1. Introduction
The effective altruism (EA) movement, as its name suggests, encourages people both to be more altruistic, and to direct their altruistic efforts more effectively than many typically do. The EA organization Giving What We Can, for example, asks people to pledge to donate 10 percent of their income to support efforts considered to be among the most effective at making the world a better place. Since most people, including most well off people, donate significantly less than 10 percent of their income to charity, Giving What We Can is clearly asking a large number of people to increase their altruistic efforts. And since most people direct much, if not all, of the money that they do donate to organizations that, on any plausible view, do less good per dollar that they receive than the organizations recommended by EA charity evaluators,Footnote 1 EA organizations are clearly committed to the view that there are strong reasons for people to direct whatever money they do give more effectively.
Among the core commitments of EA, then, are, first, a commitment to the view that greater altruism (for example, in the form of increased donations to organizations that make the world a better place) is to be encouraged, and, second, a commitment to the view that directing altruistic efforts more effectively is to be encouraged. When formulated in this way, such that they entail only that it is better, morally speaking, if people donate more, and donate more effectively, these commitments will strike most people as at least relatively uncontroversial. If, however, the commitments are understood as representing moral obligations to which well-off people are subject, so that, for example, those who donate less than 10 percent of their income, and/or those who donate to less effective organizations, are acting wrongly, many more people will tend to find them objectionable. The claim that well off people are obligated to donate more, and perhaps significantly more, than most actually do, for example, will be rejected by many on the grounds that it is objectionably demanding.Footnote 2 And the claim that giving to less effective rather than more effective organizations is impermissible will be rejected by many as inconsistent with individuals’ right (or prerogative) to direct their altruistic efforts toward causes that they care particularly about, even if donations to those causes do less good than donations directed elsewhere (Lichtenberg Reference Lichtenberg2015; Srinivasan Reference Srinivasan2015; Krishna Reference Krishna2016; Boesch Reference Boesch and Woodruff2018).
Since EA is a social movement that is best understood as structured around a set of core commitments, including, centrally, philosophical commitments,Footnote 3 it is important for both theoretical and practical reasons for those sympathetic to its central aims to consider precisely how these core commitments should be understood. In particular, it is important to consider what effective altruists (EAs) must think about the content of our moral obligations in order to be able to make the claims about (for example) the ethics of charitable giving that are essential to the movement.
Theron Pummer has recently claimed that the commitment to effectiveness in giving is significantly more important to EA than the commitment to increased giving. On his view, although EAs do aim to encourage people to be more altruistic, and specifically to give more to charitable organizations than they typically do, ‘by far their most central recommendation is that we give effectively, that is, to whatever charities would do the most good per dollar donated’ (Pummer Reference Pummer2016: 78, italics in original). In addition, he argues that individuals can be subject to obligations to give effectively that are conditional – specifically, he claims that there are many cases in which, although it is permissible for an individual to refrain from giving to charity at all, if one is in fact going to give, it is wrong to give to a less effective organization rather than a more effective one (Pummer Reference Pummer2016: 78; see also Horton Reference Horton2017). On a view with this structure, it might be wrong for a well off person to give $40,000 to a charity that trains seeing-eye dogs for blind people in the United States, given that she could have given instead to a charity that provides surgeries that reverse the blindness-causing effects of trachoma for people in Africa, even if it would have been permissible for her to, for example, spend the same $40,000 on a down payment for a new house for herself and her family.Footnote 4
There are a number of reasons, both theoretical and grounded in the aim of movement-building, why EAs might find the view that conditional effectiveness obligations can apply even in the absence of unconditional obligations to give appealing. Perhaps most importantly, it would allow them to at least largely avoid taking a position on the issue of morality's demandingness, and therefore allow them to appeal to people to give more effectively without endorsing or defending the widely rejected (and potentially off-putting) claim that well off people have demanding obligations of beneficence that they are, for the most part, currently failing to satisfy.Footnote 5 At the limit, EA might even be thought to be consistent with the view that there are no unconditional obligations of beneficence.Footnote 6 The movement could insist on effectiveness without committing to (any particular degree of) demandingness.
In my view, however, EAs cannot avoid endorsing at least some unconditional obligations of beneficence, and more generally cannot avoid taking at least a fairly strong position on morality's demandingness. I do not deny that there are cases, including potentially some involving charitable giving, in which conditional effectiveness obligations apply in the absence of unconditional obligations of beneficence. I believe, however, that the range of charitable giving cases of which this is true is too narrow for EAs to be satisfied that their core commitments allow them to criticize, in EA terms, ineffective giving in just those cases. If I am correct about the range of cases in which it is plausible that agents are subject to conditional effectiveness obligations, on the assumption that they are not subject to unconditional obligations of beneficence, then refraining from committing to at least fairly substantial unconditional obligations of beneficence applying to well off people would undermine the common EA claim that a great deal of actual charitable giving is morally objectionable (because it is substantially less than maximally effective). EAs would not, then, be able to claim that many actual people are morally obligated to give to different organizations than those to which they in fact give.Footnote 7
My central aim in the remainder of this article is to argue that EAs cannot, consistent with their broader aims, avoid endorsing at least a fairly demanding view about the unconditional obligations of beneficence to which well off people are subject. In at least many of the cases of concern to EAs, effectiveness cannot be entirely separated from demandingness. I will defend this claim by arguing that if one accepts any of the most plausible and commonly endorsed grounds for thinking that it is permissible for individuals to refrain from giving (more) to charity, then one cannot deny that those same grounds justify at least a great deal of ineffective giving. This has significant implications not only for how we should understand the core commitments of EA, but also for how we should think about the conditions in which individuals might be subject to conditional obligations of all kinds.
I will proceed as follows. In section 2, I will describe the arguments offered by Pummer and Joe Horton for the view that individuals can be subject to conditional effectiveness obligations, despite not being subject to unconditional obligations of beneficence. And I will note the implications that they think this has for cases involving charitable giving. In section 3 I will argue that, in a wide range of cases involving charitable giving, the kinds of reasons that most plausibly explain why individuals might not be subject to unconditional obligations of beneficence also provide grounds for rejecting the view that they are subject to conditional effectiveness obligations. This, I will claim, suggests that at least in these cases, the effectiveness obligations cannot be defended independently of unconditional obligations of beneficence, and therefore cannot be defended without, in effect, taking a position on morality's demandingness. I will conclude, in section 4, by briefly noting the approach to defending potentially demanding unconditional obligations of beneficence that I think is most promising, both for EAs and more generally.
2. Conditional effectiveness obligations without unconditional obligations of beneficence
Pummer and Horton each offer a case in which an agent can, at great cost to herself, save the lives of either one or multiple nearby people, in order to motivate the view that there can be conditional effectiveness obligations without unconditional obligations of beneficence.Footnote 8 Here is Horton's case, to which Thomas Sinclair gives the name Collapsing Building:
Collapsing Building: Two children are about to be crushed by a collapsing building. You have three options: do nothing, save one child by allowing your arms to be crushed, or save both children by allowing your arms to be crushed. (Horton Reference Horton2017: 94; Sinclair Reference Sinclair2018: 41)
As Horton (Reference Horton2017: 94) suggests, intuitively it is permissible to refuse to sacrifice one's arms in order to save the children because of the very high cost of doing so. Any view that implies that one is obligated to make this sacrifice would, intuitively, be objectionably demanding.Footnote 9 At the same time, it seems clearly objectionable to save only one of the children, given that one could save both at no additional cost to oneself (Horton Reference Horton2017: 94; Sinclair Reference Sinclair2018: 41).
On Horton's view, one comes to be subject to an obligation to save both children if one is willing to sacrifice one's arms in order to save either one or both of them (Reference Horton2017: 97). Willingness to make the sacrifice in order to achieve either the lesser or the greater good, then, is the condition that, for Horton, triggers the conditional effectiveness obligation in cases like Collapsing Building. On his view, once it is the case that one is willing to make the sacrifice, one can no longer appeal to the fact that saving the children would involve that sacrifice in order to justify refusing to save both of them. In other words, the fact that explains why it would be permissible to refuse to save any of the children cannot be appealed to in order to justify saving only one of them. Saving only one of the children could be permissible only if there were some alternative justification, independent of the fact that saving both would involve the sacrifice of one's arms, for doing so.
Horton accepts that in at least some cases there will be alternative justifications for bringing about a lesser rather than a greater good, even if it is permissible not to do either because of the cost that would be involved. Specifically, in some such cases individuals will have agent-relative reasons of sufficient strength to justify them in bringing about the lesser rather than the greater good (Horton Reference Horton2017: 98–99). For example, Horton suggests that if one can save either one's own child or two different children at the same very great cost to oneself, one plausibly has sufficient agent-relative reason to save one's own child, despite the fact that saving the two other children would be impartially better (Horton Reference Horton2017: 99).Footnote 10
Horton claims that our intuitive reactions to Collapsing Building, in combination with the fact that agent-relative reasons sometimes justify bringing about a lesser rather a greater good at equal cost to oneself, provide support for the following principle:
Optimific Altruism: If the only adequate justification for not bringing about an outcome O is that it requires a sacrifice S, and we are willing to make a sacrifice that is not significantly smaller than S to bring about an outcome that is not significantly better than O, and we do not have adequate agent-relative reasons to favor this other outcome, we ought to bring about O. (Horton Reference Horton2017: 99)
In somewhat simpler terms, this principle holds that for any sacrifice that one is willing to make, one is obligated to make it in the way that will do the most good, impartially construed, unless one has adequate agent-relative reasons for making it in one or more particular suboptimal ways. If one has adequate agent-relative reasons for making it in one or more suboptimal ways, then it is permissible to make it in those ways (though not in others).Footnote 11
Pummer endorses a principle with at least roughly similar implications, which he calls Avoid Gratuitous Worseness (weak):
Avoid Gratuitous Worseness (weak): It is wrong to perform an act that is much worse than another, if it is no costlier to you to perform the better act, and if all other things are equal. (Pummer Reference Pummer2016: 84).Footnote 12
Pummer is somewhat less clear than Horton about exactly what he takes the condition that triggers the conditional effectiveness obligation to be. It is most plausible, it seems to me, to understand him as holding either that the relevant condition is actually making the sacrifice, or that it is the fact that one is going to, or intends to, make it.
The case that Pummer describes in order to highlight the appeal of Avoid Gratuitous Worseness (weak) is the following, which he calls Arm Donor:
Arm Donor: There is one innocent stranger stuck on track A, and one hundred strangers stuck on track B. On each track is a trolley headed toward them, which will kill them unless the trolley is stopped. You can use your arm to stop one, but not both, of the trolleys, and will lose the arm if you do. (Pummer Reference Pummer2016: 83)Footnote 13
Avoid Gratuitous Worseness (weak), Pummer claims, plausibly accounts for the intuition that while it would be permissible to refrain from sacrificing an arm because of the large cost involved in doing so, it would be wrong to make the sacrifice in order to save the single person on track A (Reference Pummer2016: 83–84).Footnote 14
It is at least a bit more controversial that one is subject to a conditional obligation to bring about more rather than less good in Arm Donor than in Collapsing Building, since in Arm Donor this could be true only if it were also true that one is unconditionally obligated to save more rather than fewer people in cases in which doing so would be costless. In order to be consistent with any account of effectiveness that EAs might endorse, however, any argument that aims to show that well off people can be subject to conditional obligations to give to more effective rather than less effective charities by extending plausible views about nearby rescue cases must proceed on the assumption that there is an obligation to save the greater number. This is because one of the important factors that EAs consider when assessing the relative effectiveness of different charities is the number of people who are benefitted by their work – a charity that prevents one case of blindness for every $40,000 it spends is less effective, on any account of effectiveness that EAs might endorse, than one that prevents 2,000 cases of blindness for every $40,000 it spends.
If we assume that there is, all else equal, an obligation to save (or otherwise aid) the greater number, then, as Pummer and Horton suggest, the principles that they endorse have significant implications for at least some cases of charitable giving. Optimific Altruism implies that if one is willing to give an amount of money to charity, then she is obligated to give it to the charity that will do the most good with it, unless she has adequate agent-relative reasons to give it to one or more other particular charities (in which case she is permitted to give to those charities, but not to other suboptimal ones) (Horton Reference Horton2017: 102). And Avoid Gratuitous Worseness (weak) implies that giving to a less effective charity is wrong whenever giving to a more effective one would be no more costly to the donor (Pummer Reference Pummer2016: 88–89).
Clearly, then, at least some less than optimal giving will count as impermissible according to both Pummer and Horton's principles. In addition, the principles seem both to be independently plausible and to provide plausible grounds for the intuitive verdicts about cases like Collapsing Building and Arm Donor. Whether they can provide grounds for thinking that EAs can avoid taking a position on morality's demandingness, and can instead defend enough of their central claims by appealing to conditional effectiveness obligations, depends on whether enough cases involving charitable giving are such that the principles imply that giving less effectively rather than more effectively is impermissible.
3. Justifications for refusing to give and justifications for giving less effectively
In order for it to be the case that an agent is subject to a conditional obligation to give to a more effective charity rather than a less effective one, it must be the case that there are sufficient grounds for rejecting the claim that she has an unconditional obligation to give, yet insufficient grounds for holding that she is permitted to give to the less effective organization. And in order for that to be the case, the considerations that provide sufficient grounds for the permissibility of refusing to give must not provide grounds of similar strength for the permissibility of giving to the less effective charity. In addition, there must be no other, independent grounds to which the agent can appeal that would justify giving to the less effective charity. In order to attempt to determine how wide the range of charitable giving cases in which individuals are subject to conditional effectiveness obligations is, then, we must consider: (1) what the most plausible grounds for permissions not to give are; (2) how often those grounds will fail to provide grounds of similar strength for giving less rather than more effectively; and (3) what other grounds might justify giving less rather than more effectively.
The most plausible ground for thinking that it is permissible for an agent to refrain from giving some amount of money to charity is that doing so would involve costs for her of some kind and amount that are important enough, given the relevant facts (including how much good she could do by giving),Footnote 15 that accepting those costs would be supererogatory. In Collapsing Building and Arm Donor, the high costs that the agents would endure if they chose to sacrifice one or both of their arms plausibly make it supererogatory to save anyone. Similarly, in many cases agents might be permitted to keep an amount of money rather than donating it due to the costs that would be involved in giving it up. Importantly, the relevant costs should be understood in terms of considerations that are plausibly morally important in themselves, such as an agent's well-being, projects, or personal attachments.Footnote 16 Having less money in one's bank account, for example, is not in itself a morally relevant cost. If having less money would not in any way negatively affect a person's well-being, or her ability to pursue her projects or advance the interests of people about whom she cares especially, then it is not plausible that she can appeal to the ‘cost’ of giving up the money in order to ground a permission to refrain from donating it.
There are clearly some cases in which the considerations that plausibly ground a permission not to give do not also justify giving to a less effective rather than a more effective organization. There are two types of case that are perhaps especially important for thinking about the potential reach of appeals to conditional effectiveness obligations, both for EAs and more generally. In cases of the first type, agents give to a less rather than a more effective organization within a cause area that they care particularly about. Consider, for example:
Guide Dogs: Anil is committed to the cause of providing seeing-eye dogs to blind Americans. He has no personal connection to any of the organizations that work on this issue (for example, none of the organizations has provided a dog to anyone whom Anil knows personally). Charity A provides one dog to one blind American for every $45,000 it spends. Charity B provides one dog to one blind American for every $40,000 it spends. Anil gives $1,000 to Charity A.
If we assume that it would have been permissible for Anil to spend the $1,000 on, for example, a more directly self-regarding personal project, then it seems clear that this is a case in which the grounds that make it permissible for him to refrain from donating the money at all (for example, his commitments to his projects, including the project of providing guide dogs for blind Americans) do not also justify his donating to Charity A rather than Charity B. And there do not seem to be any other grounds that could plausibly justify doing this.Footnote 17 Anil, then, has a conditional obligation to donate to Charity B rather than Charity A – given that he donates (or is going to donate, or intends to donate, or is willing to donate), he is obligated to donate to Charity B rather than Charity A (assuming that these are the only relevant options).
A second type of case in which the considerations that plausibly ground a permission not to give do not also justify giving to a less effective rather than a more effective organization occurs when individuals give unreflectively to organizations that focus on causes to which they have no special commitment. Consider, for example:
Whim-Giving: Bettina sees an ad on television for Charity C, which highlights the work that it does in a particularly heartwarming way. She does not care about the kind of work that this charity does any more than she cares about the kinds of work done by many other organizations. There are many organizations that, on any plausible view, are more effective at making the world a better place, including Charity D. Bettina gives $100 to charity C.
If we assume that it would have been permissible for Bettina to spend the $100 on a personal project, then this is also a case in which neither the grounds that make it permissible for her to refrain from donating, nor, it would seem, any independent grounds, justify donating to Charity C rather than (for example) Charity D.Footnote 18 Like Anil, then, Bettina is subject to a conditional obligation to give more rather than less effectively.
In Guide Dogs and Whim-Giving, it would be no more costly to the agents, in any morally relevant sense, to give more rather than less effectively.Footnote 19 This is why both Optimific Altruism and Avoid Gratuitous Worseness (weak) imply that Anil and Bettina are conditionally obligated to give more rather than less effectively. On the other hand, many cases in which individuals give to less rather than more effective charities are such that it would be more costly to them, in some plausibly morally important way, to give more effectively (Pummer Reference Pummer2016: 82, 91–93; McMahan Reference McMahan and Woodruff2018: 83). In particular, in many cases it would be more costly for reasons that are clearly of the same kind as those that most plausibly explain why it is permissible for individuals to refrain from giving entirely.Footnote 20 Consider, for example:
Project-Based Giving: Cong has as one of his most valued projects helping to provide music education to students in the middle-class suburban district in which he grew up. He volunteers in the music programs at several of the schools in the district, and in addition provides private lessons to interested students for free. Motivated by his commitment to this project, he donates $1,000 to the music program at his old high school.
On any plausible view, Cong's donation is far less than maximally effective. If he had instead given the money to one of the organizations recommended by GiveWell, the amount of good that his donation could be expected to do would be significantly greater. Nonetheless, if we assume that it would have been permissible for him to spend the money on one of his other, more directly self-regarding personal projects – that is, if we assume that he had no unconditional obligation to donate it – then it seems clear that we must accept that his donating it in order to further his valued personal project is permissible. The very same kind of reason that, by hypothesis, explains his lack of an unconditional obligation to give – namely that, at least within certain limits, individuals are permitted to devote resources to the pursuit of their personal projects – provides a reason of at least equal strength in favor of the permissibility of donating to the music program.
Pummer acknowledges that in any case in which a person cares more about the work of a less effective charity than she cares about the work of a more effective one, the same kinds of reasons that ground permissions to refrain from giving will also count in favor of a permission to give to the less effective charity (Pummer Reference Pummer2016: 82). He also notes that Avoid Gratuitous Worseness (weak) does not rule out the permissibility of giving to a less effective rather than a more effective charity in any case in which the donor cares even a small bit more about the work of the less effective organization (Pummer Reference Pummer2016: 92). Avoid Gratuitous Worseness (weak), then, implies that agents have conditional obligations to give more rather than less effectively in only two kinds of cases: (1) cases like Guide Dogs, in which donors are committed to giving within a (perhaps highly specific) cause area; and (2) cases in which they care no more about the work of a less effective organization than they care about the work of a more effective one.
If EAs could only insist that donors ought to give more rather than less effectively in cases of these two kinds, this would limit the scope of their efforts quite a bit more than they might hope, and quite a bit more than is suggested by much EA advocacy and argument. Pummer suggests that we might accept a stronger version of Avoid Gratuitous Worseness, and thus hold that individuals can sometimes have conditional obligations to give more rather than less effectively, even when giving more effectively would be more costly for them than giving less effectively (Reference Pummer2016: 93). If such a stronger principle could be defended, it might extend the scope of conditional effectiveness obligations significantly, thereby providing grounds for objecting to much more less effective charitable giving than the weak version.
In order to provide intuitive support for a stronger version, Pummer claims that in a version of Arm Donor in which saving the one hundred on track B would require suffering a mild sore throat in addition to the loss of one's arm, it still seems that if one is going to sacrifice one's arm, it would be wrong to save the one on track A rather than the one hundred on track B (Reference Pummer2016: 93). In order to accommodate this intuition, he suggests that we might accept:
Avoid Gratuitous Worseness (still pretty weak): It is wrong to perform an act that is much worse than another, if it is slightly costlier to you to perform the better act, and if all other things are equal. (Pummer Reference Pummer2016: 93)
While this principle has plausible implications for cases like Pummer's revised version of Arm Donor, however, it has other implications that are much less plausible. And some of these implications make it clear that it cannot be understood, as Pummer intends, as generating only conditional effectiveness obligations.Footnote 21
Consider the following case:
Charity E: Charity E funds surgeries that reverse the blindness-causing effects of trachoma for people in Africa. It provides one surgery for every $20 in donations that it receives. Dante gives $20 to Charity E.
By giving this $20, Dante helps one person regain her sight. On any plausible view, however, doing this is much worse than helping both that person plus a second person regain their sight. Because of this, so long as giving $40 would only be slightly more costly to Dante than giving $20 (which for virtually any well off person will be the case), Avoid Gratuitous Worseness (still pretty weak) implies that Dante's giving $20 (or intending to give it, or being willing to give it) makes it the case that he is obligated to give $40. If we assume that it would have been permissible for Dante to spend the initial $20 on a personal project, it seems quite implausible that his giving that amount can make it the case that he is obligated to give even more.Footnote 22
In fact, however, Avoid Gratuitous Worseness (still pretty weak) does not even allow us to assume that Dante is permitted to spend the initial $20 on a personal project. This is because it would be much better to donate it to Charity E, while (we can assume) being only slightly costlier to him.Footnote 23 This makes it clear that any strengthening of Avoid Gratuitous Worseness (weak) renders it no longer merely a principle that can generate conditional obligations.Footnote 24 Instead, any strengthened principle will generate at least some unconditional obligations to accept modest costs in order to do a significant amount of good.
Furthermore, even the ‘still pretty weak’ version will imply that no matter how much a person donates (or intends to donate, or is willing to donate) to an organization like Charity E, she will be obligated to donate more, so long as doing so would be only slightly more costly to her. This is because helping n people regain their sight is much worse than helping those people plus an additional person regain their sight, for any value of n.Footnote 25 Any principle with implications of this kind is not merely a plausible extension of a principle like Avoid Gratuitous Worseness (weak). And because the principle itself entails that we can have obligations to give effectively that are not conditional on our giving (or intending to give, or willingness to give) in the first place, it cannot be appealed to in order to suggest that the scope for conditional effectiveness obligations is sufficiently extensive that EAs can rely on those obligations to provide support for enough of the claims that they want to make about the ethics of charitable giving.
Pummer claims that although giving less than maximally effectively will sometimes be permissible for the same kinds of reasons that refraining from giving is permissible, ‘in many cases it would be wrong . . . to give a sum of money to charities that do less good than others you could have given to instead, even if it would not have been wrong . . . not to give the money to any charity at all’ (Pummer Reference Pummer2016: 78; see also 82, 91). There is one sense in which, even in light of my argument thus far, this seems true. We can imagine an extremely wide range of cases that share the morally relevant features of Guide Dogs and Whim-Giving, and in all of those cases it seems correct that the relevant individuals will have conditional obligations to give more rather than less effectively.
Pummer, however, intends his claim as one about real-world cases of charitable giving. He states that it is ‘largely an empirical matter’ how prevalent cases with the relevant structure are (Pummer Reference Pummer2016: 91). Once we recognize that the ‘weak’ version of Avoid Gratuitous Worseness is the only version that implies only conditional obligations, however, it seems implausible that the range of real-world cases in which there could be conditional effectiveness obligations without unconditional obligations of beneficence is all that extensive. And it seems even more implausible that many of the cases of less effective giving that EAs tend to be especially critical of fall within this range.
First, although cases with the relevant features of Whim-Giving might be at least fairly common, most such cases will involve relatively small amounts of money, especially in comparison with the amounts that are given in cases that that share the relevant features of Project-Based Giving. It seems clearly inadequate for EAs to be able to criticize those who give $20 to an organization after seeing an ad, when they could have given to a more effective one instead, but unable to criticize, in EA terms, those who give large amounts to their wealthy alma maters because they care particularly about those institutions.Footnote 26
Second, I suspect that cases with the relevant features of Guide Dogs are at least somewhat less common than we might initially think. Although there are cases in which multiple charitable organizations work on the very same cause, if we understand causes in a sufficiently fine-grained way to cover everything about an organization's work that donors might care about (for example, in which communities they operate), the range of such cases will at least be fairly limited.Footnote 27 Of course, there will certainly be cases in which individuals do not care about the fine-grained details of different organizations’ work, and donate to organizations that less effectively advance the values that they do care about than others to which they could have donated instead. But again, if EAs are limited to criticizing those who give less rather than more effectively within whatever cause areas they happen to care about, this seems clearly inadequate, and clearly inconsistent with much actual EA advocacy and argument (Berg Reference Berg2018).
In order to make many of the central claims about the ethics of charitable giving that EAs tend to endorse (for example, that those who give large sums to wealthy universities that they are particularly committed to rather than to organizations recommended by EA charity evaluators act wrongly), it is necessary to argue that individuals either ought to care more than most do about things like improving the lives of the global poor, and to act accordingly, or else that they ought to donate (more) to causes like global poverty alleviation, even if they do not in fact care as much about that cause as they do about others that are less important, impartially speaking. And, for reasons that should now be clear, arguments to this effect cannot be made by appeal to conditional effectiveness obligations of the kind that plausibly apply in cases like Collapsing Building and Arm Donor. Instead, the view that well off people have unconditional obligations of beneficence to donate effectively must be defended. And that means that EAs cannot avoid taking at least a fairly strong position on morality's demandingness.Footnote 28
4. Conclusion
If my argument in section 3 is correct, then EAs cannot make many of their central claims about the ethics of charitable giving without accepting that well off people have potentially demanding unconditional obligations of beneficence. At least some advocates of EA have, however, been reluctant to endorse this view.
I believe that this is a mistake. There are a number of plausible grounds for thinking that well off people have demanding unconditional obligations to direct a portion of their resources where they will do the most good. In my view, the best approach to defending this claim involves arguing that individuals lack even a presumptive moral entitlement to resources that they possess in virtue of benefiting from the operation of unjust global economic institutions. If they do not have a presumptive moral entitlement to these resources, then they cannot plausibly claim that they are entitled to exercise discretion over what is to be done with them.Footnote 29 And if they are not entitled to exercise such discretion, then a strong case can be made that they are obligated to direct the resources to which they are not entitled where they will do the most good.
There are, of course, many different ways that a view of this kind might be developed, and a range of challenges that could be raised against the approach that I have suggested. These issues must be left to be dealt with elsewhere. I hope to have shown, however, that EAs cannot avoid endorsing and defending potentially demanding unconditional obligations – they cannot insist on effectiveness without accepting (some degree of) demandingness.Footnote 30
