insofar as wealthy people tend to have a wealthy lifestyle, and people who have a wealthy lifestyle tend to want to sustain it, they are likely to be less transparent with their values and reasons for beliefs and actions, and thus less autonomous.

4. Autonomy-Based Limitarian Distributive Justice

The preceding section has expounded five mechanisms which suggest that, above a certain wealth ceiling, a person’s having more material resources does not always increase her autonomy, and might even be detrimental to it, at least for some rich people. Let us move slightly further and examine the normative implications of autonomy-based political philosophy if the negative impacts of wealth on autonomy prove to be true. One way to address these negative impacts consists in preventing people from having too much in order to protect their autonomy. Distributive policies, through which a specific allocation of material wealth and resources is achieved, are a possible tool for this purpose. If wealth undermines the development and exercise of autonomy in several ways, preventing the effects of excessive wealth on autonomy will provide a reason—albeit not a decisive one - to justify a limitarian distribution. In a limitarian distribution, citizens would be prevented to acquire or receive too much material resources in order to protect their autonomy.19 Such limitarian distribution could be achieved by a “limitarian tax”, that is, a 100% wealth and income tax above a certain wealth ceiling.

4.1. Restoring the Liberal Paternalist Balance Between the Rich and the Poor

At this point, the reader is likely to worry that we are deriving too rapidly a controversial normative proposal from empirical conjectures. Political philosophers and policy-makers cannot read in rich people’s mind to determine with certainty whether they are really autonomous or not, and might thus unjustifiably treat them as non-autonomous. But when we shift our attention from the rich to the poor, however, it appears that some liberal philosophers, economists, and policy-makers seem willing to endorse policies and practices that presuppose that some categories of (non-rich) citizens need some help in order to exercise and develop their autonomy. Examples of such policies include the public provision of goods and services in kind rather than in cash.20 Few liberal-minded philosophers, economists, policy-makers or citizens advocate replacing the public provision of education or health with their equivalent in cash.²¹ Although the public provision of some collective goods may be justified on efficiency grounds,²² in-kind provision of goods such as housing or food subsidies seems to a large extent motivated by paternalistic concerns (e.g. Musgrave 1959; Thurow 1976; Currie and Gahvari 2008). As Thurow puts it:

Consumer sovereignty refers here to autonomy. Now, although in-kind aid presupposes that some individuals are not sufficiently autonomous, it does not presuppose that all beneficiaries are not autonomous. It suffices that only some of them are to justify mildly paternalistic policies. Hence several liberal-minded thinkers find acceptable to treat autonomous citizens as non-autonomous agents in order to make sure they less autonomous fellow citizens do not jeopardise their future well-being and autonomy. What is striking, however, is that we seldom consider applying the same line of thought to rich people.23

If we believe that mildly paternalistic policies (such as in-kind provision) aimed at the poor are justifiable in a liberal democracy, despite the fact that some members of the target group are sufficiently autonomous, we should be willing to consider mildly paternalistic policies aimed at the rich. Of course, a seemingly genuine liberal alternative is to have no paternalistic policies at all, be it for the rich or for the poor. The problem is that, although such policy succeeds in treating people as if they were sufficiently autonomous, it may fail at securing real autonomy in the long run: the capacity for autonomy comes into degrees and continues developing throughout human life.

These considerations speak in favour of not just a limitarian distribution of wealth, but in favour of a distributive pattern combining limitarianism and sufficientarianism. Sufficientarianism contributes to secure autonomy for several reasons. Material poverty and important economic inequalities hamper independence, liberty, and access to an adequate set of options. Poverty and inequalities jeopardise a person’s bargaining power and make her subject to the will of others. Moreover, when extremely wealthy citizens have a decisive influence on the outcome of supposedly democratic processes, the “fair value” of political liberties is not guaranteed anymore (Rawls 2001, 148–150). The parallel between poverty and wealth suggests a coherent limitarian theory of justice is likely to endorse sufficientarianism as well.

4.2. How Much Is Too Much?

Readers may wonder how much is too much. Giving a number is difficult because the impact of material wealth on an individual’s autonomy depends on a variety of factors, including economic conditions such as inflation and current standards of living. In addition, the capability approach has taught theorists of distributive justice that differently abled people need different amounts of resources to achieve the same standard of living (e.g. Robeyns 2011). Thus, the wealth ceiling for a person with a long-lasting health condition to be able to develop and exercise her autonomy must be higher than for healthy individuals. However, it is important to offer ways to identify the level of wealth that should be limited to provide guidance as well as to enable us to test our intuitions.24

If the mechanisms described in the preceding section prove to be true, they will provide some guidance to establish the appropriate wealth ceiling. The first mechanism suggests a person owns too much once she need not taking into account material constraints in her practical deliberation anymore. The problem is that the extent to which one has to take such constraints into account does not just depend on one’s level of wealth. It also depends on the financial costs of one’s ambition, on the availability of publicly funded provision of goods and services (if a person’s ambitions include attending university, this presumably will be less costly in countries that provide state-funded higher education), as well as on her own perception of how rich she is. After all, Walt Disney’s multimillionaire duck, Uncle Scrooge, remains overly preoccupied with material constraints. Perhaps the best rule of thumb would be to make a survey and ask people how much one must possess to be financially comfortable.

The second mechanism, the adaptive preferences mechanism, provides more straightforward guidance when it is combined with the third and the fourth one. The third mechanism suggests that a person is too rich when her wealth is such that she becomes accustomed to a level of comfort and luxury it would be difficult to renounce. This suggests the wealth ceiling must be quite low since the average, middle-class standard of living in Western countries is likely to be difficult to abandon for the average level of comfort in other parts of the world. Hence the wealth ceiling might be barely higher than the basic “poverty” threshold of material resources a person needs to develop and exercise a sufficient level of autonomy (note that such threshold would presumably be much higher than the poverty line put forward by the World Bank).

The fourth mechanism is triggered by inequalities of social status, which are related to inequalities of wealth and income. This means that, in order to prevent this, society should attempt to move closer to equality of social status. As a result, the wealth ceiling should be such that the gap between this ceiling and the autonomy-based poverty threshold does not lead to a significant inequality of social status. A complication is that inequality of social status is not only due to inequality of material income and capital but also to inequalities of social and cultural capital.

The fifth mechanism suggests that the limitarian distribution of wealth must attempt to prevent cognitive dissonance. Here, cognitive dissonance is involved when there are conflicts between a person’s willingness to keep her share of material resources and her ability to assess different normative and empirical views of distributive justice on the basis of their true merits (rather than on their tolerance for this person’s preference for keeping what she has). This suggests that the wealth ceiling should adjust so as to prevent the rich from getting more than their fair share. However, an important clarification is needed here. It is not exactly the conflict between a person’s preference for keeping her money and a view of distributive justice philosophers and economists could deem valid that triggers cognitive dissonance. Cognitive dissonance is more likely to be triggered by the possible conflict between a person’s preference for keeping her money and widespread views of distributive justice among laypersons since such views are more readily available. Therefore, the autonomy-based wealth ceiling should be close to people’s beliefs about how much is too much.

The result of this brief discussion seems to be twofold. First, the wealth ceiling is likely to be close to what laypersons think the wealth ceiling should be. One way of knowing that is to conduct a survey on people’s beliefs about the wealth ceiling. This task has recently been undertaken by Ingrid Robeyns (Robeyns 2018). Another way is to take inspiration from existing political campaigns and proposals, which attempt to echo electors’ intuitions about the issue. In 2013, a Swiss popular initiative labelled “1:12” proposed a law that would have prohibited firms from offering salaries more than twelve times higher than the lowest salary (the initiative was ultimately rejected). In 2017, French presidential candidate Jean-Luc Mélenchon proposed a 100% tax on yearly incomes above €400 000 (about $460 000).25 However, if we wish to use the €400 000 income ceiling as a rule of thumb, we should bear in mind that further adjustments would be needed to take into account inflation, purchasing power parity, international differences in standards of living, provision of public goods and services as well as interindividual inequalities of capabilities. Thus, the wealth ceiling for disabled persons or for those who live in expensive cities like New York, London or Paris might be higher—but this would also depend on the availability of public goods such as healthcare facilities or affordable and efficient public transportation. Second, the gap between the autonomy-based wealth ceiling and the (probably high) autonomy-based sufficiency threshold is likely to be narrow. This is not because inequality is considered bad in itself: this paper derives limitarianism from the value of autonomy, not from the value of equality. This is because the amount of material wealth everyone has access to should be sufficiently high to secure independence, adequate options as well as the proper development and exercise of mental and critical abilities, but also sufficiently low to prevent habituation to a high level of comfort and luxury, fear of a drop in status and cognitive dissonance. From this perspective, the €400 000 wealth ceiling might be already too high.26 The wealth ceiling might thus be situated somewhere between €400 000 and the amount of money each individual would possess in a strictly resource egalitarian society, and this ceiling would be adjusted for purchasing power and capability parity.

5. Autonomy-Based Limitarianism and Legitimate Coercion

Thus far, the article has suggested empirical hypotheses in support of the theses that, above a certain wealth ceiling, a person’s having more material resources does not always increase her autonomy and might even be detrimental to it. Then, it has argued that, if these theses prove to be true, a possible normative implication would be the implementation of a limitarian distribution of wealth beyond a certain ceiling. Achieving a given distribution of wealth requires unpopular measures such as taxation. To secure a limitarian distribution of wealth, a 100% tax rate on wealth and income above a determinate threshold might be needed. However, such a proposal raises a puzzling issue. On one hand, taxes seem to involve the use of the state’s coercive power to protect people’s autonomy. On the other hand, coercion is damaging to autonomy. Are taxes aiming at promoting the autonomy of the taxpayer necessarily problematic from a liberal perspective? To address this question, I shall draw from Joseph Raz’s seminal analysis of the relationship between coercion and autonomy.

Raz warns us that coercion as a method of encouraging people to act for their own good is suspect: “we are all too familiar with the danger of exaggerating the degree to which people’s well-being can be promoted in flat contradiction to their formed judgments and preferences” (Raz 1986, 151). Yet he affirms liberals should not exaggerate the evils of coercion. Coercion may legitimately be used “to secure natural and social conditions which enable individuals to develop an autonomous life” (Raz 1986, 156). Coercion is a notion that involves both descriptive and evaluative dimensions (Raz 1986, 148–157). Descriptively, coercion occurs when the coercive agent threatens the coerced agent to make her worse off if she does an action A the coercive agent wants to prevent her from doing, and when the threat is effective (the belief the threat will materialise is part of the reasons for the coerced agent not to do A). There are also two evaluative dimensions of coercion: a threat is a coercive one if (i) it invades the coerced agent’s autonomy and (ii) the fact that someone acted under coercion counts as a justification or a complete excuse for her action (Raz 1986, 150).

One might wonder whether a tax is really coercive. In The Morality of Freedom, Raz suggests that taxes, as well as subsidies, are noncoercive means the state can legitimately use to promote certain ideals (e.g. Raz 1986, 416). Taxation does not involve an overt threat. If this is correct, then we must not worry about the illegitimate use of the coercive power of the state to secure a limitarian distribution through taxation. However, taxation is not as inoffensive as it seems to be. It is coercive in the sense that it manipulates the choice menu and the costs and payoffs associated with each option (Waldron 1988, 1142). This means taxpayers’ decision to save, give or earn money is altered by the fact that the state’s action has attached new consequences to these options. Yet this need not be always morally problematic. Some taxes pursue goals that justify coercion. Is it the case for the limitarian tax? To address this question, let us have a closer look at the idea that coercion invades autonomy. Drawing on Raz’s analysis, it appears that coercion can invade autonomy in three ways. First, coercion reduces the quantity and the quality of the options available to the coerced agent. Coercion eliminates an option without creating a desirable alternative. Second, even when its effects on the agent’s ability to freely choose the life she values are negligible, a coercive act remains problematic if it insults the coercive agent’s autonomy by treating her as a non-autonomous agent. Third, coercive acts interfere with autonomy because they deliberately change the agent’s reasons for acting as she does. Even milder forms of coercion erode autonomy because they increase the opportunity costs of acting against the will of the coercive agent and modify parts of the reasons for the agent to act as she does.

Is a limitarian coercive measure like a tax problematic in one of these respects? Regarding the reduction of available options, the limitarian tax might actually increase the range of options accessible to the agent. If the empirical hypotheses expounded in Section 3 are correct, excessive wealth develops dispositions such that the rich cannot seriously consider otherwise valuable options anymore. Limitarian coercive measures could create desirable options for the formerly rich, such as the option to become a professional baker.

What about the second way in which coercion invades autonomy? Does a limitarian tax invade the autonomy of those subjected to this tax by failing to treat them as autonomous agents? The autonomy-minded limitarian could say: a state which implements coercive measures does not express disrespect for the autonomy of its citizens if these measures are precisely motivated by a concern for individual autonomy (Raz 1986, 156–157).

If it is true that wealth undermines autonomy, a limitarian tax does not seriously invade autonomy in the first two respects. However, I submit that a limitarian tax justified by a concern for autonomy would interfere with the autonomy of the rich in the third respect because it deliberately changes their reasons for acting as they do. A coercive tax changes the context of choice and thereby the reasons to choose an option rather than another (Waldron 1988, 1145–1146). A limitarian tax would induce the rich to become less rich not because they genuinely believe excessive wealth undermines their autonomy, but because they want to avoid paying even higher fines. Encouraging people to exercise their autonomy through financial penalties does not make them understand or commit to the autonomy-based reasons which motivate these penalties. Most likely, if the rich do not endorse the goal these fiscal policies attempt to promote, they will merely try to get around them and hide their money in tax heavens. Furthermore, recall that the account of autonomy the paper’s argument relies on is relational. Treating a group of people as agents who are not fully autonomous would fail to meet one of the conditions of autonomy. An autonomy-promoting society should enable the agent to regard herself as capable and authorised to define her commitments and act in accordance with them. These conditions include self-trust, self-respect as well as being recognised and treated by others (including state agents) as an autonomous agent. Even mandatory autonomy-promoting education requires educators to recognise and treat children as much as possible as autonomous agents, taking into account their age and developmental stage.

An autonomy-based limitarian tax may thus invade rather than promote the rich’s autonomy in one respect, that is, in that it may prevent the rich from acting upon their own, authentically formed reasons. But it is not sure this consideration should detract autonomy-based normative theories from giving further consideration to limitarian coercive measures. Firstly, one might conjecture that, through the five mechanisms suggested above, excessive wealth might be more damaging to the formation of autonomous conceptions of the good life than the limitation of freedom involved by wealth taxes. But this is conjectural. Secondly, if we consider the limitarian tax not as a stand-alone policy, but as a complement to other policy measures, its problematic aspects dissipate.

Consider how the autonomy-promoting limitarian tax would fit in with a more general theory of distributive justice.27 Even though the primary purpose of a limitarian tax is not redistributive, the argument deployed here suggests the fiscal implications of distributive justice might, as it happens, concur with the promotion of autonomy. Liberal, autonomy-based, theories of justice can respond to the libertarian objection that redistributive taxation is an illegitimate use of coercive power (e.g. Nozick, 1974: 167–174) along the following lines. The use of coercive power is legitimate if it is necessary to prevent the coerced agent from causing harm to others (Raz 1986, 412–420). From an autonomy-based perspective, a person is harmed when her autonomy is diminished. If a sufficientarian distribution of wealth is necessary to secure autonomy, redistributive compulsory taxation aiming at securing such distribution is morally legitimate. If Rawlsian egalitarian distributive principles (or other distributive principles, such as sufficientarian ones) were agreed upon by reasonable citizens, taxation could be a permissible tool to bring about the legitimate distribution of wealth. Whether we endorse a monist, autonomy-based, conception of distributive justice, or a pluralist one, combining autonomy and equality, the use of coercive power through compulsory taxation to bring about the just distribution of wealth seems legitimate. Yet this line of argument cannot, by itself, invalidate the claim that redistributive taxation requires interference with the autonomy of the taxpayer. It can only justify such interference either on the grounds that redistribution ultimately increases the autonomy of the beneficiaries or on grounds other than autonomy.28 In other words, if we accept that redistribution interferes with taxpayers’ autonomy, the justificatory burden inevitably falls on the shoulders of the liberal champions of redistributive taxation. They must provide sufficient justifications to show that redistributive taxation either increases other people’s autonomy or protects and promotes other values (such as equality), which have to be balanced against the value of autonomy. But, thanks to the limitarian argument, they might not have to do that. If the standard assumption that wealth is always beneficial to autonomy is not true, as suggested in Section 3, then redistributive taxation does not interfere with autonomy (at least above the limitarian lines). Challenging this standard assumption lightens the burden of justification liberal champions of redistributive tax-and-transfer schemes must bear. And the distributive justice-based legitimation of coercive taxation provides autonomy-based limitarianism with an escape from the challenge that coercion might not be the best way to enforce autonomy.29

Another line of defence against the worry that coercion damages autonomy suggests the limitarian tax could be considered a catalyst rather than a constraint. Autonomy-based limitarianism is likely to imply that educational institutions should encourage the development of limitarian dispositions. Educators should discourage the pursuit of wealth beyond what is necessary to secure adequate autonomy. Such educational practices, if consistent with the development of deliberative and critical thinking skills, would equip learners with the capacity to find, reflect on, endorse and possibly challenge the autonomy-based reasons for not becoming too rich. Would such educational policies dispense society with limitarian taxation? Perhaps not in a society like ours, in which wealth and materialistic pursuits are highly valued. In such a society, it might be appropriate to prevent people from becoming too rich through compulsory taxation even though they endorse the philosophical arguments in favour of limitarianism. In such a society, expecting people to act on limitarian reasons may be excessively demanding, because it would require them to run counter to established social norms. Those who received an education aiming at cultivating limitarian dispositions might fully agree with limitarian values, and yet find themselves unable to act upon them. Phenomena such as commercial advertising, peer pressure, expectations from partners and children, the way the labour and consumption market are organised, how the urban environment is shaped, hinder our capacity to live autonomously. This is why education, or at least school education, may not suffice to discourage the pursuit of material wealth in our societies. Whereas if we were just prevented from becoming to wealthy and thereby to adopt expensive lifestyles, we could reconcile the values we have reasons to have with our patterns of behaviour. Therefore, if it is coupled with genuinely autonomy-promoting education, compulsory limitarian taxation can thus help people to live in accordance with their own reasons. Taxation would be a catalyst, not an interference.

The result of the discussion of the first potential objection to autonomy-based the limitarian tax may be summarised as follows. The discussion draws from Raz’s account of the relationships between autonomy and coercion. A limitarian tax is a coercive measure. A coercive measure is morally problematic if it invades autonomy or if it could count as a justification or as a complete excuse for the coerced act (Raz 1986, 150).30 The limitarian tax seems unlikely to have repercussions on whether the actions of those who would pay this tax would be justifiable or excusable. Hence a limitarian tax is above all justified if it does not invade autonomy. If the thesis that wealth may restrict the rich’s options is true, a limitarian tax does not invade autonomy by restricting the rich’s options. Nor does it fail to treat the rich as an autonomous agent since the tax is motivated by a concern for the rich’s autonomy.31 Nevertheless, the tax invades the rich’s autonomy in a third way, by deliberately changing their reasons for action. But such consideration should not detract us from giving autonomy-based limitarianism further consideration. First, autonomy-based reasons may work as a complementary justification for tax-and-transfer schemes aiming at securing an egalitarian or a sufficientarian redistribution. Second, if limitarian taxes are implemented in conjunction with limitarian education, they will work as a catalyst by helping people to live in accordance with the values they acquired through education.

6. The Incentive Objection

Suppose society becomes limitarian thanks to a combination of fiscal policies and educational practices. A limitarian society might face the following challenge. Insofar as members of such limitarian society would remain self-interested maximisers,32 a widespread limitarian ethos may cause them to substitute leisure for income once they have reached a given threshold of wealth. Now, suppose some people have more economically productive capacities than others. If such people ceased to be attracted by the prospect of high earnings, one of the incentives to contribute significantly to the creation of wealth by working more would disappear. Those who are below the sufficientarian threshold, would not benefit any more from their efforts.

The challenge seems akin to the problem addressed by the theory of optimal taxation.33 However, the theory of optimal taxation does not have autonomy as its main focus. The theory of optimal tax progressivity assumes that policies should aim to increase social welfare. A theory of justice derived from the principle that autonomy is of primary moral importance does not have social welfare as its primary goal. Therefore, such theory draws different conclusions with respect to the extent to which the high-skilled should be motivated to work hard in order to contribute to the less lucky citizens. The social and economic objective is to secure the material capital society needs to guarantee everyone the capacity to develop and exercise autonomy.

If what matters is securing the economic, social and political condition of autonomy, the high-skilled should only be motivated to work to the extent that it is necessary to secure such conditions. It is unnecessary to incentivise them to create wealth beyond what is needed to secure autonomy. In Section 4, I have suggested these conditions include a sufficientarian-limitarian distribution of wealth. They would also include the material costs of securing other conditions of autonomy development, such as the implementation of an autonomy-promoting educational system. How much social wealth would be necessary to secure autonomy remains a question that I have no space to adequately address here. One might guess that, in such a society, individual shares of material wealth might not be necessarily very high. However, the aim of developing a reasonably ambitious level of autonomy might require significant investments in collective goods such as educational institutions of various sorts. Provision of adequate health care and of a financial safety net might also be needed. Therefore, we cannot exclude that securing autonomy might require a quite high social product. If so, autonomy-based distributive justice theorists have to reflect on the morally permissible and desirable ways to induce the most competent to create high levels of social products without being able to earn more than the wealth ceiling.

7. Conclusion

This article has attempted to support the plausibility of two empirical hypotheses regarding the impact of excessive wealth on autonomy: (i) above a certain wealth ceiling, a person’s having more material resources does not always increase her autonomy; (ii) above such wealth ceiling, material possession might even be detrimental to the development and the exercise of rich people’s autonomy, or at least some rich people. Starting from a relational account of autonomy, this paper has discussed five mechanisms that may challenge the standard assumption that material wealth always increases autonomy. Such mechanisms even suggest excessive material wealth could be detrimental to autonomy in a number of ways. The article has also suggested (Section 4.2) that these mechanisms may provide helpful guidance to determine a wealth ceiling. The next part of the paper has investigated whether, if it is true that excessive wealth undermines autonomy, autonomy promotion may justify coercive measures such as taxation. It might be the case in one respect: coercive measures alone tend to fail to commit those subjected to them to their rationale. However, it appears that, if a limitarian tax is coupled with either redistributive purposes or autonomy-promoting educational policies, concerns raised by the effects of coercion of autonomy might dissipate. In particular, if limitarian fiscal policies aligned with educational practices, autonomy-promoting taxation would become a device that would help people to act upon what they have reason to value rather than a coercive measure. But the problem of incentives might remain: in a limitarian society, the material incentive to contribute more to the social product would be less strong. The extent to which this would be problematic would depend on the extent to which securing sufficient autonomy would require creating high levels of wealth.

Acknowledgement

Versions of this article have been presented at the Political Philosophy and Philosophy of Law Video Workshop jointly organised by the universities of Cordoba and Louvain, at the Workshop in Economics & Philosophy (Université catholique de Louvain), at the Political Theory Group (University of Glasgow) and at the Philosophie et théorie économique seminar (University of Reims). I am grateful to all the participants for their helpful comments and questions, and in particular to Antoinette Baujard, Jean-Sébastien Gharbi, Brian Girvin, Axel Gosseries, Cyril Hédoin, Carl Knight, Louise Lambert, Lucas Misseri, George Pavlakos and Pierre Van Zyl. I also wish to thank two anonymous reviewers for the detailed and insightful comments they provided on earlier versions of the manuscript. All errors are my own. The work reported in this publication had benefited from a grant from the Fonds Spécial de Recherche (FSR) of the Université catholique de Louvain. This chapter was first published in Ethical Theory and Moral Practice, 2018, 21: 1181–1204.

References

Angelescu, Laura. 2014. Rich People, an Overview. In Alex C. Michalos (Ed.), Encyclopedia of Quality of Life and Well-Being Research (pp. 5566–5569). Dordrecht: Springer, Netherlands. https://doi.org/10.1007/978-94-007-0753-5_2521

Atkinson, Anthony B., Piketty, Thomas & Saez, Emmanuel. 2011. Top Incomes in the Long Run of History. Journal of Economic Literature, 49, 3–71. https://doi.org/10.1257/jel.49.1.3

Benson, Paul. 2005. Taking Ownership: Authority and Voice in Autonomous Agency. In Joel Anderson & John Christman (Eds.), Autonomy and the Challenges to Liberalism (pp. 101–126). https://doi.org/10.1017/CBO9780511610325.007

Blanchflower, David G. & Oswald, Andrew J. 2004. Well-being over time in Britain and the USA. Journal of Public Economics, 88, 1359–1386. https://doi.org/10.1016/S0047-2727(02)00168-8

Bourdieu, Pierre. 1979. La distinction. Critique sociale du jugement. Paris: Éditions de Minuit.

Bourdieu, Pierre. 1984. Questions de sociologie. Paris: Editions de Minuit.

Brickman, Philip, Coates, Dan & Janoff-Bulman, Ronnie. 1978. Lottery winners and accident victims: Is happiness relative? Journal of Personality and Social Psychology, 36, 917–927. https://doi.org/10.1037/0022-3514.36.8.917

Casal, Paula. 2007. Why sufficiency is not enough. Ethics, 117, 296–336. https://doi.org/10.1086/510692

Christman, J. 1987. Autonomy: a defense of the split-level self. The Southern Journal of Philosophy, 23, 281–293. https://doi.org/10.1111/j.2041-6962.1987.tb01623.x

Crouch, Colin. 2003. Commercialization or citizenship: the case of education. In Commercialization or Citizenship: Education Policy and the Future of Public Services. Fabian Society, London.

Csikszentmihalyi, Mihaly. 1999. If We Are So Rich, Why Aren’t We Happy? American Psychologist, 54, 821–827. https://doi.org/10.1037/0003-066X.54.10.821

Cummins, Robert A. 2000. Personal Income and Subjective Well-being: A Review. Journal of Happiness Studies, 1, 133–158. https://doi.org/10.1023/A:1010079728426

Currie, Janet & Gahvari, Firouz. 2008. Transfers in Cash and In-Kind: Theory Meets the Data. Journal of Economic Literature, 46, 333–383. https://doi.org/10.1257/jel.46.2.333

Dart, Tom. 2014. Texas Teenager Suffering “affluenza” Avoids Jail for Second Time. In The Guardian, 6 February 2014, https://www.theguardian.com/world/2014/feb/06/texas-teenager-affluenza-escapes-jail-second-time

Diener, Edward. 2000. Subjective Well-Being: The Science of Happiness and a Proposal for a National Index, American Psychologist, 55, 34–43.

Diener, Edward & Biswas-Diener, Robert. 2002. Will Money Increase Subjective Well-Being? Social Indicators Research, 57, 119–169. https://doi.org/10.1023/A:1014411319119

Diener, Edward & Oishi, Shigehiro. 2000. Money and happiness: Income and subjective well-being across nations. In Edward Diener & Suh, Eunkook M. (Eds.), Culture and subjective well-being (pp. 185–218). Cambridge: MIT Press.

Diener, Edward, Horwitz, Jeff & Emmons, Robert A. 1985. Happiness of the Very Wealthy. Social Indicators Research, 16, 263–274. https://doi.org/10.1007/BF00415126

Dworkin, Gerald. 1988. The Theory and Practice of Autonomy. Cambridge University Press.

Easterlin, Richard A. 1973. Does Money Buy Happiness? The Public Interest, 30, 3–10.

Easterlin, Richard A. 1995. Will raising the incomes of all increase the happiness of all? Journal of Economic Behavior & Organization, 27, 35–47. https://doi.org/10.1016/0167-2681(95)00003-BEasterlin 2001

Eckenroth, Danielle. 2015. Wealthy Justice: The Role Wealth Plays in Sentencing and in the Affluenza Defense Notes. New England Journal on Criminal and Civil Confinement, 41, 443–462.

Elster, Jon. 1982. Sour grapes—utilitarianism and the genesis of wants. In Amartya K. Sen & Bernard Williams (Eds.), Utilitarianism and Beyond (pp. 219–238). Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511611964.013

Festinger, Leon. 1962. Cognitive Dissonance. Scientific American, 207, 93–106. https://doi.org/10.1038/scientificamerican1062-93

Frank, Robert H. 1999. Luxury fever: why money fails to satisfy in an era of excess. New York: Free Press.

Frankfurt, Harry G. 1988. The Importance of What We Care About: Philosophical Essays. Cambridge University Press. https://doi.org/10.1017/CBO9780511818172

Frey, Bruno S. & Stutzer, Alois. 2002. What Can Economists Learn from Happiness Research? Journal of Economic Literature, 40, 402–435. https://doi.org/10.1257/002205102320161320

Fricker, Miranda. 2007. Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press.

Friedman, Marilyn A. 1986. Autonomy and the Split-Level Self. The Southern Journal of Philosophy, 24, 19–35. https://doi.org/10.1111/j.2041-6962.1986.tb00434.x

Gosseries, Alex. 2011. Qu’est-Ce Que Le Suffisantisme? Philosophiques, 38, 465–491. https://doi.org/10.7202/1007460ar

Hume, David. 1751. An Enquiry Concerning the Principles of Morals. London: Printed for A. Millar.

Khader, Serene. J. 2011. Adaptive Preferences and Women’s Empowerment. Oxford: Oxford University Press.

Kollmuss, Anja & Agyeman, Julian. 2002. Mind the Gap: Why Do People Act Environmentally and What Are the Barriers to pro-Environmental Behavior? Environmental Education Research, 8, 239–260, https://doi.org/10.1080/13504620220145401

Mackenzie, Catriona. 2014. Three Dimensions of Autonomy. In Andrea Veltman & Mark Piper (Eds.), Autonomy, Oppression, and Gender (pp. 15–41). New York: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199969104.003.0002

Maurin, Éric. 2009. La peur du déclassement: une sociologie des récessions. Paris: Seuil.

McLeod, Carolyn. 2002. Self-trust and reproductive autonomy. Cambridge: MIT Press.

Mirrlees, James A. 1971. An Exploration in the Theory of Optimum Income Taxation. The Review of Economic Studies, 38, 175–208. https://doi.org/10.2307/2296779

Monin, Benoît. 2008. Cognitive Dissonance. In W.A. Darity (Ed.), International Encyclopedia of the Social Sciences. (pp. 599–601) Farmington Hill, MA: Macmillan.

Morton, Jennifer M. 2011. Toward an Ecological Theory of the Norms of Practical Deliberation. European Journal of Philosophy, 19, 561–584. https://doi.org/10.1111/j.1468-0378.2010.00400.x

Musgrave, Richard Abel. 1959. The theory of public finance: a study in public economy. New York: McGraw-Hill.

Myers, David G. 2000. The Funds, Friends, and Faith of Happy People. The American Psychologist, 55, 56–67. https://doi.org/10.1037/0003-066X.55.1.56

Nozick, Robert. 1974. Anarchy, State and Utopia. New York: Basic Books.

Nussbaum, Martha C. 2000. Women and Human Development: The Capabilities Approach. Cambridge: Cambridge University Press.

Oshana, Marina. 2006. Personal Autonomy in Society. Aldershot, Hants, England; Burlington, VT: Ashgate.

Pinçon, Michel & Pinçon-Charlot, Monique. 2016. Sociologie de la bourgeoisie. Paris: La Découverte.

Rawls, John. 1999. A Theory of Justice, Revised Edition. Oxford: Oxford University Press.

Rawls, John. 2001. Justice as Fairness: A Restatement. Cambridge, MA: Harvard University Press.

Raz, Joseph. 1986. The Morality of Freedom. Oxford: Clarendon Press.

Robeyns, Ingrid. 2011. The Capability Approach. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/archives/win2016/entries/capability-approach/.

Robeyns, Ingrid. 2017. Having Too Much. In J. Knight & M. Schwartzberg (Eds.), NOMOS LVII: Wealth. Yearbook of the American Society for Political and Legal Philosophy (pp. 1–44). New York: New York University Press. https://doi.org/10.18574/9781479849291-003

Robeyns, Ingrid. 2018. Where do people put the riches-line [Blog post]. Retrieved from: http://crookedtimber.org/2018/01/06/where-do-people-put-the-riches-line/.

Sen, Amartya. 1985. Commodities and Capabilities. Oxford: Oxford University Press.

Slemrod, Joel. 2006. The Consequences of Taxation. Social Philosophy & Policy, 23, 73–87. https://doi.org/10.1017/S0265052506060171

Spears, Dean. 2014. Decision costs and price sensitivity: Field experimental evidence from India. Journal of Economic Behavior & Organization, 97, 169–184. https://doi.org/10.1016/j.jebo.2013.06.012

Thurow, Lester C. 1976. Government Expenditures: Cash or In-Kind Aid? Philosophy & Public Affairs, 5, 361–381.

Waldron, Jeremy. 1988. Autonomy and Perfectionism in Raz’s Morality of Freedom Symposium: The Works of Joseph Raz. Southern California Law Review, 62, 1097–1152.

Westlund, Andrea C. 2009. Rethinking Relational Autonomy. Hypatia, 24, 26–49. https://doi.org/10.1111/j.1527-2001.2009.01056.x

White, Stuart. 2010. A Modest Proposal? Basic Capital vs. Higher Education Subsidies. British Journal of Politics & International Relations, 12, 37–55. https://doi.org/10.1111/j.1467-856X.2009.00392.x