Utilitarianism Is a Form of Egalitarianism

1. Introduction

Utilitarianism holds that morality gives agents one aim: to maximize overall well-being. “Overall well-being” is the sum of the well-being of each individual, weighted equally. Utilitarians need not assume that well-being is easy to measure or commit to any particular account of it. To be a utilitarian, one simply has to think that well-being—whatever it is—ought to be maximized.

Because it assigns equal weight to each person’s well-being—following Bentham’s dictum “each to count for one, and none for more than one” (attributed to him by Mill 1861/2008: 199), there is a sense in which utilitarianism affirms the equality of persons (see also Sen 1987). In Bentham’s day utilitarianism stood in opposition to more traditional, hierarchical moral theories. In contemporary philosophy, however, utilitarianism is often contrasted with egalitarianism. Egalitarians, in the contemporary sense, not only affirm the equal worth of individuals, but care about who gets what and how we relate to one another. Utilitarianism, it seems, is uninterested in such things, so long as overall well-being is maximized.

In this paper I will argue—contrary to the received philosophical view—that utilitarianism is a form of egalitarianism. In §2, I motivate the thought that egalitarians need not value the equal distribution of well-being; however, all egalitarian theories are concerned about inequalities of wealth. In §3, I give two reasons to think that utilitarianism shares this concern, and will consistently provide reasons for egalitarian redistribution. That it recognizes other reasons does not preclude its being a form of egalitarianism, since all plausible egalitarian theories are pluralistic, as I explain in §4. Utilitarianism’s preference for wealth redistribution is contingent upon certain empirical facts. This too does not preclude utilitarianism from being a form of egalitarianism, since recent egalitarian theories, such as Martin O’Neill’s “non-intrinsic egalitarianism” (2008, hereafter “NIE”), which I introduce in §5, share this feature. In §6 I argue that utilitarianism is no more vulnerable to the charge of contingency than NIE. Like O’Neill’s theory, one reason that utilitarianism favours more equal distributions of wealth is the value it places on the sorts of social relations preferred by “relational egalitarians.” However, unlike NIE, it does not favour these relations intrinsically. In §7, I argue that this difference is of little importance. I therefore conclude in §8 that utilitarianism ought to be considered a form of egalitarianism.

It has long been argued that utilitarianism often implies egalitarian policies. One of the reasons I draw on for this claim, diminishing marginal utility, has been a staple of the tradition going back to Hume (1751/1965: 44), through Sidgwick (1998: 110), and the early twentieth-century social democrats (Jackson 2004)—and more recent thinkers such as Hare (1981: 9), Sen (1987), and Broome (1995: ch. 9). The other reason, the adverse effects that inequality itself has on well-being, is recognized and incorporated into kinds of utilitarianism, in different ways, by Hare (1981: §9.8) and Broome (1995: 198). The contributions of this paper are to bring together these points in the light of recent developments in both empirical social science and political philosophy. Empirical results give further evidence for utilitarianism’s egalitarian credentials with respect to both of the reasons that philosophers previously hypothesized. Meanwhile, the philosophical literature on “relational egalitarianism” (Anderson 1999; Scheffler 2003; Wolff 1998) has revised our understanding of the value of equality. Drawing upon the two, I argue not only that utilitarianism shares some of the policy implications of egalitarian theories, but that it ought to be considered a form of egalitarianism itself.

This is not merely a matter of terminology. In popular political discourse, utilitarian arguments are often raised against egalitarian demands. In philosophy, there are people who reject utilitarianism on the grounds that it cannot account for the value of equality. Bernard Williams writes that utilitarianism is “indifferent to issues of justice or equity” (1973: 137), and the two most influential liberal political philosophers of the twentieth century, Rawls and Dworkin, agree that utilitarianism is flawed in this regard (Dworkin 2000: 328–330; Rawls 1971: 22–33).¹

In recent times, this view has pushed several philosophers who are relatively sympathetic to utilitarianism to revise the theory to accommodate distributive concerns—for instance, we have the priority views of Parfit (1991) and Buchak (2017); the sufficiency views of Crisp (2003) and Frankfurt (1987); and Ashford’s non-aggregative individualist utilitarianism (2005). Meanwhile, contemporary utilitarians in the effective altruist movement often neglect the importance of egalitarian concerns, devoting their energies instead to questions of population ethics, existential risk and individual altruistic duties.²

If I am right, utilitarianism should not be rejected for failing to be egalitarian, but should instead be engaged with as an egalitarian theory—and utilitarians should take equality seriously.

2. The Currency of Egalitarianism

It might be thought that the claim that utilitarianism is egalitarian is a non-starter, because utilitarianism values only the sum and not the distribution of well-being. According to utilitarianism, a given amount of well-being is as good whether it accrues to a billionaire or a maid. Moving units of well-being from some persons to others does not change the amount of well-being there is, hence (it might be thought) purely distributive differences between two social arrangements cannot make for differences in overall well-being, and therefore utilitarianism is indifferent between distributions.

It is true that for fixed amounts of overall well-being, utilitarians do not care how that well-being is distributed.³ There are people—”welfare egalitarians”—who take this to be fatal to my thesis that utilitarianism is a form of egalitarianism. They believe that well-being is “the currency of egalitarianism”; that is to say, that egalitarianism favours more equal distributions of well-being.⁴ Thus, it is a necessary condition for a theory’s being a form of egalitarianism that it prefers more equal distributions of well-being. I disagree. The following example shows that the distribution of well-being is not what we care about when we care about equality.

WEALTHY MAN: A wealthy man, respected by all and with great power over his employees, suffers from regular migraines. This severely affects his well-being: aside from the pain itself, he becomes unable to focus on his work and his bad mood strains his close relationships. His life is now worse (in terms of well-being) than that of the woman who cleans his office: although she has one-thousandth of his wealth and her opportunities are restricted by racial and gender discrimination, she is in good health, has a happy and loving family and spends her small amount of free time pursuing an inexpensive hobby which gives her a lot of joy.

If we care about the equal distribution of well-being, then we should think that it would be in some way good to redistribute well-being from her to him. Since money and leisure time make a positive contribution to well-being for both of them, it should be an improvement, from the point of view of equality, for her wages to be reduced and for him to keep the difference, or for her to work unpaid overtime doing more chores for him. Now, I take it that most of us find such a policy unattractive. That we find it unattractive does not mean that we do not care about equality—in fact, it seems to be our concern for equality, in part, that such a policy offends. This shows that insofar as we care about equality, we care about something other than the equal distribution of well-being. In this we seem to agree with egalitarian movements in real-world politics. Feminists, socialists and anti-racists take themselves to be engaged in a struggle for equality. But they are unlikely to endorse redistribution towards the boss. A boss’s headache, however painful, is not a pressing matter for egalitarians, and greater exploitation of the poor by the rich is not something they should favour.⁵

This argument is not a conclusive rejection of welfare egalitarianism: there are, I’m sure, further things that a welfare egalitarian could say in defence of equalising well-being, especially in combination with other principles. However, it does suggest that welfare egalitarianism fails to capture (all of) the spirit of egalitarianism: thus, to impose as a necessary condition for a theory’s being a form of egalitarianism that it prefers more equal distributions of well-being seems too restrictive. Utilitarianism should therefore not be ruled out by such a condition.

What does unite all egalitarian theories? One thing all egalitarian theories are concerned about is the distribution of wealth. Some take the currency of egalitarian justice to be wealth itself (O’Neill 2008; 2017), such that justice requires us to have equal amounts of wealth. I am not making this claim. Rather, I am claiming that all egalitarians have a preference for more equal distributions of wealth—possibly instrumentally rather than intrinsically, and typically, rather than in every conceivable case. This goes for those theorists who take the currency of egalitarian justice to be bundles (such as resources or primary goods) of which wealth is part (Dworkin 1981; Rawls 1971). It also goes for those who take well-being, freedom (van Parijs 1997), capabilities (Sen 1987), or “access to advantage” (Cohen 1989) to be the currency of egalitarian justice, since inequalities in wealth tend to cause inequalities in these things. It goes, crucially, for real-world political movements who take equality to be their goal. I will argue that it also goes for utilitarianism.

Some egalitarians deny that there is anything that justice requires us to have equal amounts of. “Relational egalitarians” (Anderson 1999; Scheffler 2003; Wolff 1998) point out that many of the things condemned and advocated on egalitarian grounds are not matters of distribution: consider prejudice, exclusion and disrespect. These do not concern what we have, but how we relate to one another. As Elizabeth Anderson puts it: egalitarianism’s “proper positive aim is not to ensure that everyone gets what they morally deserve, but to create a community in which people stand in relations of equality to others.” (1999: 288–289) We will revisit such theories later in this paper. For now, let it be noted that relational egalitarians should also be concerned with the distribution of wealth. How wealth is distributed affects the relations in which people stand to one another—consider how relations between workers and bosses (or men and women, or white and black people) would change if they were equally wealthy.⁶

Thus, every egalitarian philosophy worthy of the name is concerned—either directly or derivatively—with the distribution of wealth. In contrast to the concern for the distribution of well-being, this concern matches our egalitarian intuitions about WEALTHY MAN. Wealth should be redistributed, for the sake of equality, to the cleaner (who has less of it) from the boss (who has more of it), and not vice versa. A focus on the distribution of wealth also accords with egalitarianism in real politics. The French Revolution’s call for “égalité” was a demand for—amongst other things—a redistribution of the wealth of the aristocracy and church.⁷ Socialists characteristically favour redistributive taxation and the collectivisation of property. Movements for racial and gender equality draw attention to wealth gaps between whites and people of colour, and between men and women.

This suggests that it is a necessary condition for a theory to be a form of egalitarianism that it recognizes reasons to favour more equal distributions of wealth. (I do not claim it is a sufficient condition, or, to repeat, that wealth is “the currency of egalitarian justice.”) In the next two sections I will show that utilitarianism meets this condition.

3. Diminishing Marginal Utility and Relative Wealth

We might think that utilitarianism is unconcerned with the distribution of wealth—as with well-being, utilitarians care about the total rather than the distribution. Just as a single person (other things equal) is better off the more wealth she has, perhaps utilitarianism will hold that a society is better the more wealth it has, however it is distributed.

This would be so only if total well-being were directly proportional to total wealth. Utilitarianism would then hold that the just distribution of wealth is one in which the most wealth exists. However, such proportionality does not hold. Different distributions of the same sum of wealth generate different sums of well-being. Thus, utilitarians do care about who wealth accrues to—in contrast to their indifference about the distribution of well-being. Moreover, there are at least two good reasons for utilitarians to prefer more equal distributions of wealth. The first is the diminishing marginal utility (DMU) of wealth. The second is the effect on well-being of relative wealth and poverty.

Something has DMU if (and only if) an additional unit of it would produce more well-being in the hands of someone with fewer units of it to start with, than it would in the hands of someone with more. If wealth has DMU, then utilitarians will care to whom wealth accrues: they will prefer it to accrue to those who have less of it, because it contributes more to their well-being.

Does wealth have DMU? It is intuitively highly plausible, especially at the extremities of the global wealth distribution. Imagine that the total value of Jeff Bezos’s Amazon stock increases by $500. His well-being would hardly, if at all, be altered; he very probably would not even notice. Imagine that a professor receives a $500 bonus. She might fly business class rather than economy to her next conference or stay in a fancier hotel when she arrives. These things would probably make her life a little better. Imagine that a Bangladeshi factory worker receives $500. This would roughly double their annual income. They would be able to put another child through school, eat more nutritious food, pay off the medical debt that has been haunting them or fix the leak in their roof. Their life will be made far better.

Empirical studies typically confirm that wealth has DMU within countries, and throughout the wealth distribution. Wealth (or, as is more typically measured in such studies, income) has a positive but diminishing relationship with “subjective well-being”—that is, people’s self-reports of their well-being (Clark et al. 2018: 2; Layard, Mayraz, & Nickell 2008: 1856)—and with health outcomes (Mackenbach et al. 2005: 287; Bueno 2011; Mortensen et al. 2016; Ecob & Davey Smith 1999). I do not mean to take these metrics as definitional of well-being, and therefore these studies are suggestive rather than conclusive. People may be mistaken about their own well-being, and good health is not the whole of well-being. However, on any plausible account of well-being self-reports will be at least a rough indicator of it, and health will correlate positively with it.

There is good reason, then, to believe that wealth has DMU. This gives utilitarians a reason to prefer wealth to accrue to the poor than to the rich—that is, to prefer more equal distributions.

The point about DMU is well-known. It is found in many classic utilitarian texts (Hume 1751/1965: 44; Sidgwick 1998: 110). However, marshalling the empirical evidence remains important. This evidence can inform us of the rate at which the marginal utility declines; it also helps to rebut arguments such as Harry Frankfurt’s (1987: 27), which posits “utility thresholds,” or points in the scale of wealth at which an extra dollar pushes one over the threshold and thereby increases one’s well-being to a greater extent than it would were one lower down the scale.⁸ Next I want to make a more novel argument to a similar, but more strongly egalitarian, conclusion. This stems from observations about the well-being effects of relative wealth and poverty.

Hare observes that a human tendency to envy those with more will tend to reduce the welfare of those who have less, relatively speaking (1981: §9.8). Thus, utilitarians have reasons for reducing inequalities. Broome reaches a similar conclusion through noting that “unfairness … is plainly a personal harm” (1995: 198). I argue that differences in relative wealth have adverse effects on well-being without appeal to envy or fairness, but rather through empirical social scientific findings about well-being, health and the “psychosocial” effects of inequality.

Many studies (Blanchflower & Oswald 2004; Clark et al. 2018: ch. 2; Clark, Frijters, & Shields 2008; Layard, Mayraz, & Nickell 2010) suggest a positive relationship between subjective well-being and relative wealth—that is, not simply how many dollars you have, but how many dollars you have relative to others in your society (either to the mean, median, your neighbours [Luttmer 2005], or the wealthiest group). The more unequally wealth is distributed in a society, the greater the relative wealth of the rich, and the greater the relative poverty of the poor. The differential importance of relative versus absolute wealth is disputed, though it is clear that relative wealth is important.⁹ Empirical work on health also suggests a connection with relative wealth. There is a body of evidence that most aspects of health follow a “social gradient”: that is, they are worse for poorer people than for richer people (Marmot 2015). And there is good reason to think this is not simply because of absolute poverty, but because of relative poverty (the inverse of relative wealth). Absolute poverty could explain why those with secure jobs and housing, who can afford medical care, education and three meals a day have better health and social outcomes than those without those things. But the social gradient for these problems runs all the way through society—it is not only that the poor suffer them more than the rich, but that the rich suffer more than the super-rich. Absolute wealth cannot plausibly account for differences between individuals who all have enough of it to meet these basic needs. As Marmot asks: “Why should educated people with good stable jobs have a higher risk of dropping dead than people with a bit more education or slightly higher-status jobs? Is living in a five-bedroom house with three bathrooms better for your health than “crowding” the spouse and two children into a four-bedroom house with only two?” (2015: 1–2).

The fact that the relatively poor fare worse than the relatively rich does not, by itself, give utilitarians a reason to condemn wealth inequality. It could be that by closing the gap between the two, one makes the relatively poor better off but the rich worse off to a similar or greater extent, thereby doing nothing positive for the overall sum of well-being. What is needed to provide utilitarian reasons for wealth redistribution is that relative poverty is worse for overall well-being than the relative wealth that is its corollary.

There is some evidence that relative poverty reduces the subjective well-being of the poor more than relative wealth increases the subjective well-being of the rich (Blanchflower & Oswald 2004; Clark et al. 2008). This is also consistent with the finding (Alesina, Di Tella, & MacCulloch 2004; Oishi, Kesebir, & Diener 2011) that the subjective well-being of poorer people is more adversely affected by income inequality. Looking at outcomes other than subjective well-being reports also gives us reasons to think that the adverse effects of relative poverty outweigh any benefits with respect to well-being. For instance, Richard Wilkinson and Kate Pickett find that “almost all problems which are more common at the bottom of the social ladder are more common in more unequal societies” (2009: 18). In less equal societies, there is more relative poverty than in a more equal society: the poor are relatively poorer and the rich relatively richer. The social problems Wilkinson and Pickett discuss—violence, incarceration, low educational attainment, poor mental and physical health, early pregnancies and drug abuse—all plausibly adversely affect well-being, on most reasonable conceptions of well-being. The fact that they are more common in more unequal societies, that is, societies with higher rates of relative poverty and relative wealth, suggest that the harm done by the former outweighs the good of the latter.

What is the mechanism by which relative poverty affects subjective well-being and gives rise to health and social problems? Wilkinson and Pickett, along with Marmot, appeal to “psychosocial” mechanisms: relative poverty affects our social relations, which affect us psychologically and thus affect our health and well-being. The social relations in question are those with which relational egalitarians concern themselves: exploitation, domination, hierarchy, marginalisation and prejudice. Relative poverty puts one at risk of exploitation and domination by the relatively wealthy. This reduces the control one has over one’s life, causing greater levels of stress, which Marmot (2015) shows is a major factor in many health problems—as well as being a source of pain in itself. Furthermore, economic inequalities are sustained by (and sustain) structures that grant some individuals higher status than others. As inequalities grow, so will the felt significance of such status differences (Wilkinson & Pickett 2009: 44). If you are relatively poor, such expressions will adversely affect your self-esteem, and provide another source of stress—status anxiety. And relative poverty also alienates one from the relatively wealthy: they lead such different lives that cross-class friendships become difficult and cultures diverge. Since the relatively wealthy are likely to dominate cultural production, the relatively poor are likely to find themselves erased, marginalized or stereotyped in mainstream discourses.¹⁰

Wilkinson and Pickett argue that from what relative poverty does to our social relations, various health and other problems follow; as we have seen, these problems seem likely to reduce well-being, on any plausible account of it. But note that even if those problems are not as closely associated with relative wealth and poverty as Wilkinson and Pickett suggest, we might still think that exploitation, marginalisation and so on make life worse, whatever their further effects on health, social problems, or subjective well-being. Utilitarians need not be committed to the hedonistic view that an individual’s well-being is determined by their pleasures and pains alone. They may also think that it is affected by their place in social relations. If relative poverty leads to people standing in bad social relations, therefore, that itself would be reason for utilitarians with such an axiology to be concerned about relative poverty and thus prefer more equal distributions of wealth.

As with relative wealth and poverty themselves, one might ask with respect to these relations, although they make life worse for those on one end of them, do they not make life similarly better for those on the other end? If so, utilitarians could not condemn them. There are several responses to this. Firstly, it could be that the dominators and exploiters, whilst they are happier than those they dominate and exploit, are not happier than they could be in a more egalitarian society. The rich are not immune to status anxiety and alienation from their fellow humans, for instance, whilst in a more egalitarian society these things would be mitigated for everyone. Secondly, for some effects of relative poverty, there will always be more suffering these effects than could be on the other end: there are more of us being exploited and dominated than there could be exploiters and dominators, by the nature of these relations.¹¹

So it seems that relative poverty is likely to adversely affect overall well-being, on many conceptions of well-being: it damages the subjective well-being and health of the poor; it gives rise to inegalitarian social relations; and these effects are likely to cause problems that are not outweighed, in terms of well-being, by their advantages to the wealthy. The negative well-being effects of relative poverty provide further reason for utilitarians to prefer more equal distributions of wealth.

Considerations of relative wealth and poverty open the utilitarian to more strongly egalitarian possibilities than DMU alone does. DMU suggests that redistributing a given amount of wealth from rich to poor will increase overall well-being. DMU also implies that in some cases egalitarian redistribution increases overall well-being even when it decreases the overall sum of wealth. But it could not justify redistribution that made everyone poorer.

However, if wealth has DMU and relative poverty matters to well-being in the way I have suggested, then egalitarian redistribution will increase overall well-being in some such cases. Consider the case of an unequal distribution. We make the rich a lot poorer, and the poor a little poorer. Making the rich a lot poorer may be bad for the rich, and making the poor a little poorer is bad for the poor. But making the rich a lot poorer is good for the poor, because it increases their relative wealth. By doing so it may also improve the quality of social relations and reduce the prevalence of social problems. These effects could make for an increase in well-being, even though the redistributive policy has made every person poorer in terms of wealth. As Goodin puts it: “If your stock and use of resources actually impinges upon others’ enjoyment of theirs, then it is easy enough to see how restrictions on your stock and use of resources might enhance overall social enjoyment – what you lose through such restrictions may be more than compensated by what others gain through them” (1995: 250).¹²

In this regard, utilitarianism may mandate more equal distributions than other principles which might be thought to be either forms of egalitarianism or at least more egalitarian than utilitarianism is. Firstly, take a maximin principle of wealth, which holds that distributions ought to maximize the lot of the worst off.¹³ Secondly, take a principle that says wealth always matters, but matters more the poorer the person to whom it accrues, such that there is a positive but diminishing marginal moral value to goods.¹⁴ These principles will always license inequalities that make everyone wealthier; utilitarianism, as we have just seen, will not.

Far from being indifferent to the distribution of wealth, then, DMU and relative poverty mean that utilitarians have reasons to prefer more equal distributions of wealth. Unlike some other theories, it can provide such reasons even where this reduces the total stock of wealth, and where it makes everyone poorer.¹⁵

4. Pluralism

I have shown that the effects of DMU and relative wealth give utilitarians reasons to favour more equal distributions. But there may be countervailing reasons: reasons for utilitarians to favour inequality. It could be that permitting a degree of wealth inequality increases the total wealth of a society, for example, by providing incentives for individuals to work hard or take risks. In this case, given that overall well-being tends to increase with total wealth, that increase could in theory be enough to outweigh the negative effects of inequality on well-being mentioned in the previous section, and utilitarians may prefer to permit such inequality.

This is just to say, though, that utilitarians will prefer some less equal societies to some more equal ones. And this is true of all plausible egalitarian theories. Rawls’s Difference Principle, often referred to as part of the “liberal egalitarian” view, endorses inequalities that improve the conditions of the worst-off (Rawls cites the incentive effects of inequality as an example). All plausible egalitarian theories are pluralistic: they place some moral weight on distributive equality, but endorse other values as well. Thus, where a more equal society instantiates less of those other values, such egalitarians may prefer the more unequal option. The alternative would be to prefer societies that were perfectly equal to those which were at all unequal, even where the former were places of much greater suffering, rights violations, sin, and so on.

Plausible egalitarian theories recognize that there are reasons to prefer more equal societies, and that these reasons are sometimes outweighed. Utilitarianism does exactly this. In the previous section I showed why utilitarians should believe there are reasons to prefer more equal societies. That these may sometimes be outweighed, for the utilitarian, does not preclude utilitarianism from being an egalitarian theory, but puts it in the same boat as plausible egalitarianisms. The fact that all other plausible egalitarian theories are pluralistic will be returned to below.

One might think that this is a sleight-of-hand on my part. I seem to have established that sometimes more equal distributions produce more well-being, and when they do utilitarianism favours them. But couldn’t this be reversed? Sometimes less equal distributions produce more well-being, and when they do utilitarianism favours them. There is, I think, a justification for placing utilitarianism on the egalitarian side of the line. The mechanisms by which more equal distributions lead to greater well-being are not accidental; they stem from very general truths about the nature of people and well-being. The most important aspect of people’s well-being is that their basic needs are met—thus, wealth has diminishing marginal utility. People are better off engaged in society as equals—thus, relative wealth matters. Sometimes these tendencies are outweighed. But they are deep, general and permanent.

The most important way in which these tendencies are outweighed is that people will produce more of what is conducive to well-being if they are incentivized by the prospect of unequal rewards. I doubt that this claim holds as generally as the claims about DMU. People can and do work inspired by other motives—look at the vast volunteer-produced Wikipedia, for instance, or (less optimistically) those who kill and die out of patriotic duty. Of course, unequal rewards can still add incentives alongside these other motives. But importantly we can imagine societies in which such motives were more widespread, so that sufficient productive work was undertaken without the need for unequal incentives.¹⁶ It is harder to conceive of people whose welfare has a convex relationship to wealth, so that moving from malnourishment to three meals a day did less for them than moving from three meals a day to champagne and caviar.

Such a claim is difficult to prove, and is ultimately amenable to empirical evidence as well as philosophical intuition. Perhaps human selfishness is as deep-rooted as the importance of basic needs to human welfare. However, there is a further thing the utilitarian can say for treating the latter as the more profound fact: motivation by unequal incentives is suboptimal. This is because even in a society in which unequal incentives are necessary to motivate people to work, invest and innovate, DMU and the psychosocial effects of relative wealth still operate, so any mechanism that increases inequality has downsides for overall well-being. By contrast, the society imagined by John Stuart Mill, in which motivation is supplied not by the promise of unequal reward, but by “competition, who can do the most for the common good” (1848/1965: 205), would, if possible, be superior from a utilitarian perspective.¹⁷ Some place for unequal incentives may be unavoidable. But there is a general utilitarian argument to replace such motives, where possible, with more altruistic ones, given the effects of inequality mentioned above.

The reasons pushing utilitarianism towards equality, then, are grounded in general facts about people, whilst the reasons pushing in the other direction are grounded in facts that may not be so deep, but, even if they are, are facts that utilitarians have reasons to regret, and to reduce in scope. Thus it is a better description of utilitarianism to say that it consistently gives reasons for more equal distributions, though these are sometimes outweighed, than the inverse.

5. Non-Intrinsic Egalitarianism

Those who accept the claims of this paper so far, but still wish to draw a contrast between utilitarianism and egalitarian theories often appeal to the fact that utilitarianism’s reasons to prefer equality are instrumental—they are had in pursuit of another value, overall well-being—whilst egalitarians value equal distributions in themselves, or “intrinsically” (see, for instance, Temkin 2001). However, there are egalitarian theories that are similar to utilitarianism in this regard.

As noted in §2, many egalitarian theories do not take wealth to be the ultimate currency of egalitarianism. They might take capabilities, resources, access to advantage, freedom or opportunity to the thing that of which individuals should have equal amounts. As I argued there, these theories will still typically prefer more equal distributions of wealth, as wealth inequality tends to cause inequalities in those other things. These theories, therefore, only value equal distributions of wealth instrumentally. However, they do value the equal distribution of something intrinsically, namely, whatever thing they take to be the currency.

In contrast, relational egalitarian theories do not do this. They take the fundamental value of equality to be found not in distributions but in certain social relations. As I noted, they will also typically prefer more equal distributions of wealth, since inequalities in wealth tend to cause the social relations they dislike. (Further support for this latter claim is found in the previous section’s discussion of the effects of wealth inequality on the psychosocial.) To relational egalitarians, equal distributions of wealth (or anything) are instrumental to the value of better social relations, just as to utilitarians they are instrumental to the maximisation of overall well-being.

The form of egalitarianism advocated by Martin O’Neill (2008), which he calls “non-intrinsic egalitarianism” (NIE), can be interpreted as a clear spelling out of relational egalitarianism’s implications for the distribution of wealth. In the next three sections I show that utilitarianism is just as egalitarian a theory as NIE. The remainder of this paper compares these two theories, revealing more about what it takes for a theory to be a form of egalitarianism, and ultimately supporting the view that utilitarianism has what it takes.

For O’Neill, unequal distributions of wealth make for worse states of affairs. This is explained by the consequences that usually arise from inequality. (O’Neill uses “equality” and its cognates to refer to the distribution of wealth, as I will from here, unless otherwise stated.) These consequences are (2008: 121–123):

• (a)

  the avoidable frustration of basic needs,

• (b)

  status hierarchies,

• (c)

  the domination of one group by another,

• (d)

  weakened self-respect of the poor,

• (e)

  servility and deferential behaviour,

• (f)

  the undermining of healthy fraternal social relations and attitudes.

NIE holds that inequality is bad because it leads to (some subset of) consequences (b)–(f) (2008: 123)—which are examples of the social relations to which relational egalitarians are averse. (O’Neill also believes that inequality is bad because it leads to (a), but he does not believe this qua non-intrinsic egalitarian, but rather as an independently plausible and not distinctively egalitarian claim.) If inequality occurred without leading to such consequences, NIE would find no fault with it. NIE, then, endorses equality conditional upon the empirical claim that inequality has these consequences.

Utilitarianism also holds that inequality is bad insofar as it has certain consequences, namely:

• (g)

  less overall well-being than there could otherwise be.

If inequality occurred without leading to (g), utilitarianism would find no fault with it.¹⁸ Utilitarianism’s commitment to equal distributions, then, depends on an empirical claim that distributive inequality has this consequence. In the previous two sections I showed that this empirical claim is well-founded: inequality usually will lead to (g).

So, utilitarianism seems to have the same structure as NIE with regard to distributive inequality: such inequality is bad when and because it leads to a certain consequence. Furthermore, the consequences are not entirely independent. O’Neill’s (b)–(f) are effectively the concerns whose psychosocial effects form a large part of the reason to think that inequality tends to reduce overall well-being (see §3). Therefore, both NIE and utilitarianism explain the badness of distributive inequality at least in part through relational egalitarian concerns.

Despite these similarities, O’Neill argues that utilitarianism is only “weakly” egalitarian, in contrast to NIE, which is “strongly” egalitarian (2008: 125). In the next two sections, I will show that his argument for this conclusion fails: utilitarianism is as egalitarian a theory as NIE.

6. Contingency and Reliability

Like utilitarianism, NIE endorses distributive equality for the sake of other goals, and conditional upon contingent empirical claims. This does not, according to O’Neill, make it any less strongly egalitarian. He writes:

the connection between distributive egalitarianism and these broader egalitarian goals and values [i.e. (b)–(f)], although in some sense contingent, is not a weak one. If this connection were weak, then we might … question whether Non-Intrinsic egalitarian views are strongly egalitarian at all. Yet it seems plausible to think that it is a deep social fact that we can realize the values embedded in the egalitarian considerations (b)–(f) only where substantial inequalities of condition have been eliminated … . If this “deep social fact” really does obtain, then Non-Intrinsic egalitarianism will reliably mandate the elimination of inequalities of condition … . (2008: 131)

Utilitarianism, according to O’Neill, is only weakly egalitarian. Therefore, we might expect that the connection between distributive inequality and (g) is weaker than that between distributive inequality and (b)-(f); that the social fact that equality correlates with overall well-being is less “deep” than the fact that it correlates with improvements with respect to (b)–(f). Is this true? O’Neill explains the depth of his deep social fact thus:

firstly, reductions in inequality almost always bring about improvements in states of affairs of the sort favoured under considerations (b)–(f); secondly, such improvements are generally possible only where inequalities are reduced, and greater distributive equality is achieved. (2008: 131)

Now, if O’Neill’s claims here are true, there is nothing to stop a utilitarian agreeing with them. Furthermore, since (b)–(f) are effectively those concerns whose psychosocial effects are inimical to well-being, improving states of affairs with respect to (b)–(f) is one way to improve them with respect to (g). The utilitarian can then draw this conclusion from O’Neill’s claims: there is a way in which things can be improved with respect to (g) (the increase of well-being via improvements with respect to (b)–(f)) which (1) almost always follows from the reduction of inequality, and (2) is generally possible only where inequalities are reduced.

It might seem that NIE is nevertheless more egalitarian than utilitarianism. For NIE’s concerns seem to track distributive inequalities more closely: NIE does not merely say that there is a way to improve things that necessitates reducing inequalities, but that this is the way to improve things, since well-being could be increased in many other ways but the values encoded in (b)–(f) could not. But this appearance neglects the fact that NIE is a pluralistic theory. As O’Neill recognizes (2008: 143–144), it is not plausible to hold that only (b)–(f) matter. NIE, fully spelled out, will have to admit that it is good to (for instance) cure a rich man’s headache, although this has no effect on (b)–(f). Thus NIE’s adherents also think that there is a way in which things can be improved (improvements with respect to (b)–(f)) which usually follows from and is generally only possible where inequalities are reduced. For both them and utilitarians this is just one way amongst others in which things can be improved.

In fact, I do not think O’Neill’s claim that only the reduction of distributive inequality can improve things with respect to (b)–(f) is true. There is very likely a strong correlation between distributive inequality and (b)–(f). But we can conceive of ways in which improvements could be made with respect to (b)–(f) without the distribution of wealth being touched. Servility and deference could be reduced by changes to norms and structures which could occur whilst wealth distribution is unchanged. O’Neill himself writes, in a review of their 2009 book The Spirit Level, that Wilkinson and Pickett’s

emphasis… is perhaps excessively on the way in which these policies could lead to greater income equality through compressing wage differentials. An alternative approach that de-emphasized income inequalities to some degree could give more emphasis to the ways in which various forms of economic democracy can act in more direct ways to bolster individual status and self-respect, and to transform the character of social relations at work, through granting more power to individual workers. Thus… their approach prevents them from seeing the full range of egalitarian arguments that might be deployed in favour of such policies. (O’Neill 2010: 405, italics his)

The implication is that measures such as workplace democracy could improve things with respect to egalitarian criteria, though they may not involve any redistribution of wealth or income. But then, he should not worry that the possibility of such cases would undermine the claim of NIE to be a kind of egalitarianism. It seems that his view is that egalitarians should want to eliminate or reduce (b)–(f), and this gives them reason to eliminate or reduce distributive inequalities. The fact that it could give them reason to promote non-redistributive policies, such as workplace democracy, does not threaten the egalitarianism of the view. Utilitarianism has the same attitude to such cases. It can endorse improvements with respect to (b)–(f), which will almost always lead to increases in overall well-being, whether they are associated with changes in the distribution of wealth or not. If this doesn’t threaten the egalitarian credentials of O’Neill’s view, then it doesn’t threaten the egalitarian credentials of utilitarianism either.

These cases aside, it seems plausible that one good way of improving things with respect to (b)–(f) will be closing inequalities of wealth. This is the grounds of O’Neill’s claim that NIE will “reliably mandate” such policies. Will utilitarianism be less reliable on this score?

There will be circumstances in which utilitarianism does not endorse closing inequalities even when doing so would lead to improvements with respect to (b)–(f). Utilitarianism would fail to endorse such policies in cases where overall well-being decreases as equality increases (that is, “levelling down” cases). We might think that cases of improvements with respect to (b)–(f) without improvements with respect to overall well-being are rare. Anyway, in such cases, inequalities typically lead to other values which pluralist non-intrinsic egalitarians ought to be concerned with.¹⁹ Take, for example, a country that is rapidly industrialising its economy. It may be necessary for this transformation that wealth is unequally distributed. If this transformation satisfies the utilitarian, it does not lead to (g)—that is, it produces the greatest possible sum of well-being. If the pluralist non-intrinsic egalitarian values the sum of well-being as well as (b)–(f), they may ultimately endorse this inequality. And even if they don’t value the sum of well-being, they could still endorse it. For if the greatest possible sum of well-being is produced, this is because other values are met, such as: suffering is reduced; and/or people enjoy greater freedom; and/or new technologies are invented; and/or needs are better satisfied, and so on. A plausible pluralistic NIE will value at least some of these things alongside concern for (b)–(f). In that case, if the value of these things is great enough, non-intrinsic egalitarians could endorse inequality in this case—even whilst continuing to believe that (b)–(f) are bad. To know how reliably a pluralistic non-intrinsic egalitarian will mandate reducing inequalities, one must know their other values, and their method of resolving conflicts between them. Without knowing this, we can just say that both utilitarians and non-intrinsic egalitarians will agree that reducing inequality is often (but not always) all-things-considered good, and that this is largely because of the considerations about (b)–(f).

There are also circumstances in which utilitarianism will mandate the closing of inequalities and NIE will not. This is obvious for some versions of pluralistic NIE: consider a view that favoured distributive equality on the grounds of (b)–(f), but also respected Nozickean entitlements to private property, making redistributive taxation impermissible. In a society with large inequalities and in which most resources were held privately, this form of pluralist NIE would not mandate very much reduction of distributive inequalities; utilitarianism is likely to mandate more. The general point, again, is that there is no reason to think that utilitarianism will be less reliable in mandating equality than a pluralistic egalitarian theory, and all egalitarian theories must be pluralistic to be plausible.

Even if NIE were not pluralistic, there are cases in which utilitarianism would mandate reductions in inequality that such “pure” NIE would not. Consider a version of the Divided World case introduced by Parfit (1991: 6). In this version of the case, there are two groups of people who have no interactions and are unaware of each other’s existence, and one group is much wealthier than the other.²⁰ As O’Neill says, “none of the egalitarian reasons covered by (b)–(f) can obtain” in such cases, so pure NIE does not mandate closing the inequality (2008: 136). But utilitarianism could. Given diminishing marginal utility, (g) would probably obtain: if the better-off group had less and the worse-off group more, overall well-being could be greater.

It seems, then, that the social fact linking distributive inequality to overall well-being is no less deep than that linking it to O’Neill’s concerns for (b)–(f). Furthermore, whilst there could be cases in which NIE mandates more equal distributions than utilitarianism, they may be rare, particularly if NIE takes a plausible pluralistic form. There are also cases in which utilitarianism advocates more equal distributions than pluralist and pure NIE.

7. Distinctiveness and Intrinsicness

O’Neill claims that concern for consequences (b)–(f) is “distinctively egalitarian” (2008: 124–126). What makes these concerns distinctively egalitarian, for O’Neill, is that they have to do with life in a “society of equals.” If status hierarchies or servile behaviour suddenly arose in such a society, it would cease to be a society of equals. In contrast, if every individual there became afflicted by chronic pain, then it would be worse with respect to (g), but no less a society of equals. So, (g) is not a distinctively egalitarian concern. Does this provide a reason to judge that utilitarianism is not a form of egalitarianism?

No, because utilitarians also have concerns for (b)–(f). As argued above, improvements with respect to (b)–(f) will tend to lead to increases in overall well-being. So, utilitarianism is concerned with (b)–(f), and if these concerns are distinctively egalitarian, then utilitarianism has distinctively egalitarian concerns. For sure, utilitarians care about things other than (b)–(f)—for example chronic pain, existential risk, and overall well-being itself (i.e. (g))—that are not distinctively egalitarian. But as O’Neill believes that egalitarians should be pluralists, he does not think that having non-egalitarian concerns compromises one’s egalitarianism.

Perhaps, however, O’Neill would argue not that utilitarianism does not share these distinctively egalitarian concerns, but rather that it does not have these concerns in a distinctively egalitarian way. O’Neill—despite the name he gives his theory—believes in the intrinsic badness of consequences (b)–(f) (2008: 130). Utilitarians do not: they believe that, like all consequences, (b)–(f) are bad only insofar as they lead to less overall well-being than there could otherwise be.²¹

This difference is real, but it is not important. The main significance of considering some consequences to be intrinsically bad, rather than bad because they fail to maximize well-being, is that one can condemn states of affairs that instantiate those consequences even when they maximize well-being. For the difference between NIE and utilitarianism to be important, then, there must be cases in which inequalities lead to (b)–(f) but not to (g). These are the cases such as the industrialising country mentioned above. But, as we saw there, given their pluralism, NIE is not necessarily committed to more egalitarian policies here than utilitarianism—and there are also cases in which utilitarians will condemn inequalities that non-intrinsic egalitarians, pluralistic or pure, will not.

One thing that might be said for the stronger egalitarian credentials of NIE is that in cases of levelling down, even where both utilitarianism and (pluralistic) NIE give the same answer, the latter always places some disvalue on (b)–(f) whilst the former does not.²² Thus, O’Neill can hold that it is always regrettable that (b)–(f) occur, even in levelling down cases where they ultimately, on the grounds of other values, endorse states of affairs where (b)–(f) occur.

It might be claimed that utilitarians, in contrast, could not regret the occurrence of (b)–(f) when they form part of the state of affairs that maximizes well-being. Surely utilitarianism cares only about maximising well-being—how could utilitarians see any part of an optimific state of affairs as regrettable?

The claim that utilitarians cannot regret any aspect of an optimific state of affairs would, if true, cause a problem for utilitarianism across a wide range of cases. Consider the trolley problem (Foot 1967). Utilitarianism (with common sense) suggests switching the trolley to save five lives at the cost of one. The death of the one is surely seen as something to regret—as is typically expressed by utilitarians talking of it as “a cost,” or something to be “traded off.” But the death of the one is part of the outcome that maximizes well-being. So, if utilitarians cannot regret any aspect of an optimific state of affairs, they cannot regret the death of the one.

I propose that the utilitarian can respond to this worry as follows. One ought to bring about the inequality and the death of the one, in the circumstances described, since doing so would maximize well-being. However, it would be better if the circumstances were not as described. It would be better, in particular, if one could reap the rewards of industrialisation, or save the five, without increasing inequalities or killing the one, respectively. Utilitarianism implies that these circumstances would be better than those in the original cases because more well-being could be realized in them: other things equal, saving five people without killing the one or industrialising an economy without increasing inequalities would increase well-being more than saving the five and killing the one or industrialising and increasing inequalities. This is what makes killing the one or increasing inequalities “costs”: they are costs relative to what we could realize in better circumstances.

This seems to encapsulate a common aspect of regret in cases where one does what one ought: the thought that “it would have been better if I hadn’t had to do that.” In fact, non-utilitarians are likely to share exactly this thought when switching the trolley or (as a follower of NIE would) pursuing inegalitarian policies in the levelling down cases described above. It is hardly controversial that it would be better if nobody had been in the way of a runaway trolley, and if economic development and egalitarian social relations were always mutually compatible. This shared intuition, and thus this kind of regret can be explained, I have suggested, with the resources of utilitarianism: the reason that it would be better if I hadn’t had to do that is that in such circumstances, overall well-being would be greater. So utilitarians can regret aspects of what, in their circumstances, is the optimific state of affairs.

These considerations, together with the arguments of §6, show that there is no important difference between NIE and utilitarianism, even though the latter does not disvalue (b)–(f) intrinsically. Both can have distinctively egalitarian concerns (b)-(f). Both rely on contingent empirical claims to mandate the reduction of inequalities of wealth, and where those claims go through for O’Neill they are highly likely to go through for the utilitarian. Both will, sometimes, endorse inequality that produces (b)–(f), and both will see this consequence as a cost. The times at which they do so might differ, but there is no guarantee that NIE, given its pluralism, will endorse inequality less often than utilitarianism does. The fact that one and not the other theory values (b)–(f) intrinsically, then, does not seem to issue in important differences. Both are equally forms of egalitarianism.

Furthermore, there may be advantages to not valuing egalitarian concerns intrinsically. In particular, if one values them because of their effects on well-being, one has a principled way of establishing the degree to which they are valuable: they are valuable to the degree to which they affect overall well-being. O’Neill writes that “Any plausible view of the value of equality, which acknowledges the force of considerations (b)–(f), will also acknowledge the great significance of those considerations” (2008: 144). But we might want an argument as to why these considerations have such force. This is especially pressing for a pluralistic NIE, since it acknowledges other concerns which might conflict with egalitarian ones. If we cannot show that (b)–(f) have significant force relative to those other considerations, then we might worry whether we have given an argument for egalitarianism at all (Brown 2014). Utilitarians can provide a rationale for the relative significance of egalitarian concerns: they can appeal to the arguments of §3 which showed that equality has large effects on well-being. Therefore, the utilitarian can say, since overall well-being is the measure of moral significance, equality is morally significant. The utilitarian can also work out how much weight these concerns have relative to others in cases of conflict—by assessing the impact on well-being of those other concerns and comparing it with the impact on well-being of egalitarian concerns.

8. Conclusion

To show that utilitarianism is not a form of egalitarianism, one would have to find criteria for a theory to be a form of egalitarianism that utilitarianism fails to meet. Here is a review of possible such criteria, and a summary of this paper’s arguments with respect to them.

a. Egalitarian theories favour more equal distributions of well-being. This is true of some egalitarian theories, but as a necessary condition for a theory’s being egalitarian it is flawed: it does not capture our intuitive commitment to equality, or the aims of real-world egalitarian movements, as is shown by WEALTHY MAN.

b. Egalitarian theories consistently provide reasons to favour more equal distributions of wealth. Given DMU of wealth and the effects of relative poverty, utilitarianism meets this criterion. Those reasons are sometimes outweighed, but this is true of all plausible egalitarian theories.

c. Egalitarian theories favour more equal distributions of something intrinsically. This is not true of utilitarianism, but neither is it true of NIE or relational egalitarianism.

d. Where egalitarian theories do not favour more equal distributions of something intrinsically, they favour them reliably, in virtue of deep social facts. Utilitarianism meets this criterion at least as well as NIE.

e. Where egalitarian theories do not favour more equal distributions of something intrinsically, they favour egalitarian social relations. Utilitarianism meets this criterion, given the effects on well-being of inegalitarian social relations, suggested by empirical studies and various plausible theories of well-being.

f. Where egalitarian theories do not favour more equal distributions of something intrinsically, they favour egalitarian social relations intrinsically. Utilitarianism does not meet this criterion, but this is unimportant, since it can still mandate policies to mitigate or eliminate them (in most cases), account for the regrettableness of their instantiation (in all cases) and explain their great significance.

I conclude, then, that we should treat utilitarianism as a form of egalitarianism. No longer should utilitarians ignore egalitarian concerns, or utilitarianism be rejected for ignoring them.

Acknowledgements

Thanks to Michael Otsuka, Jonathan Wolff, Hilary Greaves, Véronique Munoz-Dardé, Joe Horton, Han van Wietmarschen, Douglas Lavin, Teru Thomas, Kacper Kowalczyk, and audiences at the Salzburg Conference for Young Analytic Philosophy and at the UCL/Leipzig Philosophy Workshop, for comments on versions of this paper. It has been through the hands of many reviewers and editors. I thank them all for their constructive suggestions, in particular those at Ergo.