With respect to social justice, what matters most is that people have enough. Sufficientarianism says that our first (or perhaps only) distributive priority is to relieve want.

This article develops and defends a revised sufficientarianism. I'll argue that it furnishes the best specification of a general humanitarian ideal of social justice: our main moral concern should be helping those who are badly off in absolute terms. I begin by motivating the sufficientarian claim that the demands of justice change with respect to individuals once they meet a particular threshold of resources; I then offer an original specification of that threshold (section I). Next, I address the most pervasive objection to sufficientarianism, namely that sufficientarian regimes will channel all of society's resources towards elevating people to the sufficiency threshold regardless of the gains foregone by those above the threshold (sections II–III).

I

The common core of the family of sufficiency views is an acceptance of the claim that benefiting individuals below a threshold of sufficiency has moral priority over benefiting individuals above the threshold. Some endorse the stronger claim that benefiting individuals beneath the threshold of sufficiency has lexical priority over all rival distributive goals (that is, it unconditionally defeats all rival distributive goals).Footnote ¹ The distinctive feature of sufficientarianism is its claim that obligations of distributive justice lessen or perhaps even dissolve with respect to individuals once they've met some threshold of adequate resources or welfare.

Part of the case for sufficiency involves showing that it avoids counterintuitive implications that its major rivals do not. I'll briefly discuss worries about three principles that might similarly be called ‘humanitarian’ in virtue of their tendency preferentially to benefit the worse-off: egalitarianism, prioritarianism and the difference principle. By no means is my aim to defeat any of these conceptions decisively. I simply want to note objections to these views that appear compelling and then suggest that sufficientarianism is largely exempt from these objections.

Egalitarian views hold that equality, somehow understood, is itself a value. Egalitarians typically focus on neutralizing inequalities in welfare, resources or capabilities that arise through luck.Footnote ² One common challenge to egalitarianism is the levelling-down objection: egalitarianism implies reason to favour distributions that harm the better-off and benefit none.Footnote ³ Levelling down increases equality – even if it harms those who are already poorly off in absolute terms. Others object to luck egalitarianism on the grounds that a satisfactory theory of distributive justice must have something to say about the (un)acceptability of poverty even when it is as a result of an individual's choices. Worries about choice-based deprivation lead some to suggest that luck egalitarianism should be supplemented with a sufficiency constraint.Footnote ⁴

Prioritarianism holds that a gain is more morally valuable the worse-off the gainer.Footnote ⁵ It avoids the levelling down objection because it doesn't regard equality as a value in itself. Prioritarianism's aim, in Richard Arneson's words, ‘is to help the person who in absolute terms is badly off, by the relevant standard of judgment, not to bring about some preferred relation between what one person has and what others have’.Footnote ⁶ However, there are reasons to doubt the view. For one, prioritarianism implies reason to transfer resources to the rich – as long as the transfers are funded by the ultra-rich. This is a counterintuitive implication and it suggests that prioritarianism misunderstands the point of transfers.Footnote ⁷ Their point seems to be, in Arneson's own words, ‘to help the person who in absolute terms is badly off’. More significantly, prioritarianism allows for what Arneson calls the ‘tyranny of aggregation’.Footnote ⁸ Our moral concern for a person declines smoothly as her well-being increases. Yet this commits prioritarianism to securing a tiny benefit for many absolutely well-off people over securing an enormous benefit for a few poorly off people as long as enough people each receive the tiny benefit. To borrow Roger Crisp's example, it seems wrong to forgo relieving a few people's severe pain for the sake of providing chocolates for many already well-off people.Footnote ⁹

Lastly, John Rawls's difference principle asks us to maximize the socio-economic well-being of the worst-off.Footnote ¹⁰ It avoids the tyranny of aggregation by assigning this goal lexical priority: the interests of the worst-off unconditionally trump other interests. However, one concern about the difference principle is that it doesn't tell us how to arbitrate claims among the worst-off class. Are the interests of the worst-off among the worst-off prioritized above the better-off among the worst-off? Are they given lexical priority? If not, a question arises about why lexical priority is justified for the worst-off class itself but not individuals within the worst-off class. Furthermore, it is unclear that there's a principled threshold that defines the ‘worst-off’ class. For example, does the cut-off occur at the bottom 1 per cent, 5 per cent, or 10 per cent? Regardless of where one locates the cut-off, it's not obvious what would justify this threshold rather than some other threshold. This problem is critical because those who are excluded from the worst-off class have their socio-economic interests unconditionally trumped – and in some cases, trumped by those who are only marginally poorer.

To reiterate, my point here is not that the above principles are incapable of being patched in ways that enable them to overcome these problems. Rather, I want to show that sufficientarianism fares well in comparisons with these principles, as it's either exempt from these problems entirely or can resolve them in a straightforward way. So let's start refining a sufficiency view. There are different conceptions of sufficientarian justice.Footnote ¹¹ Harry Frankfurt suggests that resources should be allocated to minimize the number of individuals below the threshold of sufficiency.Footnote ¹² Others, such as Roger Crisp, claim that the condition of individuals below the threshold matters too. Crisp favours prioritarianism below the threshold: the worse-off an individual, the greater priority we ought to assign to advancing his or her interests.Footnote ¹³

We can see immediately that sufficientarianism avoids the levelling-down objection that besets egalitarianism. There is no imperative to equalize relative standing. Moreover, unlike prioritarianism, sufficientarianism will not demand transfers to the rich, as it claims that obligations of distributive justice change with respect to individuals at some threshold of resources or welfare. Interestingly, Arneson himself motivates prioritarianism against egalitarianism by noting that we lack reason for egalitarian transfers between absolutely wealthy parties. ‘After all,’ Arneson writes, ‘the gap between rich and poor could also appear in a contrast between rich and superrich, but the moral imperative of transfer of resources to aid the rich seems far from compelling or even nonexistent.’Footnote ¹⁴ As noted, however, prioritarianism of the sort endorsed by Arneson does imply (ceteris paribus) an obligation to transfer resources from the superrich to the rich. Arneson's intuition that we lack a moral imperative to favour the rich at the expense of the superrich seems to rely covertly on a sufficientarian idea, viz. that there really is a point at which it can be said that someone has enough, at least as far as special distributive obligations are concerned. I think that prioritarianism's initial appeal is because of its commitment to helping people who are badly off; however, the ‘mere’ rich are not badly off, and so we lack reason to prioritize their interests. An intuitive advantage of sufficiency, then, is that it caps prioritarian concern.

However, this move to truncate prioritarianism raises a question. Perhaps priority to the worse-off ought to terminate somewhere, but where exactly? Candidate thresholds abound, but none seem to satisfy. Crisp writes that ‘compassion for any being B is appropriate up to the point at which B has a level of welfare such that B can live a life which is sufficiently good’.Footnote ¹⁵ Martha Nussbaum says: ‘The job of a decent society is to give all citizens the (social conditions of the) capabilities, up to an appropriate threshold level.’Footnote ¹⁶ Frankfurt claims that ‘to say that a person has enough money means that he is content, or that it is reasonable for him to be content, with having no more money than he has’.Footnote ¹⁷

Yet these specifications are vague. Paula Casal claims that sufficientarians face a dilemma between an arbitrary or ambiguous threshold. A determinate threshold of sufficiency can be reached only by selecting an arbitrarily precise point at which to stop prioritizing individuals’ interests (e.g. the possession of n dollars’ worth of goods). Non-arbitrary thresholds (e.g. the capability for ‘truly human functioning’Footnote ¹⁸) seem more capable of substantiating the importance of sufficiency, but they do not yield determinate criteria for structuring economic institutions.Footnote ¹⁹ Casal writes: ‘Given the importance sufficientarians attach to individuals having “enough,” perhaps the most pressing problem they face is to specify that idea in a principled manner that provides determinate and plausible guidance for distributive decision makers.’Footnote ²⁰

Most notably, the pivotal concept of ‘contentment’ is underspecified.Footnote ²¹ The notion of being content with what one has plays a central role in Frankfurt's formulation of sufficiency.Footnote ²² Robert Huseby has recently proposed revisions to the principle of sufficiency, arguing that a ‘maximal’ sufficiency threshold ‘equals a level of welfare with which a person is content’.Footnote ²³ Both accounts are vulnerable to the indeterminacy objection. The problem, as Casal puts it, is specifying the notion of ‘contentment’ in a principled manner that provides determinate and plausible guidance for distributive decision-makers. Prevailing accounts have not done that.

Before trying to address this objection, let me try to soften it. I don't think that working out a satisfactory account of economic sufficiency is a problem for sufficientarians alone. Recall the earlier worry that luck egalitarianism will tolerate choice-based poverty. We've seen that some theorists have suggested patching luck egalitarianism with a sufficiency constraint to avoid this sort of result. Indeed, contemporary political philosophers generally agree that distributive justice calls upon political institutions to ensure that their citizens have sufficient income.

Casal has recently noted that most theorists who are not expressly sufficientarian nevertheless agree that satisfying the sufficiency condition is an important moral concern. She says that ‘except among Hobbesians and libertarians, it is increasingly difficult to find views that do not accept some version’ of this thought.Footnote ²⁴ Even some Hobbesians and libertarians argue for an adequate social minimum.Footnote ²⁵ So the need to articulate an account of socio-economic sufficiency is not confined to those who work explicitly within the sufficientarian framework. Many contemporary political philosophers need to figure out how to define a sufficiency threshold.

In concert with most of the prevailing articulations of sufficientarianism, I will focus on defining a threshold of sufficient resources.Footnote ²⁶ My aim is to revise and defend an account of sufficiency as a fundamental principle of justice. The account takes inspiration from Frankfurt's view; accordingly, the first task is to specify a point at which people can be said to have enough resources, where this is understood in terms of being content with what one has.

So how might we go about the task? Arneson and Casal suggest that sufficientarians must rely on ‘intuition and reflective equilibrium’Footnote ²⁷ or ‘intuitionist stipulations’Footnote ²⁸ to define a threshold. However, I think that sufficientarians have resources beyond intuitions. Ample evidence indicates that beyond a certain point, the returns to increasing one's material resources level off dramatically in terms of life satisfaction. Consider the following from social scientists:

Researchers in psychology, economics and political science contend that material accumulation reaches a point of diminishing returns in terms of life satisfaction, and the diminishing returns may be steep.Footnote ³² Improvements in satisfaction brought about by income are additive up to a certain point, at which they start to tail off appreciably.Footnote ³³ Researchers generally agree that there is a definite point of diminishing marginal returns to income – a point at which the subjective returns to income level off.Footnote ³⁴

The point of diminishing marginal returns to income is a non-arbitrary, unambiguous threshold of sufficiency.Footnote ³⁵ This research furnishes reason to think that there is a principled way to bring specificity to the notion of ‘contentment’. That individuals’ concern for income diminishes appreciably beyond a specific threshold speaks in favour of prioritizing the goal of elevating them to that threshold and then subsequently relaxing priority. So sufficientarians might sensibly follow Crisp in endorsing prioritarianism below the sufficiency threshold, as opposed to simply minimizing the number of individuals below the threshold.Footnote ³⁶ Perhaps, then, sufficientarianism is best understood as truncated prioritarianism.

To be sure, self-reported satisfaction is an imperfect measure of well-being (although any measure of well-being is bound to have imperfections).Footnote ³⁷ However, the latest analysis of the life satisfaction data overturns many standard assumptions about subjective well-being and may soften some of the traditional worries about self-reports. For example, recent evidence indicates that individuals in poverty do perceive their lives as improving with increased income, which should go some way towards alleviating worries about adaptive preferences.Footnote ³⁸ Furthermore, self-reported well-being correlates strongly with non-self-reported metrics of welfare.Footnote ³⁹

Moreover, self-reported satisfaction is a crucial measure of well-being if one works within a broadly liberal framework. A conception of justice that aspires to be justified from each individual's own point of view must attend to how individuals’ lives go from the inside. By prioritizing citizens’ economic interests until the importance of additional income levels off by the citizens’ own lights, sufficientarian regimes respect their citizens’ own judgements about how their economic lives are going.

II

Now let's turn to the issue of aggregation. Recall the tyranny of aggregation: prioritarianism is committed to securing a tiny benefit for many people above the threshold over securing an enormous benefit for few people below the threshold as long as enough people receive the tiny benefit.Footnote ⁴⁰ (The same is of course true of utilitarian conceptions, which sometimes appeal to the diminishing marginal utility of wealth to accommodate sufficientarian judgements.) The individual benefits to those with ample resources may do little to benefit each, but when aggregated, they outweigh a substantial gain to someone poorly off.

The worry can be brought out by revisiting the earlier claim that we lack a moral imperative to favour the rich at the expense of the superrich. It seems that the reason why we lack this imperative is because there is a point at which it can be said that someone has enough, at least as far as special distributive obligations are concerned. In turn, if there is a point at which someone can become sufficiently wealthy so as to lose any special redistributive entitlements, then we ought to object to prioritarianism's commitment to favouring the interests of (potentially many) individuals above this point against the interests of those below this point to whom we do have special distributive obligations.Footnote ⁴¹

Sufficientarianism, or at least those versions that assign lexical priority to the interests of those below the threshold, is immune to the tyranny of aggregation. Gains to those below the threshold unconditionally trump gains to those above the threshold. Yet the very features of sufficientarianism that inoculate it against the tyranny of aggregation seem to render it vulnerable to a crucial objection, what we can call the tyranny of disaggregation: sufficientarianism implies a duty to channel all (or a huge amount) of the surplus resources held by individuals above the sufficiency threshold to better slightly the condition of one individual below the threshold. The tyranny of disaggregation brings out the opportunity costs incurred by those whose interests are not assigned priority. Aggregative principles impose excessive costs on those from whom the moral calculus demands sacrifice, but non-aggregative principles can impose major costs on those who would have gained handsomely from aggregation.

This is a persistent objection to sufficientarianism, an objection sometimes considered decisive. Some sufficientarians find the imposition of huge burdens on those who command sufficient resources acceptable.Footnote ⁴² I disagree. To appreciate just how disconcerting this possibility is, consider Peter Singer's formulation:

Singer seems right: that is an absurdity. So I'll try to address the objection rather than concede it.

The sufficientarian theory of distributive justice I have been developing is not vulnerable to this criticism because sufficientarianism will only allow for the tyranny of disaggregation when understood as a theory of allocative justice. According to Rawls's account, the problem of distributive justice is this:

Two related differences between distributive justice and allocative justice are relevant for present purposes. First, the focus of distributive justice is dynamic and its subject is the structure of social cooperation over time; the focus of allocative justice is static and its subject is the allocation of a given bundle of goods at a particular time. Second, a theory of distributive justice attends to the production and distribution of goods, whereas a theory of allocative justice attends only to the latter.

Both defenders and critics of sufficientarianism tend to interpret sufficiency as a criterion of allocative justice. For this reason, the tyranny of disaggregation objection may be on target with respect to some of the prevailing sufficientarian conceptions. Defenders such as Frankfurt and Crisp focus on cases of allocative justice, such as how to allocate a fixed amount of food to a starving population or fine wine to wealthy individuals;Footnote ⁴⁵ similarly, critics like Arneson and Casal discuss cases such as the allocation of medical resources to dying patients and donations for disaster relief.Footnote ⁴⁶

Missing in all of these examples is a discussion of the conditions that facilitate the cooperation necessary to produce the goods in question. The tyranny of disaggregation objection enjoys intuitive plausibility when we restrict our focus to snapshots of social cooperation. It implicitly treats the supply of goods as given and, as a related point, disregards the effects of today's reallocation on tomorrow's supply of goods.

If sufficientarianism's concern were allocative, i.e. to produce a pattern of holdings that distributes a given bundle of goods to optimize the socio-economic condition of those below the sufficiency threshold at a particular time, perhaps the theory would imply an obligation to channel all the surplus resources held by those above the threshold to a single person below the threshold. Yet I propose that we understand sufficientarianism as a theory of distributive justice for society's basic structure; as such, it concerns the structure of social cooperation and terms of production over time. Thus, the relevant question is whether a policy of channelling away all of the surplus resources held by those above the threshold, e.g. by instituting a 100 per cent marginal tax rate on all income above the threshold, is apt to realize sufficientarian aims better than alternative policies.

Although determining the optimal rate of taxation is an empirical matter, there are undoubtedly feasible institutional alternatives that better realize sufficientarian goals than this one. Optimizing the socio-economic condition of individuals below the threshold will require a healthy labour supply. Implementing a 100 per cent marginal tax rate on all income above the threshold (along with an entitlement up to the threshold level) is a recipe for enervating the supply of labour.

The decreasing supply of labour under increasing rates of taxation is a constraint that theories of distributive justice, but not allocative justice, must accommodate. A sufficientarian theory of distributive justice will be sensitive to the conditions conducive to the production of the goods needed to improve the socio-economic condition of individuals below the threshold – to the likely incentive and expectation effects generated by alternative institutional arrangements. A policy that would channel all the surplus goods held by citizens who have attained sufficiency to one citizen would decrease the number of goods available for everyone by decreasing the number of goods produced. This cooperative scheme would in turn allow more people to fall below the threshold (or do worse by those below the threshold) than alternative schemes. Thus, this scheme would fail by sufficientarian standards themselves.

III

An important worry about the above reply remains.Footnote ⁴⁷ It seems as though sufficientarianism fails as a fundamental principle if there are any conditions in which it yields intuitively unacceptable consequences – even if the conditions are highly unlikely given facts about human psychology and economic behaviour. The proposed counterexample thus defeats sufficientarianism.

One response is to follow sufficientarians like Martha Nussbaum and Elizabeth Anderson in denying the sufficiency principle lexical priority.Footnote ⁴⁸ Perhaps sufficiently large gains in other values can outweigh gains in sufficiency (which receives extra weight). This reply leaves the distinctive elements of sufficientarianism intact: there remains a principled and determinate sufficiency threshold and benefiting individuals below this threshold has priority over competing aims. Moreover, this revision would not diminish the case for sufficientarianism relative to its rivals, because they are likely to require similar revisions.Footnote ⁴⁹

Sufficientarians might also resist the apparent counterexample: it appears decisive at first blush, but not upon further reflection.Footnote ⁵⁰ They can appeal to the tradition in moral theory which holds that we should not trust our intuitions about cases in which features that are typically present in our moral experience and relevant to our moral judgement – for example, incentive effects – are stipulated to be absent or to function in highly atypical ways.Footnote ⁵¹

In their recent work on the neural substrates and normative significance of moral intuition, James Woodward and John Allman emphasize that much moral knowledge is tacit, learned from lived experience.Footnote ⁵² Decision-making in social contexts is frequently driven by implicit learning. This sort of learning can cause us to experience an appropriate intuitive response in the relevant context, but without explicit reasoning or even awareness of the processes that produce the intuitive response or the factors to which it is responding.Footnote ⁵³ Our intuitive judgement is often implicitly informed by relevant considerations. Thus, we frequently lack explicit access to the factors shaping our intuitive judgement, and so we should be wary of examples that require us to explicitly adjust our judgement for the absence of these factors.

This reply resembles a common utilitarian reply to exceptional counterexamples.Footnote ⁵⁴ So before addressing the apparent counterexample to sufficientarianism, it will be useful to consider a similar counterexample to utilitarianism. Woodward and Allman discuss Bernard Williams's example in which Pedro tells Jim that he plans to execute twenty randomly selected villagers in reprisal for their anti-government activity. If Jim shoots one himself, Pedro will spare the other nineteen; if Jim refuses, Pedro will kill all twenty.Footnote ⁵⁵

Here the act-utilitarian recommendation to kill the one may seem counterintuitive. However, it's not clear that we should trust our intuitive judgement about this case, at least as it is commonly described. Woodward and Allman draw our attention to some of the features the scenario must stipulate away to compel the act-utilitarian to endorse the killing of the innocent person. We must be assured that Pedro – a man who proposes to execute twenty randomly selected villagers – will not break his promise to refrain from further killing, that capitulation will not embolden Pedro to issue more of these threats in the future, and so on. The problem is that these sorts of considerations are typically salient features of real-world moral decision-making. Our typical judgements about how to respond appropriately to people and their actions are informed by a range of factors that include the likely incentive and expectation effects produced by our response.

Because the influence of these factors on our intuitive judgement is often tacit, learned implicitly from our ordinary experience, we should distrust our intuitions about how to react to Pedro's offer in the stylized example. The example instructs us not to respond to many of the factors to which our moral intuition has been trained to respond over the course of our lives. Attempts to deliberately adjust for the absence of these factors are problematic given that we often lack explicit awareness of their influence on our moral judgement. Woodward and Allman write:

When learning is a result of feedback processes rather than explicit deliberation, we typically lack the kind of introspective access that would allow us to isolate and assess only those features of a situation to which we wish to respond.

Turning now to the apparent counterexample to sufficientarianism, dramatic increases in the marginal tax rate typically produce deleterious incentive effects. And as noted, incentives are a pervasive feature of real-world decision-making; indeed, evidence shows that people's ordinary intuitions about how income should be distributed are sensitive to the likely incentive effects of the distributive policy.Footnote ⁵⁷

Thus, we have grounds for distrusting our intuitive evaluations of a distributive principle's implications for circumstances in which the response to incentives differs dramatically from the response experience has conditioned us to expect. Because there is reason to doubt our ability to explicitly adjust our intuitions to control for this difference, we should suspend judgement about whether the counterexample is successful. Woodward and Allman argue that we should be sceptical of intuitions about

We shouldn't be unnerved if a principle has apparently counterintuitive implications when psychological or economic conditions differ dramatically from the conditions to which our intuitions are adapted.Footnote ⁵⁹ Perhaps strange implications are appropriate for these strange conditions.Footnote ⁶⁰ Perhaps not. It seems sensible to suspend judgement about the matter.Footnote ⁶¹

In closing, I have attempted to substantiate sufficientarianism by reconstructing it as a criterion of distributive justice for society's basic structure. This revised account appears capable of defeating the outstanding objections lodged against the theory. Thus, sufficientarianism emerges as a serious competitor to more developed theories of distributive justice.Footnote ⁶²