Utilitarianism is in some important respects egalitarian, as I shall explain. One of the most significant of these respects is that it condemns as seriously wrong the inequality that marks the contemporary world. But it values equality in wealth or resources only instrumentally, and many egalitarians think that this misses some of the moral and political significance of this kind of equality. They object also to the specific distributive implications of utilitarianism. On the face of it, utilitarianism favours shifting resources from those who have expensive needs towards those who can extract more utility from them, who may be in every way better off. This sort of objection has seemed to many, especially in political philosophy, to be fatal to utilitarianism.1 This is one reason why its stock is even lower among political philosophers than among moral philosophers. When considered as a theory of justice utilitarianism is condemned even more widely. John Rawls and Robert Nozick—who in important ways have opposed views about justice—join forces in their criticism of utilitarianism as a theory of justice.2 Participants in later debates amongst egalitarians about the precise form of the best egalitarian theory were inclined to dismiss utilitarianism as not really even a candidate theory of justice.3 These critics had a point. It is not merely that each of them espouses a theory of justice that is at odds with utilitarianism, for the theories they espouse are also at odds with each other. The deeper point is that utilitarianism may seem not to have the conceptual resources to distinguish between justice on one hand and morality or goodness on the other. If the utilitarian theory of justice is simply that justice requires that social institutions maximize utility, for example, then its theory of justice coincides with its theory of the value of institutions. Nothing seems to be added by saying of the best institutions that they are also, for that reason, just. The complaint is then not so much that utilitarianism has the wrong theory of justice, but that it lacks a theory of the distinctive social virtue of justice at all.4 This chapter elaborates a more complex utilitarian theory of justice, according to which justice requires respect for moral rights. This enables us to distinguish justice from goodness, and so to explain the distinctiveness of justice as a virtue from a utilitarian point of view. Of course, the result is not a theory of justice that will satisfy all of those with egalitarian intuitions. But it will enable a more accurate estimation of the force of the challenges to utilitarianism from equality and from justice. 7.1 Distributive Justice Our focus is on distributive justice in particular. This is a matter of justice in the distribution of benefits and burdens among people.5 This topic is broader than justice in the distribution of wealth and money. Among the relevant burdens and benefits are rights and liberties, powers, and social standing. Even defined in this broad way, distributive justice is only one kind of justice. For tractability we will assume that it makes sense to discuss it separately from the other kinds, such as justice in war, or epistemic justice, or retributive justice. There are, no doubt, important connections between these topics, and a full treatment of any of them may have to discuss one or more of the others.6 But our aim is much more modest than this: we wish only to see whether utilitarianism can include a theory of distributive justice that is both distinct from its theory of the goodness of institutions and worth taking seriously. One question is whether utilitarians should want to develop a theory of distributive justice at all. They could, instead, seek to answer questions about the proper distribution of benefits and burdens by appeal to their general theory of rightness and goodness. For example, they might say that an allocation of benefits and burdens is good insofar as it promotes well-being, and that an action which alters the distribution of benefits and burdens is right insofar as it makes the outcome as good as possible. Why should they want to say anything about justice, in addition?7 One answer is that convictions about distributive justice are an important subset of people’s ethical convictions. Any ethical theory that seeks to explain people’s convictions on the whole had better be able to explain convictions about justice, or to explain why they do not need explaining. Offering a theory of justice is one way to try to meet this demand. However, we might also wonder whether people are right to have convictions about justice. Do they play any important role not played by convictions about goodness or rightness? There is no agreed or uncontroversial way of distinguishing justice from other normative concepts. Some try to distinguish it by saying that it applies to institutions rather than actions, though that runs the risk of construing it too narrowly.8 Others try to distinguish it by virtue of its alleged connections with the legitimacy of states or with rightful coercion.9 Others claim that justice is a matter of respect for people’s rights.10 These views may be combined in various ways. I will construe distributive justice as consisting in respect for people’s rights in relation to the distribution of benefits and burdens. This fits with several commonplace claims about justice. For example, it fits with the idea that to have been treated unjustly is not merely for a wrong to have occurred, but to have been wronged. It also fits the idea that justice can be claimed or demanded, as of right. It fits the idea that justice is a matter of impartiality, since rights are also supposed to have that feature. It fits the idea that justice is an important part of morality, though not the whole of it. Finally, it fits the idea that justice is internally complex, in the sense that what justice requires depends on a balance of competing considerations—since what rights someone has also depends on a balance of competing considerations.11 The downside of thinking of justice as consisting in respect for rights is that it risks construing justice too broadly. If Rita promised to lend Marcel her lawnmower, Marcel may have acquired a moral right to use it; but if Rita does not in the end lend it to Marcel, it may seem odd to say that Rita has behaved unjustly, or that Marcel has suffered an injustice.12 One response to this would be to concede that the account is too broad, and to refine it so that it claims that justice is a matter of respect for moral rights by institutions. But this refined view seems too narrow: intuitively, individuals and their actions can also be unjust. This suggests a different response. If individuals and their actions can be unjust, perhaps we should accept that Rita’s failure to lend Marcel her lawnmower was indeed unjust. The oddness involved in saying that it was unjust might reflect pragmatic considerations, deriving from the fact that it is not a very important injustice. We might be reticent to describe it as an instance of injustice for the same reason that we would be, in fact, reticent to describe it as having violated Marcel’s moral right to use the lawnmower. Both descriptions are overblown and inappropriate in normal circumstances, whether or not they are true. If this reply fails to satisfy, we could hope to find some other restriction—short of the restriction to institutions—to distinguish injustice from other infringements of rights. Either way, we will proceed on the assumption that justice is either entirely a matter of respecting rights, or that respect for rights is a central requirement of justice. The distinctive ethical significance of justice will then be either wholly or partly explained by the distinctive ethical significance of rights. According to the view we have been developing, rights are one main source of constraints on the promotion of well-being. They are ethically significant in their contribution both to determining which actions are right and to the moral psychology of virtuous agents. So we can distinguish distributive justice from other parts of morality, and utilitarians have some reasons to offer their own theory of it. If they do so they will be pitting their theory against a wide range of sophisticated contending views. These include egalitarian views of various sorts, according to which justice requires equality of well-being, or of opportunity for well-being, or of resources, or something similar to equality in one of these dimensions, such as sufficiency or priority for the worse-off.13 They also include the view that justice requires distribution of goods according to moral desert, so that it is unjust if sinners fare better than saints.14 Further, they include Nozick’s historical entitlement theory, according to which justice places constraints on the way that goods are originally appropriated and on the ways in which goods are transferred, but not on the resulting patterns of distribution of goods.15 Utilitarians must hope to offer a theory of distributive justice that is, in comparison with these other views, sufficiently plausible to be worth taking seriously. 7.2 Justice for Utilitarians Critics of utilitarianism typically assume that it must endorse one of two theories of distributive justice. The first of these is the simplest possible theory, according to which justice requires that goods are allocated in whichever way maximizes utility. According to this theory, ‘justice’ is a property of allocations of goods, and it is co-extensive with maximum goodness. This theory is not at all plausible. It does not assign any distinctive role to the concept of justice, and its implications about justice are wildly out of step with ordinary convictions. Suppose, for example, that Ellen is about to donate her life savings to a very effective charity, when Joe comes along and steals the money and donates it to a slightly better charity, thereby making the outcome better. This theory implies that Joe’s action made the world more just. This is, I assume, starkly at odds with ordinary intuitions about justice. Utilitarians are committed to claiming that Joe’s action made the world better. Depending on their theory of rightness, they may also claim that this makes Joe’s action right. But it is very implausible for them to claim that, in addition, he made the world more just. This is to co-opt the concept of justice in a way that is both unconvincing as a claim about what justice consists in, and adds nothing worthwhile to the existing utilitarian claims about this case. We should reject this very simple theory. In his discussion of utilitarianism as a theory of justice, John Rawls attributes a much more plausible claim to utilitarians. Rawls writes: The main idea is that society is rightly ordered, and therefore just, when its major institutions are arranged so as to achieve the greatest net balance of satisfaction summed over all the individuals belonging to it.16 Generalizing slightly—and ignoring Rawls’s methodological concern with only a single society—we can say that according to this utilitarian theory of justice, a society is just if and only if and because its ‘major institutions’ maximize utility. In contrast to the previous claim, this theory characterizes justice as a property of societies, which exists when their institutions are maximally good. Justice is then a matter of the goodness of institutions, not of allocations. One advantage of this theory over the previous one is that it lets rights enter the picture. It is not certain, but quite plausible, that utility-maximizing institutions will assign rights to people, and it could be that these rights will match reasonably well our ordinary convictions about what justice requires. Rawls goes on, of course, to give reasons why he thinks this theory of justice will not be satisfactory all things considered.17 But whatever our view of the force of those criticisms, this theory is much more plausible than the simple theory we just considered. It may well be able to explain the conviction that Joe acts unjustly in stealing Ellen’s life savings, for example, even though he makes the world better in doing so. Whether it can do this depends on some details in the formulation of the theory, together with some hard to assess empirical claims about which institutions would maximize utility. The details of formulation concern how to spell-out the idea of ‘best institutions’. One issue is the set of relevant alternatives—best out of which set of possible institutions?—and another is how to characterize the consequences of institutions. Depending on how we answer these questions, the theory Rawls attributes to utilitarians could end up at any point on a spectrum from recommending pragmatic piecemeal institutional reform, at one end, to a highly idealizing form of ideal theory, at the other. In any case, this broad approach to thinking about justice is not, I believe, the way for utilitarians to go. It yokes justice too closely to goodness, and in doing so it suffers from a problem that also afflicts other theories of justice. The problem arises if we accept Rawls’s idea of ‘legitimate expectations’, understood in a certain way. Legitimate expectations are entitlements generated by adherence to the rules of a sufficiently just society.18 For example, if you live in a sufficiently just society and you undertake the duties specified in your employment contract, you acquire an entitlement to the wages it specifies. This is not merely a legal right; if it were, there would be nothing unjust in changes to the law that cancel your legal entitlement. Your entitlement to your wages is also a moral right. The crucial question is how just a society must be for legitimate expectations to generate moral rights in this way. Rawls does not answer this question. Most of his discussion of legitimate expectations occurs in the context of discussing ideally just societies, though at one point he makes remarks that suggest that he believes they generate moral rights also in somewhat unjust societies.19 In any case, I assume that legitimate expectations generate moral rights even in societies that are not perfectly just. If we do not assume this, we seem forced to say that arbitrary changes in the law which confound people’s expectations are not unjust when they take place in imperfectly just societies. The problem is what to say about property rights in sufficiently-though-not-perfectly just societies. We are assuming that members of these societies acquire moral rights to things in virtue of legitimate expectations. If we endorse a theory of justice which implies that, ideally, justice requires a certain allocation of property, or that it requires that a certain procedure is followed for acquiring property, we seem forced to allow that someone’s property in some item could be both a requirement of justice (because he has a moral right to it generated by legitimate expectations) and disallowed by justice (because the ideally just allocation or procedure implies that it is someone else’s property). There are some ways of avoiding this problem. One is to deny the assumption we have been making, that legitimate expectations give rise to moral rights even in imperfectly just societies. That is a high price to pay, since we rely on this assumption in explaining the injustice of arbitrary changes to the law. A second way is to distinguish senses of ‘required by justice’, to avoid the threatened contradiction. We could say, for example, that in the imagined case justice ‘non-ideally requires’ that one person have the property, but ‘ideally requires’ that someone else have it. This is, I think, the way that some theorists of justice hope to avoid this difficulty. But it is not very satisfactory, since it conflates two different issues. One issue is what justice requires in different circumstances: in the present circumstances, it requires that one person has a property right in this item; in other circumstances, it requires that someone else has a property right in it. A different issue is the nature of the requirement that justice imposes in any circumstance. The distinction between ‘ideally requires’ and ‘non-ideally requires’ on the face of it suggests that there is some defect in the way in which justice requires something when it non-ideally requires it. Rather than distinguish ideal requirements from non-ideal requirements, we should simply distinguish what is required in ideal circumstances from what is required in other circumstances. In our imperfect society—provided it is sufficiently just for legitimate expectations to generate moral rights—justice requires that you get your wages.20 This is not a second-rate requirement of justice, but a full-blown requirement of it. There is no superior way in which justice requires things, according to which it requires that someone else have this property. All that is true is that, in different circumstances, justice may require that someone else have it. This suggests a way of thinking about justice which is highly procedural and which distinguishes much more sharply between justice and goodness than does the theory Rawls attributed to utilitarians. The rights that justice consists in respecting, on this view, are not those assigned by ideal or maximally good institutions. Instead, they are the moral rights that exist in the circumstances. In actual circumstances, they are the moral rights that people actually have. These include the core moral rights that are independent of any existing legal system, together with those moral rights which are generated by the mechanism of legitimate expectations, which depend on the features of the existing legal system. Someone’s moral right not to be tortured is of the first kind, which implies that justice requires that she not be tortured regardless of whether her jurisdiction permits it. If she lives in a sufficiently just society—let us say, one in which most core moral rights are respected—then justice also requires that her legitimate expectations are satisfied. If, however, she lives in a grossly unjust society, the social rules will not generate moral rights through the mechanism of legitimate expectations. In that circumstance, justice requires only respect for the core moral rights—though prudence, efficiency, or other considerations may generate reasons for satisfying people’s expectations, even though justice does not require it. Utilitarians should, I believe, think about justice in this way. Perhaps surprisingly, it is a way of thinking according to which justice is significantly further removed from goodness than most contemporary theories of distributive justice would have it. But in identifying justice with respect for the moral rights that people in fact have, it assigns a distinctive and worthwhile role to the concept of justice. To act justly is to act within the constraints generated by people’s moral rights. It will not always be possible to maximize the good and to act justly—as the example of Joe and Ellen illustrates. This way of thinking of justice is in some ways similar to Nozick’s theory. Like his theory, it defines distributive justice as simply a matter of respecting people’s moral rights, and not as involving the production of favoured outcomes.21 Also like his theory, it implies that any society in which all moral rights are respected is perfectly just. Justice is sharply distinct from other virtues, and a perfectly just society may be far from ideal in other ways. In contrast to Nozick, however, utilitarians will tend not to think of property rights as generated ‘pre-institutionally’, but instead as generated by people’s behaviour against a background of sufficiently just institutions.22 7.3 Kinds of Equality In order to address egalitarian objections to utilitarianism we need to distinguish some different kinds of equality. This will also enable us to consider which, if any, of these kinds of equality will be requirements of justice according to the theory just outlined. One kind of equality is what Rawls calls ‘formal justice’.23 This consists of applying rules or laws to people equally, in the specific sense that like cases are treated alike—where relevant differences are defined by the rules or laws themselves. This kind of equality is inconsistent with favouring some out of whim or personal preference, but it is consistent with favouring some because they are favoured by the law or rule itself. This kind of equality is not trivial or pointless: if the rules state that no-one may park in a certain area, and we have both parked there, but I get a ticket and you do not because you are friendly with the enforcement officer, we have been treated unequally in this sense. But neither is this sort of equality all that we care about. If the rules state that friends of enforcement officers may park where they like, I cannot complain on grounds of formal justice. Utilitarians are likely to regard formal justice as important because of its instrumental value. Unless they are Rule Utilitarians and the rules in question are the rules in the ideal code, they will not regard equal and impartial application of rules as a fundamental moral requirement. Nevertheless, formal justice is constitutive of organizing things according to rules or laws at all: it is not an optional extra. Since there is often great value in organizing things that way, utilitarians will view formal justice as necessary to realize the instrumental value that is often achieved by organizing social life using rules and laws. A second kind of equality is equal importance of equal benefits and harms. Utilitarianism is fundamentally committed to this, since it treats an equal-sized benefit or loss as making an equal-sized difference to overall goodness, no matter to whom the benefit or loss occurs. In the phrase that Mill used to sum up Bentham’s view on this matter, ‘everybody to count for one, nobody for more than one’.24 This idea is entailed by the way that sum-ranking features in utilitarianism. As we have defined them, utilitarians claim that the value of an outcome is equal to the sum of the utilities it contains. This implies that an increment or decrement of one ‘utile’ makes the same difference to the goodness of an outcome, no matter to whom it falls. Equal importance of equal benefits or harms is one way of conceiving the equal importance of persons. Sometimes it is said that it fetishizes utility, as if what it really expresses is a concern that utiles are treated equally. But this criticism is misplaced. The only reason that utilitarians care about utiles is that they are bits of goodness-for-people. The reason that they treat each utile as counting the same is that they treat people as equally important. On the other hand, it is also true that equal importance of equal benefits and harms is not the only way of conceiving the equal importance of persons. Prioritarians, for example, claim that benefits and harms matter more, morally, the worse off the subject. This might be thought to conflict with the idea that each person is equally important. But it is not clear whether it does so. Prioritarians do not attach extra importance to helping the specific individuals who are worse off, thereby treating those individuals as more important than others. Instead, they attach extra importance to helping the worse off, whoever they are. The extra importance goes with the level of well-being, not the individual. They can say to each person, ‘we would have just the same concern for you, if you were as badly off’.25 And, of course, egalitarians of different stripes offer their own conceptions of the equal importance of persons.26 A third important kind of equality is equality of basic moral rights. Everyone who believes in moral rights at all agrees that different people have different sets of moral rights, since they can acquire or lose moral rights through their actions. Thus if I am rightly imprisoned, I have presumably lost, temporarily, some moral rights of free movement. If you keep your part of a just contract, you gain a moral right that others lack to the other party’s performance. But an important and widely shared belief is that there is some set of basic moral rights that are equally held by all persons.27 As we noted in Chapter 6, utilitarians who believe in moral rights are likely to think that what basic rights people have depends in some way—not necessarily in a highly sensitive way—on the circumstances. This follows from the fact that any utilitarian justification of moral rights will be, ultimately, an instrumental one. So utilitarians should accept that in at least some possible circumstances, people would not have moral rights, or not the ones we have. Therefore, utilitarians cannot support the claim that there is a set of basic rights that is possessed by all people in all circumstances. However, they have some grounds for claiming that, in cases where some people have a specific set X of core moral rights, others in the same circumstances also have X. This is a narrower kind of equality of basic rights, since it applies within circumstances, not across all possible circumstances.28 Even this narrower kind of equality of basic rights is not guaranteed; it is not written into utilitarianism’s fundamental principles. But it is supported by the same considerations that we appealed to in Chapter 6 in order to argue that the best patterns of respect for rights will not include making exceptions between cases. There we noted that the benefits of ‘respect for rights’ flow not merely from the behaviour such respect consists in, but also from the motivations it consists in. In particular, it seems likely that the motivation to depart from rules governing simple kinds of cases—such as the rule ‘do not torture’, governing opportunities to torture—will on the whole be counterproductive in utilitarian terms. If you believe that others are ready to break this rule when they judge that it would maximize utility to do so, your security will be undermined. The same seems to go for other basic moral rights, such as the right to freedom of thought or association. If this is correct, it provides grounds for thinking that utilitarianism will support the narrower kind of equality of basic rights. The reason for this is that the possible grounds for beneficially distinguishing between persons also seem to be grounds for distinguishing between circumstances in respect of the same person. If we think it counterproductive for people to be motivated to treat person S differently just in case they judge her to have property P—where P corresponds with an opportunity to promote utility by departing from the usual rule governing cases of this kind—then we should also think it counterproductive for people to be motivated to distinguish between people according to whether they have property P. At least normally, the grounds for distinguishing between people will also be grounds for distinguishing between cases. Thus the reasons not to distinguish between cases will normally be reasons not to distinguish between persons. There is, however, an important loophole. We are appealing here to hypotheses about the effects of different possible motivations. These effects operate through psychological and sociological mechanisms, not simply through chains of logical implication. If people conceive of others as divided into groups, they may be able, psychologically, to acknowledge different treatment of members of groups to which they take themselves not to belong, without suffering the harms of insecurity we have been appealing to. They may think, ‘failure to respect their rights is compatible with secure respect for mine’. Unfortunately, we have ample evidence that people draw such distinctions. For this reason, the utilitarian case for narrow equality of basic rights is incomplete. In conditions where social divisions are widely accepted, it is in principle possible for inequality of basic rights to be utility-maximizing. This is an unwelcome implication of the utilitarian theory of rights. In this respect, it fails to capture some of our deepest convictions about rights and justice. Utilitarians can seek to mitigate the problem in at least two ways. One is to point out that where there are genuine grounds for distinguishing between groups—such as in the case of the groups ‘young children’ and ‘competent adults’—we are inclined on reflection to accept inequality of basic moral rights.29 Our attachment to equal rights is not sweeping. The more problematic case is where the social division is based on mere prejudice. Utilitarians then seem forced to admit that, holding the prejudice as given, it is possible in principle that inequality of basic rights across the groups concerned would maximize utility.30 However, they can emphasize that in cases such as this we should not hold the prejudice as given, and that it is very likely that there would be strong reasons for trying to reform or eradicate the prejudicial practice. A fourth kind of equality concerns the material conditions of people’s lives in a way that goes beyond formal equality, equal importance of persons, or equality of basic moral rights. Egalitarians disagree amongst themselves over how exactly to specify the relevant concern here. Some claim that we should favour equality of welfare; others that we should favour equality of resources, or equality of opportunity for welfare, or equality in some other ‘currency’.31 They also disagree about whether it is, strictly, equality in one of these dimensions that we should favour, or instead some kind of sufficiency or some form of priority for the worse off.32 They disagree, thirdly, about the relevant sense or senses in which these things should be ‘favoured’. Some think that they are favoured in the sense of being required by justice, while others think that they are favoured in the sense of making the outcome better, and others think they are favoured in both senses.33 But there is a broad sense in which they agree with each other that some form of what we might call ‘substantive equality’ is required by justice, and that utilitarianism goes wrong by failing to recognize this.34 I have claimed that utilitarianism supports equality in the sense of formal justice, that it offers one conception of the equal importance of persons, and that with some important qualifications it supports equality of basic moral rights. The remaining question is thus whether utilitarianism is insufficiently egalitarian in the specific sense that it fails sufficiently to value substantive equality. 7.4 Utilitarianism and Substantive Equality All utilitarians, as I have defined them, are committed to sum-ranking. This is the claim that the value of an outcome is equal to the sum of the goods (and bads) it contains. For utilitarians this means that the value of an outcome is equal to the sum of well-being it contains. This is the sense in which utilitarianism is, as is often said, ‘purely aggregative’. There is no room for manoeuvre over this, if we understand ‘utilitarianism’ in the way I have defined it. However, consider the following example: (20, 40) (30, 30) World A World B Each world contains just two individuals, whose well-being is represented by the numbers. In world A, the second individual is twice as well off as the first. In world B, both are equally well off, at a level midway between the two levels in world A. There are no further consequences associated with either world, so the numbers represent all of the well-being for each world, not just the well-being in it at some time or during some period. Utilitarians are committed to claiming that these two worlds are equally good. Some take this claim to be sufficient reason to reject utilitarianism.35 That is, they reject utilitarianism because it is incompatible with the judgement that, in cases like this, the two worlds are not equally good. Since this judgement is incompatible with sum-ranking, endorsing the judgement gives us reason to reject sum-ranking—and, therefore, all forms of utilitarianism. Call this the purely axiological objection to sum-ranking. For it to be pure, we have to set aside any thoughts about possible causes or remedies of the inequality in World A—since these might activate intuitions about justice or duty. So we should imagine that the individuals in World A do not interact with each other in any way: let us say that each lives a lonely existence, separated by several centuries. To make it easier to grasp how their lives differ, suppose that they have the same, constant, level of well-being while alive, but that the first person lives for only twenty years, while the second lives for forty. Finally, for symmetry, make parallel assumptions about the individuals in World B. With these clarifications in place, the purely axiological objection to sum-ranking does not seem to me to be very powerful. It is certainly not sufficient to defeat utilitarianism by itself. Utilitarians can agree that there is an important respect in which World B is better than World A: both individuals in B live better lives than the worse off individual in A. They can also agree that, if the worse off person in A could live for thirty years rather than twenty, that would be better for him. They merely insist that these facts are balanced by the respect in which A is better than B: the better off individual in A lives a better life than the individuals in B. It is not at all obvious that sum-ranking goes wrong in making this claim. Admittedly, sum-ranking makes the same claim no matter what the ratio between the levels of well-being of the individuals, so long as the total in each world is the same. As the ratio goes up, we may be more tempted to doubt whether sum-ranking could be right. But equally, as the ratio goes up, we should be more doubtful of whether we have grasped the features of the example accurately in forming our intuitions. The second doubt undermines the first. If the purely axiological objection works, it should work for the example specified, which we can readily understand. The purely axiological objection would defeat all versions of utilitarianism were it to succeed against any, but it is too flat-footed to succeed. More interesting objections operate by appealing not simply to a verdict about a simple case, but also by offering an explanation of the relevant features of the case. One such objection is that utilitarianism goes wrong by failing to recognize the noninstrumental value of substantive equality. As noted above, ‘substantive equality’ has to be specified further—in terms of well-being, resources, opportunity for well-being, or some other currency. But faced with any objection of this sort, utilitarians will reply that the value of the specified sort of equality is purely instrumental, not noninstrumental. We will then have something of a standoff. In favour of the egalitarian side, a claim such as ‘equality of well-being is good in itself’ has some intuitive appeal, quite apart from its implications about cases. Against that, though, we have the powerful appeal of the levelling-down objection to this sort of claim.36 The levelling-down objection has led some with broadly egalitarian sympathies to favour some form of sufficiency or priority instead.37 In so doing, they side with utilitarians in claiming that equality itself has purely instrumental value. However, sufficiency and priority principles do not have as much intuitive appeal, independently of their implications about cases, as the egalitarian claims they displace. The claim that benefits matter more when they go to the worse off is appealing, I assume, primarily because it seems to explain egalitarian judgements about cases while avoiding the levelling-down objection. Taken just by itself it is certainly no more plausible than sum-ranking.38 These points suggest that we should care more about whether a theory generates the right degree of support for substantive equality than about whether it endorses a claim about value that seems, in the abstract, intuitively appealing. Insofar as we think that substantive equality is valuable, it matters greatly whether utilitarians can explain this value. Their explanation will be in terms of its instrumental value, but for the time being we are setting aside intuitions about whether the value of equality is purely instrumental or also noninstrumental. Considerations of diminishing marginal utility are sometimes used to claim that utilitarianism tends to favour one sort of substantive equality: equality of resources. Resources have ‘diminishing marginal utility’ in the sense that, usually, someone gains less well-being or utility from an additional unit of resources, the more resources she already has. This is an empirical claim with some exceptions, but it is nevertheless an important and credible generalization.39 A very simple argument that utilitarianism supports equality of resources states that, given this generalization, we can typically increase overall well-being by equalizing resources, since the loss to the rich will be outweighed by the gains to the poor. This argument sets aside the effects of inequality on productive output, and asks instead how to distribute a fixed amount of resources so as to maximize utility. Even with this caveat, however, this argument fails. The reason the simple argument fails is that the generalization that, usually, each person will gain less from each additional unit of resources does not imply that different people make gains at the same rate. Nor does it seem to be true that different people make gains at the same rate. Some people, for example, have more expensive goals, ambitions, or tastes than others. Some are also inclined to expect more of their goals, ambitions, and tastes to be fulfilled than are others. Some are more adept at pursuing their goals than others. Furthermore, some people have expensive medical or other needs not shared by others. All of these facts make it likely that different people gain personal utility at different rates—all of which is consistent with the generalization that, for each person, additional units of resources typically bring smaller gains in utility.40 In a fixed population, total utility is maximized when marginal utility is equalized. If different people are on different ‘utility curves’ (that is, the function from resources to utility is different for each person) their marginal utilities will be equalized at different levels of resources and of utility. Consider Figure 7.1. The marginal utilities of Bob and Sheila are equalized, and so their joint utility is maximized, when they are placed at points on their respective utility curves that have equal gradients. As Figure 7.1 shows, this leaves Bob with more resources and more utility than Sheila. The reason for this is that he gains more utility, in general, than Sheila does from each unit of resources—his curve is steeper than hers, at each level of resources. Redistributing resources from Bob to Sheila to equalize their amounts of resources would decrease total utility, since she would gain less utility than he would lose. Figure 7.1 C7.F1 Why the simple argument does not work Open in new tabDownload slide Why the simple argument does not work This means that an omniscient utilitarian administrator concerned to maximize the utility gained from a fixed amount of resources will, in allocating them so as to equalize marginal utilities, almost certainly give different people different amounts of resources, thereby leaving them with different amounts of utility. If substantive equality means equality of welfare or equality of resources, this administrator would not produce substantive equality. However, we should not think of substantive equality in this way. We do not choose directly between allocations of resources, but instead between institutions or general policies that have somewhat predictable effects on allocations. Second, we should not ignore the effects of distributive policies and institutions on production, or the total stock of goods. Since what matters for utilitarians is the whole future, they will also want to bear in mind issues of sustainability rather than aiming for maximum utility for currently existing people.41 Still, we do not accurately model that problem by assuming a fixed stock of goods. Third, and perhaps most importantly, there are no omniscient utilitarian administrators. I argued in Chapter 4 that it is very difficult even for an individual to know what would be best for her. The idea that any administrator could know how to maximize utility for a population of people is so unrealistic as to be not even a useful simplifying assumption when considering these matters.