Limitarianism: Pattern, Principle, or Presumption?

In this article, I assess the prospects for the limitarian thesis that someone has too much wealth if they exceed a speci ﬁ c wealth threshold. Limitarianism claims that there are good political and/or ethical reasons to prevent people from having such ‘ surplus wealth ’ , for example, because it has no moral value for the holder or because allowing people to have surplus wealth has less moral value than redistributing it. Drawing on recent literature on distributive justice, I defend two types of limitarian principles of justice. First, limitarian midlevel principles draw on the limitarian thesis to specify normative commitments for guiding institutional design and individual actions. Second, the limitarian presumption draws on that thesis to specify what a just allocation of wealth requires under epistemic constraints. Such a presumption says that without substantive reasons to the contrary, we should regard a distribution as unjust if some people ’ s wealth exceeds the limitarian threshold. Furthermore, I will argue that we must reject a possible but implausible interpretation of limitarianism as an ideal distributive pattern. Yet both as a midlevel principle and as a presumption, limitarianism can play an important role in theorizing about justice in the real world. prioritarianism, suf ﬁ cientarianism, and other distributive views could all accept the limitarian presumption when thinking about distributive justice in nonideal circumstances. Yet that is not an objection to presumptive limitarianism; at most, what is shows is that presumptive limitarianism, much like limitarian midlevel principles, can be defended from a variety of different perspectives. That only strengthens the prospects for limitarianism in distributive justice.


Introduction
In this article, I assess the prospects for the limitarian thesis that there is some wealth threshold, the 'limitarian threshold', such that someone has too much wealth if they exceed that threshold. 1 Drawing on recent literature on distributive justice, I defend two types of limitarian principles of justice. 2 First, limitarian midlevel principles draw on the limitarian thesis to specify normative commitments for guiding institutional design and individual actions. Second, the limitarian presumption draws on that thesis to specify what a just allocation of wealth requires under epistemic constraints. I will argue in favour of both limitarian midlevel principles and the limitarian presumption.
This article is structured as follows. After introducing limitarianism and the arguments supporting it (Section 2), I will first argue that we must reject a possible but implausible interpretation of limitarianism as an ideal distributive pattern (Section 3). I then argue in favour of two types of nonideal limitarianism, namely limitarian midlevel principles (Section 4) and the limitarian presumption (Section 5). I end by reflecting on the role of limitarianism in distributive justice (Section 6).

Limitarianism and Surplus Wealth
Ingrid Robeyns recently coined the term limitarianism and argued that it has a place in thinking about the demands of distributive justice. 3 She defines the view as follows: Limitarianism advocates that it is not morally permissible to have more resources than are needed to fully flourish in life. Limitarianism views having riches or wealth to be the state in which one has more resources than are needed and claims that, in such a case, one has too much, morally speaking. 4 At the heart of this defence of limitarianism lies what we may call the flourishing claim. This is the claim that above some wealth threshold having more wealth does not contribute to one's flourishing and therefore has 'zero moral weight' 5 . We have reasons to redistribute such 'surplus wealth' if that promotes some morally valuable aim(s), such as political equality or eradicating poverty.
Yet limitarianism need not commit itself to this flourishing threshold. The limitarian threshold could also signal, say, sufficiency in some other metric of advantage, or the level of the threshold could be set by investigating when allowing people to accumulate more wealth upsets some important normative concern, such as political equality or equality of opportunity. 6 Therefore, the crucial limitarian claim is that there are good political and/or ethical reasons to prevent people from having more than a certain amount of wealth. In short, limitarianism claims that people should not have surplus wealth.
The claim that people should not have surplus wealth can be justified on at least three different grounds. I will spell them out explicitly because limitarians need not tie their case too closely to one particular reason. And even those who reject one or two reasons for why surplus wealth should be redistributed might still be drawn to limitarianism because of the other reason, which broadens the scope of limitarian theorizing.
The first reason to redistribute surplus wealth could be that it has zero moral value, which simply means that nothing morally valuable can be gained from having it. On this view, all other things being equal, a world in which some people have surplus wealth is not preferable over a world in which no one has surplus wealth. I take it that this is why Robeyns says that surplus wealth has zero moral weight, for example, when she says that the 'argument for urgent unmet needs [see below, DT] is based on the premise that the value of surplus income is morally insignificant for the holder of that income' 7 .
The second reason to redistribute surplus wealth could be that it has moral value but that this value is lexically outweighed by some other normative concern(s). This does not deny that something morally valuable can be gained from having surplus wealth, nor that, all else being equal, sometimes people should be allowed to have surplus wealth. But whatever can be gained from having surplus wealth is less valuable, morally speaking, than other normative concerns.
The third reason to redistribute surplus wealth could be that in practice allowing people to have surplus wealth is less important, morally speaking, than other normative concerns; yet, at least in theory, allowing people to have surplus wealth could outweigh those concerns. For instance, someone might prefer a distribution in which one person lives in poverty but all others have surplus wealth over a distribution in which everyone lives just above the poverty threshold. This would conflict with the limitarian thesis that someone has too much wealth if they exceed the limitarian threshold. But even if one holds such a view, in the actual world so many people are below the poverty threshold that the reasons for allowing people to have surplus wealth are simply outweighed by the reasons for redistributing it.
Robeyns gives three reasons why people should not have surplus wealth. 8 The democratic argument states that extreme wealth undermines political equality and fairness in democratic procedures. 9 The needs argument states that extreme wealth should be used to meet people's urgent needs, such as by lifting them from poverty or by financing solutions to urgent collective action problems. 10 And according to the ecological argument, the wealth of the superrich should be used to finance climate mitigation and adaptation. 11 This article asks the following question: if we are concerned with political equality, meeting urgent needs, and disruptive climate change, does this justify the limitarian thesis in distributive justice that someone has too much wealth if they exceed the limitarian threshold?
Robeyns defends limitarianism in nonideal circumstances, taking the current distribution of wealth as her starting point. 12 However, Robeyns' initial formulation of limitarianism leaves open what kind of principle it is exactly. This calls for further elaboration because, as I will argue below, not all interpretations of limitarianism are equally plausible and each of them has different implications. I distinguish three ways in which limitarianism can be interpreted as a principle of justice: it can be seen as (i) a distributive pattern, (ii) a midlevel principle, or (iii) a presumption. In what follows, I will assess the prospects for limitarianism in distributive justice and argue in favour of limitarian midlevel principles and the limitarian presumption.

Limitarianism as an Ideal Distributive Pattern
We must first examine a possible but implausible interpretation of limitarianism, which I will refer to as ideal pattern limitarianism. Despite this interpretation being implausible and, to the best of my knowledge, not having any defenders, assessing that view serves two purposes: it shows why we must not be tempted to (uncharitably) interpret limitarianism as an ideal distributive pattern, and it will prove valuable later on to show why the objections to such ideal limitarianism do not apply to limitarianism as a nonideal view. 13 Ideal patterns specify what distribution of valuable goods must be achieved or pursued in a just society. In this debate, the main contenders are egalitarianism, prioritarianism, and sufficientarianism. 14 If limitarianism is interpreted along those lines, it claims that in an ideal world people should not exceed the limitarian threshold. We can interpret such ideal limitarianism as an all-things-considered view according to which it is always unjust if people exceed the limitarian threshold; or as a pro tanto view according to which distributions in which some people exceed the limitarian threshold are in at least one respect less just than distributions in which people do not exceed that threshold.
However, we must reject both interpretations of ideal pattern limitarianism. Limitarianism only claims that it is unjust to have surplus wealth under nonideal conditions, which includes, for example, the fact that the current distribution of wealth is vastly unequal, that the superrich have objectionably more political power than others, and that millions of people around the world live in extreme poverty. Limitarianism claims that having surplus wealth only becomes objectionable if we combine the idea of surplus wealth as having zero moral value or less moral value than other moral concerns with the circumstances in which we find ourselves.
Yet none of the interpretations of the moral value of surplus wealth by itself imply that people should not have such wealth; and so, in ideal circumstances, people should be allowed to have surplus wealth. This is why we must reject ideal pattern limitarianism. There is nothing unjust about a distribution in which all normative concerns are met and some people exceed the limitarian threshold. Moreover, if surplus wealth has moral value for the holder, they may in fact be morally entitled to surplus wealth provided those normative concerns are met. But limitarian views are nonideal views that only apply under specific conditions. And so, those kinds of possible distributions do not count against limitarianism because in those distributions the conditions under which limitarianism applies do not hold.
Therefore, ideal pattern limitarianism should be rejected. However, that does not mean that we should reject the pursuit of limitarian distributions. I will argue that extreme wealth can only be just if we leave aside important nonideal considerations. Limitarian midlevel principles and limitarian presumptions, which are two different ways to unfold limitarianism in a nonideal form, do take such considerations into account. They both say that in our world and possible worlds similar to it we have good reasons to defend limitarianism despite the fact that, in an ideal world, limitarianism cannot be justified. In what follows, I will discuss those specifications of limitarianism in turn.

Limitarianism as a Midlevel Principle
If limitarianism is interpreted as a midlevel principle, it claims the following: 15 Limitarian midlevel principle: no one should have wealth that exceeds the limitarian threshold.
Midlevel principles are moral principles that connect 'theory' and 'circumstance'. By theory, I mean normative foundations, such as the greatest happiness principle, a conception of autonomy, a notion of moral equality, or some procedural conception of justice. By circumstance I mean the specific policies, rules, institutions, and individual actions that characterize the status quo. The reasons adduced in defence of limitarianism, such as the democratic argument, the needs argument, the ecological argument, and the account of flourishing, can be understood as arguing in favour of limitarian midlevel principles in circumstances characterized by wealth inequality, unequal political power, extreme poverty, and disruptive climate change.
Limitarianism can draw support from what Cass Sunstein labels 'incompletely theorized agreement' 16 in which agreement exists on specific propositions or outcomes, but there is no agreement on the general theory that accounts for it. Both sufficientarians and prioritarians, for instance, can agree that justice requires the eradication of poverty and support policies and institutions which aim to do so, including limitarian policies. However, for sufficientarians the ground for such limitarianism is that the poor live below the sufficiency threshold; yet prioritarians supports limitarianism because the poor have weighted priority. Limitarian midlevel principles bypass such foundational disagreement and enable agreement about normative commitments in specific cases.
Midlevel principles specify pro tanto commitments that must be carefully balanced in light of other normative commitments and the particulars of specific cases. 17 Such principles must be assessed in light of the ability of the state to administer and enforce the policies, rules, and institutions they promote, their likely incentivizing effects, concerns about efficiency, effectivity, and public support, trade-offs with other midlevel principles, and so forth. 18 To illustrate, Marc Fleurbaey claims that 'imposing a 100 percent marginal tax rate [is] a recipe for economic collapse'. 19 If this is obviously true and clearly so for those theorizing about what justice requires, limitarian midlevel principles are unlikely to be a valuable contribution to thinking about, say, institutional schemes that optimally promote justice in income taxation (assuming that limitarianism indeed proposes a 100 percent marginal tax rate). I do not think this is obviously true at all. But even if limitarian midlevel principles would seriously hamper economic activity, such principles can still serve as a frame to shift the Overton window, and they might still move the superrich to act for limitarian reasons. 20 However, one might object that defending limitarian midlevel principles only pushes back the problem of justifying limitarianism. 21 There are two types of cases we might imagine when considering the possibility of an incompletely theorized agreement on limitarianism. The first involve proponents of different perspectives who are considering whether to converge upon a single shared conception of limitarianism. Here I have this first type of cases in mind. But another type of cases is relevant as well, namely if proponents of limitarianism disagree about what form the limitarian threshold should take. For instance, some might defend higher thresholds than others, or defend limitarian principles to guide institutions but not individual agents. However, one might question what good it is to converge upon limitarianism as a midlevel principle if there is disagreement as to what form such a principle should take in practice.
In response, note that even if there is disagreement about the exact limitarian threshold, different proponents of limitarian midlevel principles can still agree on procedures to determine that threshold, such as by voting or consulting experts. And they may prefer such a threshold over having no wealth limit at all, even if the threshold they agree upon is different from what they regard as the best threshold. The need for such agreement is simply a feature of the context in which limitarian midlevel principles are deployed. However, and importantly, there may be less disagreement about what form limitarianism should take in some important cases. Let me discuss two such cases, drawing on Robeyns' needs argument and ecological argument, to show how limitarian midlevel principles can inform institutional design and individual actions.
The needs argument states that surplus wealth should be used to meet people's urgent needs. This argument is not really controversial. Many people, for example, including egalitarians, prioritarians, and sufficientarians, believe that we have strong normative reasons to eradicate poverty. 22 And following Peter Singer's canonical work on this topic, effective altruists have argued for this claim for a long time. 23 They all agree that those who possess wealth above some high threshold have specific duties to eradicate poverty, even though they disagree about what gives rise to those duties, whether they are ethical and/or moral duties, or whether these duties should be discharged through governmental policies or individual actions. Importantly, it is not because egalitarians, prioritarians, sufficientarians, and others attach value to the limitarian threshold per se that they can agree that those who have wealth that exceeds that threshold have special moral obligations. In the context of poverty alleviation, then, limitarian midlevel principles can inform institutional design and individual actions.
According to the ecological argument, we must use surplus wealth to help address climate mitigation and climate adaptation. 24 First, the rich are responsible for a disproportionate amount of emissions compared to others and therefore have greater individual responsibility to combat dangerous climate change. Second, the industries that have allowed people to accumulate vast amounts of wealth, such as the oil industry, are often carbon intensive. Designing institutions in such a way that the superrich are responsible for a significant part of the costs of climate mitigation and adaptation can therefore be regarded as compensation for negative externalities. Third, at least some of the wealth of the superrich comes from subsidized industries that are harmful to the environment. Taken together, these three reasons, according to Robeyns, justify limitarianism in this context. And thus, when thinking about policies in the context of climate change, those who agree with these reasons can all adopt a limitarian midlevel principle in that specific context.
Hence, limitarian midlevel principles aim to bridge the gap between theory and circumstance by saying that when theorizing about what justice requires in specific circumstances, there is a pro tanto claim that no one should have wealth that exceeds the limitarian threshold. And as such a principle, the limitarian thesis can be defended in distributive justice.

Limitarianism as a Presumption
Limitarians who claim that there is a presumption in favour of limitarianism endorse the following definition: Presumptive limitarianism: without substantive reasons to the contrary, we have reasons to regard a distribution as unjust if some people's wealth exceeds the limitarian threshold.
I will argue in favour of such presumptive limitarianism in distributive justice. More precisely, justice requires a limitarian distribution of wealth unless we have substantive grounds to think otherwise. I will give three arguments for this. First, the widely held 'presumption in favour of equality' grounds a derivative 'presumption of limitarianism'. 25 Second, the idea of surplus wealth grounds presumptive limitarianism. And third, presumptive limitarianism can be derived from moral concerns such as political equality and meeting urgent needs if we factor in epistemic constraints.
Let me first clarify what a 'presumption' entails. A presumption is a risk-averse principles that aims to minimize the possible harm of a decision given the prior beliefs and evidence available to the decision-maker. Presumptions are often mistaken for substantive principles, but it is crucial to recognize the differences between them. 26 Substantive principles, such as ideal distributive patterns, tell us what we must do on the assumption that we know the relevant facts. But presumptions tell us how to act in the absence of knowledge about those facts. We can compare presumptions in distributive justice with the presumption of innocence in legal theory and the precautionary principle in environmental ethics and policy. The presumption of innocence tells us to treat someone as if they are innocent until they are proven guilty. And the precautionary principle tells us how to weigh different options in the absence of decisive evidence about what they will bring about. Similarly, presumptions in distributive justice tell us what distributive justice requires in the absence of substantive grounds to favour specific distributions.

Presumptive limitarianism and the presumption of equality
The limitarian presumption can be derived from the egalitarian presumption. Let me illustrate the egalitarian presumption with an example. 27 Suppose Jesse wants to distribute some valuable goods between Adam and Eve depending on who of them writes the longest poem. Unfortunately, however, the poems get lost before Jesse can read them, and there is no way for him to tell whether Adam or Eve drafted the longest poem. Given this uncertainty, Jesse decides to distribute the valuable goods evenly between them. This is not because he believes that they are equally deserving of itthat is, after all, something Jesse cannot know without reading the poems. In fact, he might believe that they are not equally deserving. But in the absence of the relevant information, it seems most just for Jesse to presume that Adam and Eve are equally deserving. This is the egalitarian presumption in distributive justice. Now suppose Jesse distributes valuable goods between Adam and Eve according to some substantive moral principle, such as a conception of 'desert' or 'weighted priority'. Again, however, Jesse lacks information about the extent to which Adam and Eve meet that criterion. Now consider the following distributions between Adam and Eve: On the assumption that Jesse lacks knowledge about how many goods Adam and Eva are entitled to on substantive grounds, the egalitarian presumption favours distribution A. In A, Adam and Eve can both at most be overpaid two goods or underpaid two goods. In contrast, in B and C, they can be overpaid or underpaid up to three goods. And in D and E, they can be overpaid or underpaid up to four goods. Following the presumption of equality, then, A is most risk averse, B and C are less risk averse than A but more risk averse than D and E, and D and E are least risk averse (or most risk tolerant). Because of this, it is presumptively just, according to the presumption of equality, to distribute the valuable goods equally between Adam and Eve.
If we now consider the distribution of wealth rather than of generic valuable goods, the presumption of equality holds that people should have equal amounts of wealth unless we have substantive reasons suggesting otherwise. In general, the larger Adam's share of wealth relative to Eve's share, the less just Adam's share is likely to be. This supports presumptive limitarianism by implication. Presumptive limitarianism is likely to reduce or at least constrain objectionable inequality by setting an upper threshold on how much wealth people can have.
The presumption of limitarianism is less demanding than the presumption of equality. This is because presumptive limitarianism specifies a broader range of possible distributions that are equally just. If, for example, the limitarian threshold deems that having four valuable goods or more is unjust, then, unlike the presumption of equality, it is agnostic between distributions A, B, and C. The presumption of equality, then, grounds a derivative presumption of limitarianism. But the relation is not biconditional: one can endorse presumptive limitarianism without endorsing the presumption of equality.
Alternatively, we can also think of presumptive limitarianism as a specification of what the presumption of equality requires. Presumptive limitarianism specifies what justice requires in the distribution of wealth specifically. But this is compatible with endorsing the presumption of equality as the overarching fundamental normative principle. For example, the presumption of equality might require a distribution of primary goods or capabilities that is equal, which implies, when it comes to wealth specifically, that the distribution of wealth must be limitarian.
Hence, the presumption of limitarianism can be defended as an implication of the presumption of equality in distributive justice and/or as a specification of a more fundamental presumption of equality in the context of the distribution of wealth.

Presumptive limitarianism and surplus wealth
The second argument for presumptive limitarianism takes as its point of departure the limitarian claim that some people have surplus wealth. 28 As I argued in Section 2, the idea of surplus wealth can be grounded on three different claims, namely that above some threshold wealth has zero moral value, that it is lexically outweighed by some other normative concern(s), or that, in practice, allowing people to have surplus wealth has less moral value than redistributing it. Those who agree that under one or more of those interpretations some people have surplus wealth must endorse presumptive limitarianism.
Recall distributions C and D.
Adam Eve Let us again assume that Jesse must distribute valuable goods between Adam and Eve but that he lacks the relevant information to distribute those goods on substantive grounds. Furthermore, let us assume that people exceed the limitarian threshold if they have more than three goods. If the distributions are wealth distributions, this means that in C neither Adam nor Eve has surplus wealth and that in D Adam has surplus wealth but Eve does not.
Above I argued that the presumption of equality prefers C over D because C is more equal and that this supports presumptive limitarianism by implication. But we can derive a similar conclusion from the observation that only in C no one possesses surplus wealth. If, as presumptions in distributive justice entertain, a risk-averse distribution is preferable over a risk-tolerant distribution, then a distribution that redistributes surplus wealth is preferable over a distribution that allows people to have surplus wealth. Between C and D, then, C is the most risk-averse distribution because only in C is there no surplus wealth. Therefore, the idea that some people have surplus wealth justifies the limitarian presumption.
One might object here that Adam could be really deserving of four goods, and, because of that, D is preferable over C on substantive grounds. However, if wealth above the limitarian threshold really is surplus wealth, it is difficult to see how someone could be deserving of it, morally speaking. Whatever substantive reasons we have for favouring D over C, if having more than three goods means that one has surplus wealth, those reasons cannot be that Adam is entitled to four goods. Instead, those reasons must be that allowing Adam to have more than three goods has other morally significant benefits. I will come back to this objection in Section 5.4.

Presumptive limitarianism and epistemic constraints
The third argument for presumptive limitarianism is that decision-makers often lack the epistemic grounds to apply substantive principles for distributing wealth fairly. 29 Joseph Heath, for example, argues that substantive principles concerning the distribution of labour income fail to give a plausible account of how labour income must be and is in fact distributed. 30 He concludes that markets are structurally unable to deliver 'just' wages because markets only channel labour to its best employment. And a similar case can be made for other economic resources. In an ideal market, for example, capital too is channelled to its most productive usage, where 'productive' means that it increases a specific conception of welfare.
To give another example, luck egalitarians have long since argued that it is often impossible to know what people's relative advantages and disadvantages are in the real world. This point extends to all proponents of substantive principles that require knowledge about individual's comparative standing to specify what distributive justice requires. As Richard Arneson puts it: the idea that we might adjust our distributive-justice system based on our estimation of persons' overall deservingness or responsibility seems entirely chimerical. Individuals do not display responsibility scores on their foreheads, and the attempt by institutions or individuals to guess at the scores of people they are dealing with would surely dissolve in practice into giving vent to one's prejudices and piques. 31 Hence, although justice is certainly concerned with the distribution of wealth, it is not evident that we know what justice requires regarding that distribution in the actual world on substantive grounds.
However, many people believe that what we do know is what justice more broadly requires. 32 For example, the democratic argument rests on the assumption that justice requires that political equality is secured, and such a commitment to political equality is widely shared. And the needs argument suggests that justice requires that those with urgent needs have priority. If limiting the accumulation of wealth and/or redistributing it promotes those aims, we have presumptive grounds to distribute wealth in such a way that it respects certain limits. And importantly, the democratic argument and the needs argument do not require knowledge about individual persons to specify justice in the allocation of wealth between them. We do not need information about Adam and Eve to specify what presumptive justice in the allocation of wealth between them requires. But, according to the limitarian presumption, what we do know is that a distribution between Adam and Eve in which neither of them exceeds the limitarian threshold is more likely to be compatible with political equality and meeting urgent needs than a distribution in which one of them does exceed that threshold.
Hence, if the democratic argument or the needs argument holds, presumptive limitarianism offers a plausible criterion for distributing wealth if we lack substantive grounds to favour specific distributions. And if the distribution of wealth is indeed such that it is impossible to know whether it tracks substantive principles, or if it is impossibly complex to apply those substantive principles to actual wealth distributions, presumptive limitarianism supports distributions in which people do not exceed the limitarian threshold.

Three objections to presumptive limitarianism
Let me discuss three objections to the limitarian presumption. The first objection is that presumptive limitarianism falls prey to the same objection as ideal pattern limitarianism because it may fail to secure political equality and meeting urgent needs. This is because it seems to neglect possible allocations of wealth that are to the maximum advantage of the lesser off. For example, consider the following two distributions: Adam Eve Distribution C 1 3 Distribution F 2 4 Distributions C and F differ in that the total amount of wealth in each of them is different. In C, neither Adam nor Eve exceeds the limitarian threshold of three goods. In F, however, Eve does exceed that threshold. But in F Adam is better off than in C. So which distribution should we prefer? If presumptive limitarianism renders C more just, it commits itself to the claim that people should not exceed the threshold, yet it does so at the expense of Adam who could be better off. Yet if it renders F more just, it commits itself to a distribution that allows people to exceed the limitarian threshold. This robs presumptive limitarianism of the distinctive limitarian claim that a distribution is unjust if some people exceed the limitarian threshold. Hence, presumptive limitarianism seems implausible here for the same reason as ideal limitarian patterns are implausible.
However, limitarians can say two things in response. The first is that presumptive limitarianism is irrelevant if we have substantive grounds for favouring certain distributions. If we know that redistributing surplus wealth makes those below the limitarian threshold worse off, the presumptive limitarian reason becomes irrelevant. But it is only because we know that Adam is better off in F than in C that we favour F over C. This touches upon a crucial difference between patterns and presumptions. Patterns claim that justice-relevant goals, such as securing political equality and meeting urgent needs, can be met by pursuing a specific pattern. On the contrary, presumptions specify risk-averse principles that aim to minimize the harm of possible misallocations of valuable goods in light of epistemic uncertainty. The claim here is not that presumptive limitarianism leads to the pattern that will optimally promote the justice-relevant goals, but that it is most likely to do so given the epistemic constraints in place. If there are no such epistemic constraints, however, we no longer have to take the presumption into account.
The second response is that we might in fact believe that C is preferable over F, at least presumptively, because Adam might be worse off in F. Distributions C and F only indicate how much wealth Adam and Eve have, and it seems that, from that specific perspective, Adam is worse off in C than in F because in the latter distribution he has more wealth. However, that leaves open whether F leaves Adam worse off in some other morally valuable domain (e.g. social standing, political equality, etc.) despite the fact that he has more wealth in that distribution. Though presumptive limitarianism specifies what a just allocation of wealth requires, the reasons for defending such limitarianism extend beyond a specific concern for the distribution of wealth as such.
The second objection to presumptive limitarianism is that it offers an account of distributive justice that is too minimal and, furthermore, that it is already entailed in other distributive views. Because presumptive limitarianism only focusses on the superrich, it only offers a partial account of a presumptively just distribution. However, it need not exhaust what presumptive justice in the distribution of wealth requires, and it can be combined with other presumptions as well. 33 Furthermore, it may indeed be that egalitarianism, prioritarianism, sufficientarianism, and other distributive views could all accept the limitarian presumption when thinking about distributive justice in nonideal circumstances. Yet that is not an objection to presumptive limitarianism; at most, what is shows is that presumptive limitarianism, much like limitarian midlevel principles, can be defended from a variety of different perspectives. That only strengthens the prospects for limitarianism in distributive justice.
The third objection to presumptive limitarianism is that it is redundant because there is always at least some knowledge available to decision-makers to apply substantive principles. However, presumptive limitarianism can play a role in such cases too. For example, suppose justice requires distributing wealth based on the number of hours worked and that Adam works twice as many hours as Eve. Does the fact that we know this mean that Adam is entitled to twice as much wealth as Eve no matter what distribution we end up with? That does not follow. For one thing, it is not evident that the conversion of hours into wealth is such that working twice as many hours entitles one to twice as much wealth. Furthermore, it is not evident that distributing wealth on the basis of that substantive principle must guide the entire wealth distribution. For example, Adam and Eve may already have different levels of wealth, which may have a bearing on justice regarding additional benefits. The substantive principle might offer only a partial specification of justice in the distribution of wealth, in which case presumptive limitarianism holds for the remaining economic resources.
In short, the limitarian presumption can be derived from the presumption of equality, from the idea of surplus wealth, and it can be defended as a risk-averse strategy for distributing wealth given epistemic constraints. Those reasons are not mutually exclusive, of course, and may in fact strengthen each other. Yet each of them provides a distinctive case for the presumption of limitarianism in distributive justice.

Conclusion
The limitarian thesis states that there is a limitarian threshold such that someone has too much wealth if they exceed that threshold. In this article, I have assessed three ways in which the limitarian thesis can be defended in distributive justice, namely as an ideal distributive pattern, as a midlevel principle, and as a presumption. I have argued that limitarianism must be rejected as an ideal principle and that it should be interpreted and developed along nonideal lines instead. More specifically, both as a midlevel principle and as a presumption, limitarianism can play a role in distributive justice. In particular, I have argued that without substantive reasons to the contrary, we have reasons to regard a distribution as unjust if some people's wealth exceeds the limitarian threshold. And given the current disparities in income and wealth between the rich and the poor, and in light of the accumulation of wealth in the hands of a small global elite, limitarianism can play an important role at that. 34 Dick Timmer, Department of Philosophy and Religious Studies, Utrecht University, Janskerkhof 13, Utrecht, 3512 BL, the Netherlands. k.d.timmer@uu.nl to refer to the bundle of economic resources an individual possesses. I will focus on economic limitarianism. However, limitarianism is also applicable to other valuable goods, such as emissions or natural resources.