6. The Limits of Limitarianism

Robert Huseby

1. Introduction

Limitarianism is a view of distributive justice according to which there is an upper limit, or threshold, to the amount of goods (for instance money, resources, welfare, or capabilities) that people can permissibly have.1 This view is interesting, though somewhat perplexing, since most accounts of distributive justice emphasize what people ought to have, rather than what they ought not to have.2

In this article, I assess limitarianism in both instrumental and intrinsic versions.3 I start by examining Ingrid Robeyns’s instrumental form of limitarianism, since this is the most elaborate and detailed version currently available in the literature. It emerges, however, that this theory, while in many ways plausible, is not limitarian as such. It might well have limitarian-like implications when applied under a certain range of non-ideal conditions, but it does not amount to real instrumental limitarianism. Rather, it is a view best seen as a combination of instrumental egalitarianism and sufficientarianism.4

Further, and more generally, I argue that it is hard to envision a compelling version of intrinsic limitarianism.5 Robeyns suggests (but explicitly refrains from endorsing) some possibilities, based on virtue, paternalism, and perfectionism. None of these holds much promise, however, primarily because they do not amount to intrinsic limitarianism. Rather, they appear to be based (unsurprisingly) on the intrinsic value of virtue, paternalism, and perfectionism, respectively. There might be values—for instance some virtues—that have limitarian-like features, but these are most likely of limited value to the debate on distributive justice. I also consider another possibility based more directly on the (supposed) badness of having too much. This version too turns out to be unconvincing.

Next, I return to instrumental limitarian views, and consider Danielle Zwarthoed’s proposal for an instrumental limitarianism based on the value of personal autonomy. This account also, like Robeyns’s, is in my view best understood in light of a non-limitarian principle (sufficientarianism).6 Thus, the so-far leading attempts at defending instrumental limitarianism do not seem to work qua limitarianism. In addition, any form of instrumental limitarianism, must rely on some more fundamental, intrinsic value. Since I raise serious doubts as to the possibility of there being plausible forms of intrinsic limitarianism (see above), this intrinsic value must be something else—for instance, egalitarianism, sufficientarianism, or autonomy. If these claims are sound, it is not clear that we need limitarianism as a distinct and independent principle of distributive justice.

Further, I assess the suggestion that limitarianism might be valuable as a more limited principle, as a ‘presumption’ to be relied on under certain epistemic conditions, rather than as a more general (intrinsic or instrumental) principle of distributive justice. I also briefly consider the idea that limitarianism might have value as a more impactful principle than its contenders. I argue that neither of these limitarian versions are likely to be particularly helpful.

My overall conclusion, then, is that limitarianism is of very limited value to debates on distributive justice. I should nonetheless underline that not all possible forms of (instrumental and intrinsic) limitarianism can be assessed here, and that there could be other versions that are either more plausible, or more independent, than the ones I consider.7 Nevertheless, the conclusion, such as it is, does not bode well for limitarianism.

2. Robeyns’s Instrumental Limitarianism

Limitarianism is in general characterized by the claim that “it is not morally permissible to have more resources than are needed to fully flourish in life.”8 Full flourishing is not specified in detail, but refers to having at least adequate capabilities in certain central respects. Beyond this level, it is morally impermissible to have surplus money, wealth, or other financial resources.9 Robeyns finds it “surprising that so little (if any) contemporary theorizing on justice has focused on the upper tail of income and wealth distribution.”10 Limitarianism is a response to this lack of normative research on the best off.

Robeyns’s specific version of limitarianism is non-ideal. It does not constitute an answer to the question of what justice requires “in a world with strong idealized properties.”11 It is also instrumental, because limitarianism is not seen as intrinsically valuable, but rather as necessary in our non-ideal world to realize “two intrinsic values: political equality … and the meeting of unmet urgent needs.”12 Robeyns’s defense of limitarianism thus consists in two distinct arguments, according to which the realization of these two intrinsic values require limitarian principles and policies in our non-ideal world.13 In the following, I assess these two arguments in turn.

A. The Argument from Political Equality

According to the first argument, limitarianism is necessary in order to secure political equality because “great inequalities in income and wealth undermine the value of democracy and the ideal of political equality in particular.”14 Referring to Christiano, Robeyns mentions various mechanisms through which money can be turned into political power. These include vote buying, gatekeeping, influencing of opinion, and money as an independent political power.15 On Robeyns’s view,

[t]he democratic argument for limitarianism can easily be derived from the mechanisms that Christiano outlines: Because rich people have surplus money, they are both very able and seemingly very likely to use that money to acquire political influence and power.16

One might think that legislation and other institutional measures could secure democratic equality, but Robeyns holds that even though such measures are necessary, they will not be sufficient, because

much of the political influence of rich people escapes the workings of formal institutions … Large inequalities in income, and the possession of surplus money in particular, will thus always undermine political equality.17

In this quote, Robeyns mentions both inequality and surplus money as challenges for political equality.18 But it is not clear why surplus money should be a problem. Suppose everyone was equally super-rich. Or, suppose everyone was equally wealthy and just slightly above the limitarian threshold. The surplus wealth in these scenarios does not appear to threaten political equality at all. Of course, both cases might be seen as ideal, and thus fall outside the scope of Robeyns’s theory. Consider instead a non-ideal (though stylized) world in which 10 percent of the population is (equally) super-rich, and the remaining 90 percent lack surplus money altogether. It is easy to see that the inequality between the two groups can endanger political equality. However, if surplus money “in particular” is the problem, we would have reason to be almost as worried about political equality within the group of the equally super-rich. This does not seem plausible.19 Thus, it must be the inequality that is the root problem.

It is therefore not clear why the concern for political equality provides instrumental reasons for limitarian policies. Of course, limitarian policies will increase equality in our actual non-ideal and unequal world, and this will in turn most likely increase political equality. However, the same is arguably true of sufficientarian and prioritarian policies as well. And, under many circumstances, it will also be true of utilitarian policies. Thus, the goal of political equality provides no grounds for limitarian policies in particular, and instrumental limitarianism is thus not necessary for political equality.

Instrumental limitarianism is also not sufficient for political equality, since Robeyns’s limitarianism accepts differences in riches below the threshold.20 Even if no one has surplus money, and people at the threshold (of full flourishing) thus have much less political influence than they would have had, had they been rich or super-rich, they have more political influence, at least potentially, than those whose most urgent needs are just barely met.

To be sure, it could be the case that none of the most well-known distributive principles, in instrumental renderings, is both necessary and sufficient to secure political equality. Even complete economic equality (at the level of full flourishing) might be compatible with some political inequality, just because people prioritize political influence differently, and some might even flourish through their political activity, whereas others might not.21 Thus, it is not an effective criticism to point out that limitarianism, (perhaps) like all other principles, is neither necessary nor sufficient to secure political equality.

However, as I have argued (and as Robeyns suggests at times), the core of the problem seems to be that inequalities of economic means (rather than surplus money as such) in the distributive sphere cause morally offensive inequalities in the political sphere.22 If so, the problem is really caused by a discrepancy in the holding of economic means between individuals. In light of this, the most reasonable strategy is to address the discrepancy as such. To do so, one can start at the top end, at the bottom end, or both. Robeyns chooses to start at the top end, without offering arguments for why we should not start at the bottom end (as prioritarians and sufficientarians would suggest),23 or, most plausibly, why we should not start at both ends simultaneously. In my view, instrumental egalitarianism (of money or financial resources) is clearly the most plausible principle given the problem at hand.

To illustrate, suppose the limit for full flourishing is at 10, and that A has 5, and B has 15.24 Limitarianism advocates against supra-limit holdings and implies that the distribution should be A:5 and B:10. This is not to suggest, uncharitably, that limitarianism is incompatible with further distribution below the threshold. The point is just that the theory emphasizes one particular end of the distribution, namely the “upper tail,” which in this case refers to B’s 5 surplus units. If those units are gone, limitarianism has no further implications. (Depending on the extent to which one emphasizes the sufficientarian aspect of the theory).

Something similar can be said about (instrumental) sufficientarianism. Suppose the threshold is at 10 and that there are 5 unowned available units. These units can be given to A, so that A gets 10, and B remains at 15. Sufficientarianism in itself has no further implications (though this view, too, is compatible with further distribution, once sufficiency is secured). Thus, sufficientarianism starts at the bottom end, and limitarianism starts at the top end. But limitarianism allows sub-threshold inequalities, and sufficientarianism allows supra-threshold inequalities.

Egalitarianism, on the other hand, straightforwardly suggests a redistribution of 5 units from B to A, such that they both end up having 10. Again, if, as seems plausible, the inequality is the root of the problem, presenting limitarianism (or sufficientarianism for that matter) as the solution misses the mark. This is so even if limitarianism (and sufficientarianism) will admittedly reduce the problem. The reason is that there is an available alternative, egalitarianism, that is superior, and which will reduce the problem to a greater extent, at least under a very wide range of circumstances.25

There is also a more general problem with the argument from political equality. We can assume that the world is either such that there are enough resources to allow everyone to flourish fully, or such that there are not enough resources to allow everyone to flourish fully.26 Consider the first possibility (which limitarians may reject as an ideal scenario). If there are any surplus riches left once everyone has reached the threshold of full flourishing, limitarianism could either distribute them equally, thus rendering the view egalitarian rather than limitarian, and the argument from political equality redundant, or it could demand the destruction of surplus riches, rendering the view wasteful and implausible.

On the other hand, if there are not enough resources in the world to let everyone flourish fully (which might be more in line with the non-ideal assumption), it seems to follow from Robeyns’s flourishing account that all surplus riches should be spent, in some sufficientarian fashion, to maximize either the incidence or extent of full flourishing.27 Political equality would not play any significant role in this scenario.

Thus, if there are enough resources to let everyone flourish, limitarianism is either implausibly wasteful, or indistinguishable from egalitarianism. If there are not enough resources to let everyone flourish, limitarianism seems indistinguishable from sufficientarianism. The argument from political equality can only play a distinct role, then, if it implies wasting resources in circumstances of plenty. But this is arguably not a very plausible result (and most likely outside the non-ideal scope of limitarianism).

Overall, I do not see how the value of political equality provides support for instrumental limitarianism in particular. Limitarianism is neither necessary nor sufficient for securing political equality, but, more importantly, limitarianism does not appear to be the instrumental distributive principle that is best suited to secure this important value. Rather, some form of instrumental egalitarianism (of money or resources) is the superior alternative.28 In addition, it is hard to see what role political equality could play, in non-ideal circumstances of (moderate) scarcity.

What about the value of meeting urgent needs? As noted, the limitarian threshold, on Robeyns’s view, is located at the level of full flourishing. Once this level is reached, further money or wealth will not, by definition, contribute to further flourishing. (Though it can contribute to increased welfare, capabilities, and forms of flourishing that do not accord with Robeyns’s moralized understanding of the term).

The argument from unmet needs claims that since surplus money does not contribute to people’s flourishing, it has zero moral weight, and it would be unreasonable to reject the principle that we ought to use that money to meet those urgent unmet needs.29

The level at which urgent needs are met, moreover, is significantly lower than the level of full flourishing. Thus, limitarianism (in this version) is a layered view, with two thresholds, one lower and one higher.

Since meeting urgent needs is intrinsically important, and since surplus money or wealth above full flourishing has zero value, it follows that those whose urgent needs are not met must have absolute priority over those who have surplus money. Thus, those who are flourishing fully, with money to spare, should give up their surplus to meet others’ urgent needs. This is, as far as I can see, a version of (two-level) intrinsic sufficientarianism.30

Is instrumental limitarianism the way to go if we want to secure that urgent needs are met? There are some reasons to doubt this. First, the moral impermissibility of having too much cannot in general be defended by reference to the importance of meeting urgent needs. Assume that there is a number of people whose urgent needs are currently unmet, but whose urgent needs could be met, given sufficient financial resources. Suppose also that there is a number of people who have surplus wealth. Call the amount of resources needed to meet the urgent needs X, and the amount of surplus money held by the flourishing Y.

The argument from unmet urgent needs does not imply that it is impermissible to be rich if Y > X. It only implies that it is intrinsically important that urgent needs are met, and that goods above the limitarian level have zero value. One could argue that in our non-ideal world, it is in fact the case that Y < X, but that is quite a bold empirical claim, given the amount of currently available resources.31 According to the argument from unmet urgent needs, then, it is perfectly permissible to be rich so long as all urgent needs are met (Y > X). Thus, this argument for limitarianism does not work.

One response could be that if Y < X, then it is, in effect, impermissible to be rich. But this is merely a coincidence. If, in a given situation, an equal level amounts to a level at or below the limitarian threshold, egalitarianism has limitarian-like implications. It would then be, in effect, impermissible to have wealth above the threshold, because that would necessarily imply inequality. Again, that is just coincidental, and arguably not a theoretically interesting aspect of egalitarianism. The same is true if, in a given situation, sufficientarian, prioritarian, or utilitarian distributions just happen to lead to outcomes in which no one is above the limit. In these cases, each of the principles would imply that it is impermissible to be rich, but here too, the implication would just be the coincidental upshot of the various principles applied to specific circumstances.32

Further, if it is intrinsically valuable that people’s urgent needs are met (which is plausible), and if we want to consider what policies would help realize this value, limitarianism does not strike me as very promising. Instrumental sufficientarianism is more reasonable, since such a view focuses squarely on bringing the deprived up to the relevant threshold rather than (primarily) bringing the rich down to the (limitarian) threshold. Instrumental prioritarianism or egalitarianism also appear better suited to tackle the problem, since the first explicitly prioritizes the worst off, and the second (as noted) is equally concerned with both tails of the distribution. Even instrumental utilitarianism (if there is such a thing) would be more appropriate, given some plausible assumptions about diminishing marginal utility.

Thus, neither of Robeyns’s arguments succeed qua arguments for instrumental limitarianism. Again, this does not imply that her view is implausible. Both political equality and securing a basic minimum are important moral goals. But since I am here interested in how promising limitarianism is as a principle of distributive justice, it is noteworthy that the most elaborate available account of instrumental limitarianism does not really have limitarian implications at all.

3. Intrinsic Limitarianism

On Robeyns’s definition, “[I]ntrinsic limitarianism is the view that being rich is intrinsically bad, whereas according to non-intrinsic limitarianism, riches are morally non-permissible for a reason that refers to some other value.”33 As noted, she refrains from defending intrinsic limitarianism, but she does mention some possible versions, based on virtue, perfectionism, or paternalism. It seems, however, that the prospects for intrinsic limitarianism are quite dim.

A. Paternalism, and so on

Paternalism means interfering with other people’s autonomy without their consent, for their own good.34 In the case of limitarianism, paternalism would presumably mean that some agent, perhaps the state, prevents people from having wealth above the limit, for their own good. If so, excess wealth must be bad for people (something that is suggested by Robeyns and defended by Zwarthoed).35 However, if having riches above the level of full flourishing is bad for people, and this is the reason why paternalistic measures are necessary, then paternalism is merely instrumental. There must then be some other intrinsic value that paternalism is supposed to protect or promote—such as, for instance, welfare (which is one way of understanding “being bad for people”), virtue, autonomy, or some perfectionist value. But then it is one of these ideals that are doing the job, rather than paternalism.

What about the other suggestions? Perhaps having too much can undermine the possibility of realizing certain virtues, or make it hard to achieve other valuable ideals. Moderation, for instance, can be hard to realize in conditions of excess and luxury. More generally, there could be several virtues that are incompatible with excess, and several perfectionist values that are hard to bring about in luxurious circumstances. This does not mean, of course, that any of the possible virtues or perfectionist values are plausible, but at least they could be. However, as was the case with paternalism, there is a more fundamental problem. Even if some of these values can ground a view of justice that sets an upper limit to the level of goods that people can permissibly have, this will not amount to intrinsic forms of limitarianism. The limit would be instrumental in bringing about (the conditions for) virtue, autonomy, perfectionist values, or any other non-limitarian value, just as Robeyns’s own view is (allegedly) instrumental in bringing about the meeting of urgent needs and political equality.

This notwithstanding, there might still be intrinsic values that are themselves such that it is important not to have too much of it, and virtue might be one place to look. Courage, for instance, is a virtue that is best had in optimal amounts. I will not deny the possibility that such values may exist, and that they may exhibit limitarian features.36 Even so, such values are quite different from the forms of limitarianism that are under consideration here, which focus on distributive justice. Again, on Robeyns’s view, intrinsic limitarianism is the view that “being rich is intrinsically bad”37 and not the view that there are some intrinsic values that are best had in moderation. Thus, values with limitarian-like features are not likely to play a significant role in discussions about distributive justice.

What, then, could intrinsic limitarianism be? Consider for illustration intrinsic egalitarianism. On this view, it matters in itself, that people are equally well off in some relevant respect.38 Intrinsic sufficientarianism, on the other hand, holds that it matters in itself that people are sufficiently well off.39 Intrinsic limitarianism would correspondingly be the view that it matters in itself that people are not better off, in some relevant respect, than some upper limit.40

There are two versions of this possibility. The person-affecting version holds that it is (intrinsically) bad for people to have too much money (or welfare, or some other metric). The non-person-affecting version holds that it is (intrinsically) bad if people have too much money (or welfare, or some other metric), irrespective of whether this is (also) bad for them.

Consider first the person-affecting version. On this view, having too much is bad for people, and it is bad intrinsically, in much the same way that it can be bad for people, and intrinsically so, to have too little. Here it is worth keeping in mind that the metric might play a role. Suppose the metric is welfare. It is very hard to see that it could be bad for you, and intrinsically so, to have more welfare. To have more welfare is conceptually very close to something being good for you, and possibly intrinsically so. Quite similarly, it is hard to see that having more, or better, capabilities, can be bad for you, and intrinsically so.

On the other hand, it is plausible that it could sometimes be bad for people to have too much money or wealth (and that this may be intrinsically bad). But there are some worries to consider. First, this is likely to be contextual. It is hard to believe that there is some set limit above which more money or wealth will make people’s lives go worse in some relevant respect, regardless of who these people are, what psychological traits they have, how the distribution is elsewhere in society, where and when they happen to live, and so on. But perhaps this problem could be solved by defining the limit in light of such variation, such that it is bad for a person to have so much money or resources that it is in fact bad for her given her circumstances and individual traits.

There are more important problems with the suggestion, however. If having too much is bad for people, what they have too much of, and what they get less of, must refer to two different things, for instance wealth and welfare.41 It is coherent under some circumstances to say that if you get more of X you get less of Y. But if someone suggested that the more you get of X, the less you (simultaneously) get of X, we would be hard pressed to understand the proposition. Because of this, it is difficult to see that it can be intrinsically bad to have too much. What is intrinsically bad must be that people are getting worse off in some respect. And they cannot become worse off in that respect by becoming better off in the same respect! If they are getting worse off because they have too much money, on the other hand, that just means that money beyond some threshold is instrumentally bad because it makes people’s lives go worse. The intrinsic value must be that people do not fall below some level of, for instance, welfare or flourishing. This suggests, again, sufficientarianism, rather than limitarianism.

Third, even if having too much was intrinsically bad for people (which I, as indicated, doubt), this would not (at least yet) give us sufficient reason to conclude that it is morally impermissible to have too much.42 The distance from the badness to the impermissibility must be covered by a plausible paternalistic argumentative path. Perhaps such arguments could be provided, but it does not seem likely.

What about the non-person-affecting version? On this view, it is in itself bad if people have too much, regardless of whether this is (also) bad for them. This too comes across as an unlikely suggestion. Suppose there are three possible future versions of the world, and that the limitarian threshold is at 10. In the first possible version, everyone is far below the limit, at 4. In the second, everyone is just below the limit, at 9. In the third, everyone is just above the limit, at 11. Intrinsic limitarians would say that, at least in one respect, the last world is the worst of the three. That is, the world in which everyone is best off, and equally so, is worse than a world in which everyone is much worse off, and equally so. It might be worthwhile to consider these cases with different currencies of justice in mind. But regardless of whether we assume that it is resources, welfare, capabilities, or something else that should be distributed, intrinsic non-person affecting limitarianism seems very implausible.43

What is lacking is some convincing rationale for why the limit is where it is (in the example above, the limit is, of course, randomly set), and why it is intrinsically important that people do not have too much. None of this suffices to rule out conclusively the possibility that some plausible version of intrinsic limitarianism could be developed, but I, for one, am not holding my breath.

4. Instrumental Limitarianism More Generally

So far, I have argued that Robeyns’s instrumental limitarianism is not really limitarian, and that intrinsic limitarianism, (quite) generally, holds little promise. This leaves ample room for other forms of instrumental limitarianism. One suggestion has, as noted, been offered by Zwarthoed. On her view, personal autonomy can be undermined by having too much money and financial resources. Thus, if we value autonomy intrinsically, and if it is right that autonomy can be thwarted by having too much, then we have a reason to set an upper limit to how much money or wealth people can have.

However, while having too much might undermine some people’s autonomy, this hardly generalizes.44 More importantly, autonomy generally requires sufficiency of at least resources. People who lack basic necessities will have a very hard time making autonomous decisions, whereas many who have too much are very able to take charge of their own lives. Thus, autonomy robustly requires sufficiency, and only contingently requires limitarianism. It would be difficult, in my view, to justify a general policy of limitarianism with reference to autonomy. It would also be difficult to show how the autonomy-based view could justify the impermissibility of having too much.

Robeyns also mentions several other possible forms of limitarian-inspired policies and institutions, including emissions quotas, quotas for how many children one can have, and capability ceilings defined in terms of one’s actions’ negative impact on others.45 In my view, the two first are more plausibly understood as examples of equal distribution of shares, and the last as an implementation of a version of the no-harm principle. If we think something should be distributed equally, everyone’s share is limited, of course, but this does not mean that the distribution is therefore limitarian. And if I am not allowed to harm you, there are some limits to what I can do. Even so, the no-harm principle is not limitarian in a relevant sense.

Further, as has been emphasized throughout, any form of instrumental limitarianism will necessarily have to rely on some intrinsic value. And that intrinsic value is unlikely to be of a limitarian variety, for reasons spelled out in the preceding section. Thus, limitarianism must be a means to an end. Robeyns’s view is based on equality and sufficiency, and Zwarthoed’s view is based on autonomy. One could imagine versions based on prioritarianism, virtue, or some perfectionist value. The problem is that it can be seen as a distraction to present a view as a novel principle of distributive justice,46 if this principle is in fact just an empirically contingent way of securing some further, and more familiar intrinsic principle. In other words, it is not helpful to present it as a principle alongside a set of other principles, of which there are (presumably) plausible intrinsic versions.

In addition, it is not easy to think of a plausible intrinsic value that is best promoted only by setting an upper threshold to how much people can permissibly have. In the case of political equality, urgent needs, and autonomy, as I have argued, it is at least as important to make sure that people either have equally much or that they reach some threshold (whether or not that coincides with the limitarian threshold). Again, the need for instrumental egalitarianism, sufficientarianism or perhaps prioritarianism will be at least as great as the need for instrumental limitarianism.

Moreover, as indicated above, it is hard to see how one can ground the claim that it is impermissible to have too much.47 To be sure, in light of many principles, it may under some circumstances happen to be impermissible to have too much, given the prevailing distribution. But limitarianism holds that it is impermissible to have too much without reference to the prevailing distribution,48 and without reference to the intrinsic value that limitarianism is supposed to be instrumental in realizing. Admittedly, Robeyns defines instrumental limitarianism like this: “according to non-intrinsic limitarianism, riches are morally non-permissible for a reason that refers to some other value.”49 So it might be wrong to say that instrumental limitarianism claims that it is impermissible to have too much without reference to some other intrinsic value. In other places, however, it sounds as if this is exactly the case: “In a nutshell, limitarianism advocates that it is not morally permissible to have more resources than are needed to fully flourish in life.”50

In any case, there is some tension here. If limitarianism is defined without reference to other values, it is hard to justify the impermissibility claim. Why should it just be impermissible to have too much in our non-ideal world? But if limitarianism is defined with reference to other values, it immediately looks more like a sufficientarian claim. It then says that it is important that people reach some morally relevant threshold—for instance the threshold of full flourishing—and resources beyond that point have zero value. Resources of zero value can be used to realize ideals that have more than zero moral value (meeting urgent needs, say). It is then impermissible to have surplus money whenever there is some more morally worthy cause that the surplus could be used to promote, otherwise not. But this is just what sufficientarianism says (in many versions). Like all sufficientarian views that accept an upper (negative) threshold, it implies that all supra-threshold holdings can be used for other, more weighty distributive goals, such as providing for those who are below the threshold.51 The impermissibility claim is just the implication of this view when applied under a certain range of empirical conditions.

To sum up, any instrumental form of limitarianism is valuable in virtue of some intrinsic value. If so, referring to a distributive principle as “limitarian” is not very helpful, if the relevant limitarianism is instrumental in bringing about, say, egalitarianism, prioritarianism, utilitarianism, sufficientarianism, virtue, autonomy or some perfectionist value. To be sure, instrumental limitarianism could in principle be valuable in virtue of bringing about some form of intrinsic limitarianism, but, as we have seen, it is doubtful that any form of intrinsic limitarianism is plausible. Further, none of the forms of instrumental limitarianism suggested so far, seems convincing, since the intrinsic values on which they rely appear to be better promoted by other instrumental principles. This latter conclusion, of course, applies only to the few views examined, and not to all possible forms of instrumental limitarianism. Even so, the assessment so far provides good reason to doubt that limitarianism has much to add to debates about distributive justice.

5. Limited Forms of Limitarianism

So far I have examined limitarianism as a principle of distributive justice. In doing so, I have taken limitarianism to be a contender with familiar principles and theories such as egalitarianism, prioritarianism, sufficientarianism, and utilitarianism. In my view, this is also the spirit in which (instrumental and non-ideal) limitarianism is presented and defended by Robeyns.52 However, even if my criticisms so far hit the target, it could still be the case that there is room for more limited limitarian principles, perhaps principles that can justify specific policies. Or it could be the case that limitarianism is politically useful in ways the more traditional principles are not.

Dick Timmer has recently defended limitarianism along the former lines.53 He proposes that limitarianism can be defended as a presumption that specifies what a just distribution requires under “epistemic constraints.”54 There are three arguments for this view. According to the first, “the limitarian presumption can be derived from the egalitarian presumption.”55 According to the egalitarian presumption, if we are to distribute goods between individuals in light of some moral principle, such as desert (which is the example Timmer mostly relies on), and we do not know who is more deserving, we should distribute the goods equally, because this reduces our chances of increasing injustice.56 In Timmer’s case (based on an example presented by Juha Räikkä),57 an employer is supposed to allocate 4 units of goods to two workers on the basis of desert. As it happens, the employer lacks information about the workers’ individual contribution to the job. According to the egalitarian presumption, the best way to proceed is to divide the units equally. If both workers receive 2 units, the worst mistake that can be made is that one of them is underpaid 2 units and one is overpaid 2 units. Since the principle to be realized is a desert principle, both underpaying and overpaying are a mistake. If the employer distributes 3 units to one worker and 1 unit to the other, the worst possible mistake is that one worker is overpaid by 3, and the other is underpaid by 3 (and even worse for a 4/0 distribution).58 Thus, in this situation, we should distribute equally if we want to distribute in light of desert, because this will minimize the extent of injustice that we may (inadvertently) cause.

The egalitarian presumption, Timmer claims, “supports presumptive limitarianism by implication. Presumptive limitarianism is likely to reduce or at least constrain objectionable inequality by setting an upper threshold to how much wealth people can have.”59 I am not sure, however, that this is the case. In Timmer’s example, the limitarian threshold is set at 3 units. If 4 units is all that is up for distribution, and 4 is too much, then no one should receive more than 3. Both the egalitarian and the limitarian presumption agree that a 3/1 distribution is preferable to a 4/0 distribution. But they disagree in that the limitarian presumption allows a 3/1 distribution, whereas the egalitarian presumption does not. Thus, it is not correct that the egalitarian presumption implies the limitarian presumption.

Moreover, let us consider contexts that take initial holdings into account. This aspect is not present in Räikkä’s original example, and it is easy to see why. If we want to pay workers for a particular job on the basis of desert in that particular situation, we do not need information about how well off they are initially. To be sure, their initial holdings may, for all we know, be unfair, but we can assume that information on this score is even more inaccessible than information about their relative merits in completing the job. Suppose there are still 4 units to be distributed, that the limitarian threshold is at 10, and that worker A is currently at 6 and worker B is currently at 10. The egalitarian presumption implies that the workers should still get 2 units each, so long as information about their relative (general and specific) desert is lacking, resulting in an 8/12 distribution. The limitarian presumption, of course, would have us give 4 units to A and 0 to B, resulting in a 10/10 distribution. Thus, the egalitarian presumption does not imply the limitarian presumption. The egalitarian presumption, at least in the case under scrutiny, in which desert is the fundamental principle of justice, is insensitive to initial holdings, while the limitarian presumption is very sensitive to initial holdings, and would be inexplicable if not.60

There is also a more general worry. According to the limitarian presumption, “without substantive reasons to the contrary, we have reasons to regard a distribution as unjust if some people’s wealth exceeds the limitarian threshold.”61 In my view, it is hard to understand what a limitarian threshold refers to at all, if limitarianism is not itself supposed to be a substantive principle of distributive justice. It is, however, a premise for Timmer’s discussion that limitarianism is not a substantive principle of justice, and that the limitarian presumption is valid and useful across a range of possible substantive and ideal principles that are not themselves limitarian.

The second argument says that if one accepts the limitarian claim that there is a threshold of wealth above which wealth has zero moral value (or at least little value), then one should accept the limitarian presumption. Thus, one should prefer distributions in which no one has surplus wealth62 (at least in situations in which the total amount of wealth or goods are the same). In my view, this only holds on the assumption that sub-threshold wealth or goods are morally valuable, or at least more valuable than wealth or goods above the threshold. If so, the “limitarian claim” is (again), really a sufficientarian claim. And sufficientarians would prefer distributions with as much sufficiency as possible; whenever some individuals have more than the threshold, their excess wealth should be distributed to those who are below the threshold. To say that this is a limitarian presumption, rather than a sufficientarian one, does not seem right, for reasons presented above.

“The third argument for presumptive limitarianism is that decision-makers often lack the epistemic grounds to apply substantive principles for distributing wealth fairly.”63 Timmer goes on to suggest that political equality and meeting urgent needs are both substantial requirements of justice, and not plagued by such epistemic uncertainty.64 Moreover,

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.65

This seems to me to repeat the main strands of the instrumental arguments for limitarianism, and the reply is that a limitarian presumption just seems much less useful than an egalitarian presumption, and somewhat less useful than a sufficientarian presumption. This is true even though presumptions are supposed to “specify risk-averse principles that aim to minimize the harm of possible misallocations of valuable goods in the light of epistemic uncertainty.”66 There is no reason to think that the limitarian presumption is better at minimizing misallocation, given the goals of political equality and meeting urgent needs, than either the egalitarian presumption or a possible sufficientarian presumption.

There is, however, yet another way in which limitarianism could play a role in distributive justice. According to Robeyns, normative analysis of the rich “would make it possible for philosophers to have greater impact on existing debates in society.”67 The reason is that many political parties and movements (and economists) are preoccupied with the wealth of the very rich and the super-rich.

Two points in response. First, as I am sure Robeyns would agree, impact as such is a rather empty goal, and the extent to which limitarianism can be valuable in this particular sense, relies on the extent to which limitarianism can be theoretically and morally justified. As should be clear by now, I am not optimistic on that score. Second, if the quoted claim is true, that would presumably hold for all analyses of the upper tail of the distribution, including criticisms of limitarianism, and not only those that defend some form of limitarianism. Thus, this provides no particular reason to think that limitarianism qua distributive principle, is particularly helpful in this regard.

6. Conclusion

Overall, it appears that there is very limited need for a limitarian principle of justice. Robeyns’s instrumental limitarianism, which is the version of the view that is currently worked out in most detail, is not really limitarian, even though it may have some limitarian implications under some circumstances. More generally, it is difficult to imagine a plausible version of intrinsic limitarianism, since it would be hard to explain why it should be morally impermissible, as such, to have more than some limit.

Further, other possible versions of instrumental limitarianism, such as Zwarthoed’s, seem, like Robeyns’s urgent needs account, to be more sufficientarian than limitarian. Clearly, there can be many other versions, but we do not have much reason to embrace instrumental limitarianism until we see at least one such account that is convincingly limitarian, and not sufficientarian or egalitarian. In addition, any instrumental forms of limitarianism must be based on some intrinsic value that is not limitarianism (given that intrinsic forms of limitarianism are implausible).

Lastly, I have argued that there is little reason to think that limitarianism is particularly helpful either as a presumption or as a principle providing philosophers with more impact in public debates. None of this is meant to rule out that some plausible limitarian theory could be defended, or that limitarianism could constitute a part of an overall hybrid account,68 but the evidence so far suggests that limitarianism is of very limited use in discussion about distributive justice.

Acknowledgement

Earlier versions of this article were presented at a GOODPOL workshop in 2020, the Nordic Network in Political Theory in 2020, the Norwegian Network for Practical Philosophy in 2021, and a workshop on sufficientarianism organized by the University of Liverpool in 2021. I am grateful to Didde Boisen Andersen, David Axelsen, Katharina Bohnenberger, Lars Christie, Sebastian Johansen Conte, Ben Davies, Keith Dowding, Jakob Elster, Hallvard Fossheim, Markus Furendal, Andreas Føllesdal, Ian Gough, Clare Heyward, Fredrik Dybfest Hjorthen, Cathrine Holst, Simo Kyllönen, Ortrud Lessmann, Kasper Lippert-Rasmussen, Alejandra Mancilla, Jasmina Nedevska, Frank Nullmeier, Jørgen Pedersen, Thomas Princen, Jouni Reinikainen, Magnus Reitberger, Ingrid Robeyns, Rudolf Schuessler, Liam Shields, Espen Dyrnes Stabell, Attila Tanyi, Jesse Tomalty, Dick Timmer, David Vogt, and three anonymous referees for very helpful comments and suggestions. This chapter was first published in The Journal of Political Philosophy, 2022, 30(2), pp. 230–248.

References

Christiano, Thomas. 2012. Money in politics. In: David Estlund (Ed.), The Oxford Handbook of Political Philosophy. Oxford: Oxford University Press, pp. 241–57. https://doi.org/10.1093/oxfordhb/9780195376692.013.0013

Crisp, Roger. 2003. Equality, priority, and compassion. Ethics, 113, 745–63. https://doi.org/10.1086/373954

Dworkin, Gerald. 2020. Paternalism. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy, <https://plato.stanford.edu/archives/spr2020/entries/paternalism/>

Foley, Jonathan A. 2011. Can we feed the world and sustain the planet? A five-step global plan could double food production by 2050 while greatly reducing environmental damage. Scientific American, 305, 60−5.

Gheaus, A. 2018. Hikers in flip-flops: luck egalitarianism, democratic equality and the distribuenda of justice. Journal of Applied Philosophy, 35, 54−69. https://doi.org/10.1111/japp.12198

Huseby, Robert. 2010. Sufficiency—restated and defended. Journal of Political Philosophy, 18, 178−97. https://doi.org/10.1111/j.1467-9760.2009.00338.x

Huseby, Robert. 2020. Sufficientarianism and the threshold question. Journal of Ethics, 24, 207−23. https://doi.org/10.1007/s10892-020-09321-7

Kramm, Matthias and Ingrid Robeyns. 2020. Limits to wealth in the history of western philosophy. European Journal of Philosophy, 28, 954−69. https://doi.org/10.1111/ejop.12535

Kraut, Richard. 2018. Aristotle’s ethics. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy, <https://plato.stanford.edu/archives/sum2018/entries/aristotle-ethics/>

O’Neill, Martin. 2008. What should egalitarians believe? Philosophy and Public Affairs, 36, 119−56. https://doi.org/10.1111/j.1088-4963.2008.00130.x

Parfit, Derek. 1997. Equality and priority. Ratio, 10, 202−21. https://doi.org/10.1111/1467-9329.00041

Piketty, Thomas. 2014. Capital in the 21st Century. Cambridge, MA: Harvard University Press.

Räikkä, Juha. 2019. On the presumption of equality. Critical Review of International Social and Political Philosophy, 22, 809−22. https://doi.org/10.1080/13698230.2018.1438335

Robeyns, Ingrid. 2017. Having too much. In: Jack Knight and Melissa Schwartzberg (eds), NOMOS LVI: Wealth. New York: New York University Press, pp. 1–44.

Robeyns, Ingrid. 2019. What, if anything, is wrong, with extreme wealth? Journal of Human Development and Capabilities, 3, 251−66. https://doi.org/10.1080/19452829.2019.1633734

Robeyns, Ingrid. 2022. Why limitarianism? Journal of Political Philosophy, 30, 249–70.

Temkin, Larry. 2003. Egalitarianism defended. Ethics, 113, 764−82. https://doi.org/10.1086/373955

Timmer, Dick. 2019. Defending the democratic argument for limitarianism: a reply to Volacu and Dumitru. Philosophia, 47, 1331−339. https://doi.org/10.1007/s11406-018-0030-6

Timmer, Dick. 2021. Limitarianism: pattern, principle, or presumption? Journal of Applied Philosophy, 38, 760–63. https://doi.org/10.1111/japp.12502

Volacu, Alexandr and Adelin Costin Dumitru. 2019. Assessing non-intrinsic limitarianism. Philosophia, 47, 249−64. https://doi.org/10.1007/s11406-018-9966-9

Zwarthoed, Danielle. 2019. Autonomy-based reasons for limitarianism. Ethical Theory and Moral Practice, 21, 1181−204. https://doi.org/10.1007/s10677-018-9958-7