Should the beneficiaries pay?

Introduction

Many political theorists have recently devoted their attention to the question of how the burdens following from climate change should be distributed. Broadly speaking, this debate seems to be on the right track. I subscribe to the view that there are two principles that, taken together, are well suited to address the issue. The first is the polluter-pays principle (PPP), and the second is the ability-to-pay principle (APP). This combination lies at the heart of several influential accounts of climate justice, including recent work by Simon Caney (2005, 2010) and David Miller (2009) (glossing over some notable differences).

The PPP is based on the idea that if an agent contributes to some harm, she is morally responsible for undoing that harm, or for compensating the victim (usually only insofar as she meets some further criteria for being held morally responsible). This principle is integral to most people’s conception of morality. However, as several theorists have argued, the PPP has a less extensive application to the case of climate change than one might initially think. ¹ Consequently, we must either accept that the brunt of the harm brought on by climate change must go unaddressed, or else look for some other principle that might complement the PPP.

The APP is such a principle. While it can be interpreted in several ways, the APP holds, roughly, that those agents who are capable of preventing or alleviating severe harm should do so, even if they have not themselves contributed to the harm in question (Caney, 2005, 2010; Miller, 2001, 2009). ²

Other writers also endorse, in one form or the other, the beneficiary-pays principle (BPP). According to the BPP, an agent can gain duties to address some harm if she has benefited from the action that caused it. Thus, in relation to climate change, those (but not only those) who have benefited from greenhouse gas emissions should pay the cost of mitigating further emissions and compensate those who are currently harmed by previous emissions (Baatz, 2013; Gosseries, 2004; Page, 2008, 2012). ³

The main aim of the present paper is to argue that the BPP is unjustified. In the next section I offer a definition of the principle. The third section provides some theoretical context, and argues that the BPP belongs to the domain of corrective, rather than distributive justice. In the fourth section, I briefly assess the merits of the BPP in the context of climate change, focusing mainly on Caney’s (2006) discussion of this issue. In the fifth section I consider some of the arguments that have been presented in support of the BPP, while the sixth and seventh sections are devoted to criticizing the principle. The eighth section concludes.

BPP defined

The BPP can be stated as follows:

If an agent A benefits from an action x ⁴ that she did not perform, and an agent B is harmed by x, then, if B did not perform x, A, in virtue of benefiting from x, owes B compensation (which is not to exceed A’s gain from x), insofar as no other agent has stronger duties to compensate B.

While it is easy to see the appeal of this principle, I suspect that the appeal stems largely from the fact that it gives the intuitively right answers in many cases (although for the wrong reasons, as I will argue below), and from the fact that it trades on the idea that the causal connection between the two agents (that are both affected by the action x) is morally relevant. Causal connections can indeed be morally salient, but I will claim that this is not the case in BPP scenarios.

The BPP comes in different versions. One could for instance include events, in addition to acts. Nevertheless, I think that my definition will suffice, since it is unlikely that the BPP will be justified in the case of events, but not in the case of acts. If it is justified in the case of acts, however, an extension might be considered. Further, the definition could be restricted to instances of wrongful or unjust harm. I have avoided this limitation because the BPP, as noted, is often discussed in relation to climate change, and restricting it to unjust harm would (potentially) exclude acts of emission performed at times when the emitters had no idea that what they did would be harmful to later generations. I do not want to rule out by definition that benefitting from harmful acts that are not unjust may give rise to duties of compensation. ⁵ However, since it seems that the BPP has its strongest chance of being justified in cases of unjust harming, I will mainly discuss such cases.

The BPP has also been specified in a way that severs the causal relation between the agent who is benefited and the agent who is harmed. Caney (2010) has suggested that an agent who benefits from an unjust act gains duties of compensation towards any agent who has been harmed by an unjust act, regardless of whether or not the two acts were identical. This version is wider than the above definition (I will refer to it as the generalized BPP, or GBPP). It is seemingly also less vulnerable to criticism. I will address it explicitly in the seventh section. Lastly, it has been argued that if an agent causes a harm for which she cannot be held morally responsible, she can nevertheless be held liable for her act to the extent that she benefits from it (Caney, 2010: 210; Gosseries, 2004: 40–41). This version of the BPP rests on a combination of contribution and benefitting that falls outside the scope of this paper.

Thus, while my definition of the BPP is not restricted to cases of unjust harms, I assume that if I succeed in arguing that agents who benefit from unjust harms do not, in virtue of that fact, gain duties of compensation towards agents who are harmed by the same action, I thereby show that the BPP is unjustified in most versions, excluding Caney’s wider justice-based version (which I will discuss further), and the version that includes (innocent) contribution (which I will not address).

The beneficiary-pays principle and corrective justice

The BPP is in some respects a limited principle. It is not advanced as a general theory of distributive justice (according to which the only transfers in the name of justice would go from beneficiaries of harmful act x to those harmed by x). Rather, the BPP should be seen as a form of corrective justice, aimed at rectifying injustices (Butt, 2009: ch. 2). ⁶

Further, the principle is not advanced as a primary criterion for corrective justice (according to which the beneficiary from harm to others is the agent primarily responsible for correction, and the perpetrator only secondarily responsible). No theorists, to my knowledge, argue that the BPP should replace the PPP in cases where they both apply. Rather, the BPP is best seen as a secondary principle of corrective justice. If the perpetrator of harm is present and able, she should compensate the victim. ⁷

A further question concerns the relationship between the domains of distributive and corrective justice. This relationship, however, cannot be determined in the abstract (Walt, 2006). (For ease of illustration I will in the remains of this section refer to the PPP rather than the BPP, and assume that the PPP relies on the well-known no-harm principle. Most people, and most legal systems, accept that if A harms B, A (perhaps given a set of further requirements) owes B compensation. The PPP, then, can be understood as a (primary) principle of corrective justice).

Some theories of distributive justice do not have room for corrective justice. Monistic theories, such as total outcome egalitarianism or total utilitarianism will, by definition, not allow further distributive principles. Consider total utilitarianism. On this theory, the only goal is to maximize (net) amount of welfare or utility. If A harms B, the question of whether or not A owes B compensation relies crucially on whether or not A’s compensating B will maximize welfare. If it will, then that is why compensation should take place, and corrective justice will provide no independent reasons. If, on the other hand, compensating B will not increase total welfare, compensation would violate justice. ⁸

Pluralist theories, on the other hand, are perfectly compatible with principles of corrective justice. In the following, then, I will assume a pluralist background theory of distributive justice. I will also assume that this background theory is incomplete, in the sense that it remains silent on at least some distributions. The reason is that it is easier to assess principles of corrective justice if we assume that distributive justice does not always have distinct (even if overridable) implications for the cases under consideration. To see why this is the case, assume a pluralist theory combining (non-total) utilitarianism and PPP. If we consider a case in which some harm-inflictor presumably owes compensation to some victim, it will complicate matters if the background theory of distributive justice already provides arguments for or against compensation to the victim (whichever maximizes utility).

To be sure, even incomplete pluralist theories may at times conflict with corrective justice. Consider a form of pluralist sufficientarianism, according to which the incidence of sufficiency ought to be maximized, and harms ought to be compensated. These two concerns may conflict, if for instance A’s compensating B would bring A below the sufficiency threshold. Theories that allow for such conflicts must then prioritize distributive justice over corrective justice, or the other way around. Alternatively trade-offs can be made, so that A should compensate only partly, in order to meet the demands of corrective justice to some extent, even if this brings her slightly below the sufficiency threshold.

These illustrations are relatively straightforward. There are more complicated cases, for instance luck egalitarianism (LE), which I will discuss briefly below. The main point is that evaluating the BPP as a principle of corrective justice is a quite complicated exercise, since not all of its proponents specify the notion of distributive justice that it is supposed to complement. Because of this, I will, as noted, assume a background theory of distributive justice that is both pluralist and incomplete. I will also assume that in the (majority of the) examples I introduce (mainly in the sixth and seventh sections) the distributions before and after the harm, and before and after an eventual compensation, are all compatible with the (incomplete and pluralist) background theory of distributive justice. It should be emphasized, however, that assuming some suitable background theory of justice is not sufficient to prevent whatever (perhaps monistic and complete) background theory we actually accept, from influencing our judgements of BPP cases.

The beneficiary-pays principle and climate change

According to the BPP, then, if (human-induced) climate change benefits A and harms B, and neither A nor B have caused the problem, A has a prima facie duty to compensate B. This duty might be overridden by the fact that climate change is caused by actions performed by agent C (and, presumably, others). If so, C has a duty to compensate B, and this duty is stronger than A’s duty. Of course, climate change is not as simple as this, and some of the special features of this phenomenon make (as will become clear below) the BPP especially inapplicable. But as noted, I want eventually to argue that the principle is indefensible even in less complex contexts. ⁹

Simon Caney (2006: 473) has criticized the BPP as applied to climate justice. One problem, he notes, is that most beneficiaries of historic emissions are now dead, and it would be unfair to make those beneficiaries who happen to be alive, foot the whole bill. It seems, however, that this objection can be met by adherents to the BPP. Firstly, as implied by the definition above, the beneficiary’s duty to compensate can reasonably be limited upwards, in such a way that the compensatory duty does not exceed the received benefit. Even if it might still be unjust, in one sense, that some (living) beneficiaries pay, while others (the dead) do not, this is not sufficient to let the (live) beneficiaries off the hook. If both A and B owe money to C, A’s debt does not disappear just because B fails, for some reason, to honour his. ¹⁰ However, it could be the case (depending on the size of the benefits received by the living), that parts of the burden of climate change would remain, if the total cost exceeds the cost that can legitimately be imposed on the remaining beneficiaries. This, however, does not amount to a criticism of the principle, it only points to a limitation in scope.

Secondly, Caney (2006: 474–476) refers to the non-identity problem, ¹¹ and argues that at least some (but probably many) of those actions necessary to produce the benefits that allegedly accrue to currently living people, are extremely likely to also belong to the immense set of necessary conditions for current people’s existence. Thus, just as current people cannot be said to be harmed by climate change because there is no alternative world in which they both exist and are unharmed by historic emissions, current people cannot be said to benefit from climate change because there is no alternative world in which they both exist and are not benefited by historic emissions. ¹² On this view, both harms and benefits are analysed in comparative terms, so that A is only harmed or benefited relative to an alternative world in which the beneficial or harmful act did not take place. Clearly, this only works if the alternative world contains A.

While this seems to be a forceful argument against the application of the BPP (and the PPP), it is not intended to undermine the plausibility of the BPP in general. Limited application does not equal zero application, and there is still a need to assess the principle further. (Note that the PPP is not taken to be implausible as such even though it too, as mentioned, has a limited application to climate change.)

Caney does present a more general objection to the BPP. If A’s action harms C, but both A and B benefit from A’s action, it is unclear how the burdens should be shared between them. It appears that the BPP and the PPP do not “… sit easily together” (2006: 472). Caney suggests four solutions. Firstly, we could give A the largest share of the burden, since she both caused and benefited from her action. But it is hard to see exactly how to distribute these burdens. A second option is to hold that the PPP and the BPP overlap in the sense that they both pick out the same duty-bearers. Caney dismisses this as implausible. Thirdly, one could argue that the BPP should replace the PPP, but again, this seems to be a non-starter (for reasons suggested above). A fourth possibility is to say that if the causally responsible agent is alive, she should pay, whereas if the causally responsible agent is dead, the BPP kicks in and assigns duties to living beneficiaries. Caney (2006: 472–473) is sceptical of this proposal, however, because it is hard to see why receiving a benefit generates duties in some cases, but fails to do so in others.

In my view, Caney’s responses to the second and third suggestions are entirely correct. I am, however, less convinced by his rejection of options one and four. As regards the first, it is surely hard to know exactly how to distribute the duties between an agent who is a beneficiary and an agent who is a beneficiary and a contributor to harm. But this kind of uncertainty is not uncommon in ethics. Consider a case in which an agent negligently causes an accident, while a different agent intentionally avoids preventing the accident even if she could have done so with virtually no cost or effort. How exactly duties, responsibilities, blame, and so on are to be distributed is seldom easy to answer with precision. Thus, I do not agree that this option is necessarily implausible (which is not to say that just any kind of vagueness and indeterminacy is always perfectly all right).

I have a similar reaction to Caney’s reply to the fourth alternative. To say that being a beneficiary of harm must generate duties either always or never, is to overlook other relevant possibilities. Being a beneficiary could, for instance, generate defeasible duties, defeasible in particular by duties generated by contributing to some harm. This too is quite familiar from other situations. If someone recklessly causes a small child to fall into deep water, the reckless person clearly has a duty to jump in and save the child. But if he flees the scene, then any capable bystander ought to save the child, whether the duty came into existence at the exact moment the reckless person fled, or whether it was there all along as a defeasible duty.

It seems then, that neither the first nor the fourth options are undermined by Caney’s criticisms, and that the BPP is thus not undermined as a principle with at least some application to climate change. In my view this is good news, since the criticisms seems to apply to the APP as well. One agent may cause an action and be able to compensate, and a different agent may only be able (but perhaps more so than the first agent). How are we then to determine what share of the duty each is to have? And does the duty on the part of the able non-causally responsible agent spring to life only if the causally responsible agent is unavailable? I do not think these questions challenge the APP in any fundamental way.

Arguments in favour of the beneficiary-pays principle

From what has been said so far, it appears that to give the BPP a fair hearing, we should consider it in less complex contexts. Consider two examples provided by Axel Gosseries and Daniel Butt. Both examples are intended to support the claim that being a beneficiary may give rise to duties of compensation. Gosseries’ case involves

… a world with three island communities: the US community, the European Union, and Bangladesh. We assume that there is extensive trade between the US and the EU while trade barriers prevent any economic exchanges between the two former countries and Bangladesh. Moreover … the US has a heavily polluting industry while the EU would in fact not emit any CO2 into the atmosphere.… [A]ll of a sudden, due to a totally unexpected phenomenon, the US island gets completely flooded. The whole population dies instantaneously. One day after the flood, the EU receives new US products that had been sent before the flood. (Gosseries, 2004: 43) ¹³

A further, and vital, element of this case is of course that Bangladesh is, on the whole, harmed by US emissions, while the EU is not (Gosseries, 2004: 43).

Butt’s example involves four people, A, B, C and D, living together (but having little interaction) on a remote island.

…The only crop which will grow on the island is the extremely versatile Polychrestos plant…The Polychrestos plant’s root grows underground and is harvested each autumn…[T]he island’s underground river distributes water evenly throughout the island’s soil. Nonetheless, the Polychrestos plant is a high maintenance crop…In order for each person to support herself, she must produce 200 kilos of root per year. A is a very hard-working, industrious type, whose agricultural efforts…mean that she produces 700 kilos per annum…B, C, and D are rather laid-back in their approach to agriculture, and work just five hours a day to produce the minimum 200 kilos a year…D, a rather unsavoury character, decides she does not want to work even five hours each day. Unknown to all the others, she diverts the underground river away from B and C’s sections of the land, so that her land receives all of their water…When harvest time comes,…A harvests her regulation 700 kilos. C’s land has had no water, and consequently she has no crop…It also emerges that D (no water engineer) has in fact diverted the water away from her own land as well as that of C, and B…has been a beneficiary. To her surprise, she harvests 400 kilos. D… hangs herself with a rope fashioned from the last year’s Polychrestos crop. This leaves the problem of C. Without her year’s produce, C will die unless A and B provide her with the necessary 200 kilos. How should the remedial responsibilities be distributed? (Butt, 2007: 132)

What should we say about these cases? In Gosseries’ example it is difficult to completely abstract from two factors. Firstly, the EU and Bangladesh are (after the flood) the only agents on the scene, and secondly, given the names of these agents, and given that one of them is both harmed by pollution and cut off from trading relations with the others, it is natural to suppose that the former is rather affluent and the latter rather poor. Thus, many would accept that the EU should assist Bangladesh regardless of the causal history of their respective holdings. In other words, there are many good reasons to agree with the mandates of the BPP, but many of these reasons are not derived from that principle. Almost any plausible background theory of distributive justice would demand assistance, and it is therefore very hard to see whether or not a separate principle of corrective justice is independently at work.

In Butt’s case, many would be inclined to think that A is entitled to her holdings in virtue of her industriousness and effort. All those who think responsibility is relevant to distributive decisions are likely to think that A deserves what she has, but that B only deserves half of what she has, and that C deserves 200 additional kilos of Polychrestos. This judgement would not be altered if we removed D entirely from the story, and assumed that B’s good luck and C’s bad luck was the result of two unrelated natural events. Therefore, to the extent that we agree, it is unclear if we share Butt’s judgement because of responsibilitarian reasoning or because of the BPP. In addition, this example, like Gosseries’ is set up in such a way that the agent who is harmed ends up very badly off (he will die unless someone comes to his aid). This makes it hard to separate concerns stemming from distributive and corrective justice, respectively.

It is also worth noticing here that even if we consider less elaborate examples, the BPP has normative implications that overlap to a great degree with other, widely accepted, normative principles. Consider the following simplified case.

A benefits, B is harmed, and C is wholly unaffected, by D’s past (unjust) action. D is by now dead.

In the absence of further information, LE will very often point in the same direction as the BPP. According to LE, “…it is bad - unjust and unfair - for some to be worse off than others through no fault [or choice] of their own” (Temkin, 1993: 13).

Nearly all cases that, seen in isolation, activate the BPP, also activate LE, or related responsibilitarian theories. BPP scenarios always concern agents who are harmed or benefited by factors over which they have no control. Neither agent causes anything; they are both innocent receivers of the harms and benefits in question. I do not mean to suggest that the BPP is a part of LE, or that the theories are identical in either extension or justification. The point is just that if D harms B (and D then vanishes), and A is benefited as a result, LE would demand that A compensates B (if neither agent is in any way responsible for A’s harm). Of course, this depends on whether or not A and B prior to D’s act were justly equally well off. However, most BPP scenarios either assume that this is the case, or leave it open. In such cases the luck egalitarian impulse is to demand that A compensates B. ¹⁴

Butt (2012) also has a different case, however, which is less influenced by other concerns. This case takes its cue from an example provided by Fullinwider (2002) (originally intended as an argument against the BPP).

While I am away on vacation, my neighbour contracts with a construction company to repair his driveway. He instructs the workers to come to his address, where they will find a note describing the driveway to be repaired. An enemy of my neighbor, aware, somehow, of this arrangement, substitutes for my neighbor’s instructions a note describing my driveway. The construction crew, having been paid in advance, shows up on the appointed day while my neighbor is at work, finds the letter, and faithfully following the instructions paves my driveway. (Fullinwider, 2002)

Butt varies this case in order to render it “… as favourable to the BP[P] as possible” (2012: 6).

Before I went on vacation, I decided that I wanted my driveway to be repaired. I wrote to the very same construction company, asking them to come and repair my driveway on the day after I return from vacation, for the same price as my neighbour paid. I do not pay up front. On my return, I see my new driveway, and am delighted by it. I assume that the construction crew came on the wrong date, but conclude that it has worked out for the best. There is a knock at my door. I assume it is the construction crew, wanting to be paid. I open the door with a big smile on my face, holding an envelope of cash, wanting to pay the crew. It is my neighbour. He explains the situation. He asks if I will give him the envelope. (Butt, 2012: 6)

According to Butt, it would be wrong not to hand the envelope over to the neighbour, and the lucky owner of the new driveway has “…acquired rectificatory obligations to [the neighbour] as a result of having benefited from injustice” (2012: 6).

It seems right that the person whose driveway was repaved should compensate the person whose driveway was not. But I do not agree that this is because the first person has benefited from injustice (which he, in some sense, has). In the example, the lucky neighbour actually ordered the job done. In addition, the unlucky neighbour seems to have paid the same company in advance. Why else should he have any grounds for collecting the envelope? Now, absent the neighbour, if I order my driveway repaired by Friday, and it is repaired by Wednesday, it seems that I have to pay up in any case. This holds true even if the workers during the repair falsely assume that they are repairing another person’s driveway. So, regardless of any benefiting from injustice, the lucky neighbour owes the company money. However, the company owes the unlucky neighbour money. He has paid for a driveway repair that has not been delivered. To save time, the lucky neighbour might as well hand over the envelope, rather than paying the company which will then pay the unlucky neighbour. But the lucky neighbour, in my view, should not hand over the envelope because he has a duty to do so in virtue of having benefited from injustice.

One might, however, consider a slightly different scenario. What if the lucky neighbour intended to order his driveway repaired, but failed for some reason to do so? ¹⁵ In my view, no compensation is called for in this case. As most people would agree, the malicious character who replaced the instructions is the real perpetrator, but even if he manages to get away with it, the construction company has not repaired your neighbour’s driveway, and it seems that it is their responsibility to cover the damages. And even if the company suddenly goes bankrupt, it is still not your job to compensate your neighbour. If your intention was sufficient to create obligations in such cases, it would mean that if you intended to have your driveway repaired before you went on vacation, but then, while away, changed your mind and decided to put a swimming pool in your garden instead, you would be stuck with the driveway once you got home (assuming you cannot afford both). This is not to say that no fairness considerations at all are at play here. But it does not seem that you, in particular, have obligations to your neighbour in particular, over and above obligations to countless others whose misfortune is less directly linked to your unexpected benefit.

It is hard to argue conclusively that no cases can be devised in which benefiting from injustice gives rise to duties of compensation. But it is noteworthy that many of the cases explicitly intended to support the principle can be explained without reference to the BPP at all. Of course, this could still mean either that it is the BPP that does the work, or that it would have done the work, had the other factors been absent. But there is little reason to think so, given that these other factors are well-known and widely accepted (for instance assisting people in dire need, giving up stolen goods that happen to be in your possession, honouring contracts, LE, etc.).

As I noted above, in order to assess the BPP as a principle, it is necessary to consider examples in which most, or at least many, plausible candidates for distributive justice do not already imply the same result as the BPP. Both Gosseries’ and Butt’s examples seem too noisy in this respect. But even in less elaborate examples, further concerns might seep in.

Criticizing the normative basis of the beneficiary-pays principle

What is it that is supposed to give the BPP its normative force? My impression is that there are two considerations that are seen to count in favour of the BPP in particular (and that are not derived from more general views of distributive justice). The first is the sole fact that that there is a causal connection between two agents that are harmed and benefited, respectively, by the same action (Baatz, 2013: 97). The second is that compensation from the beneficiary to the victim in some sense restores the balance between the agents, or that it defeats injustice (Butt, 2007: 11). This consideration is based on corrective justice, and not on a general egalitarian notion of distributive justice. The idea is not that there should be a balance between all agents, but that injustice creates a particularly offensive imbalance that ought to be rectified. In the following, I will consider both these justifications. In the course of my discussion I will make occasional references to the agent’s welfare levels, in order to sharpen some of the cases that I consider. As noted, however, I assume throughout that all distributions (before and after the harmful act, and with or without BPP compensation) are compatible with some pluralistic and incomplete account of distributive justice.

The generalized beneficiary-pays principle

The BPP proponent could perhaps accept the criticisms presented so far, and instead advance a broader version, according to which any beneficiaries from unjust acts have special duties to compensate any victim of injustice. ¹⁹ This seems to accord quite well with the idea that injustice calls for rectification in a way accidents do not. Further, this acceptance would also have the welcome implication that the BPP no longer relies on the questionable causal connection criticized in the above subsection. It would incorporate both the claim that the causal connection is morally insignificant, and the claim that it is implausible to think that justice is rectified only by, or especially by, compensation going from a beneficiary of a particular act to the victim of that particular act. We can call this new principle the GBPP:

Any beneficiaries of unjust acts have duties of compensation (limited upwards to the value of the benefit) towards any victims of unjust acts insofar as no other agents, such as the performers of the unjust acts, have stronger duties of compensation and the capacity to honour them.

It is likely that many of us in the Western world are beneficiaries of injustice in some sense or the other. According to the GBPP, we should, if we had the choice, rather volunteer to help victims of injustice (from political oppression, say), than volunteer to help victims of freak accidents (an earthquake, say). This is not plausible. To rank people in dire need on the basis of the causes of their misery (on the condition that the causes include neither the victim nor the potential helper) seems disrespectful towards the victims of freak accidents.

This conclusion might be challenged. As the parentheses in the above paragraph indicate, many will accept that the perpetrator of harm has a stronger reason to assist his victim rather than a nearby victim of a freak accident. Thus it is not necessarily wrong to prioritize the victims on the basis of the causes of their misery. However, the status of the perpetrator is not comparable to the status of the beneficiary (of some injustice). The perpetrator can be said to owe his victim assistance, since he is the one who through his agency made sure that the victim was harmed. The beneficiary from any injustice does not stand in the same or even comparable special relation to any victim. ²⁰

It is also worth pointing out that injustice is quite often not addressed immediately. To resort to an illustration known from other contexts, we can imagine that two strangers are stranded on two different rocks in the rising tide, that you are the only potential helper, and that you only have time to rescue one of them. You have, in the past, benefited from injustice. As far as the two imperilled are concerned, their immediate situation is solely the result of a freak accident. However, one of them has, in the past, been the victim of injustice. According to the BPP, you should save him. This is, again, I think, an implausible conclusion. The rescuer’s status as a previous beneficiary of injustice, and the one stranded person’s status as a previous victim of injustice, does not seem relevant to the question of who to save.

Caney has suggested a supplementary justification for the GBBP. Instead of relying only on the idea that injustice requires rectification, he argues that one reason to single out A (and any other agents who have benefited from some injustice) as an agent with particular duties to compensate is that those whose holdings are not the result of injustice are entitled to their wealth, unlike those whose holdings are the result of previous injustice (Caney, 2010: 216). It seems, however, that this justification too has to rely, in addition, on the justification already discussed (otherwise it would be hard to single out the right recipients). Apart from that, the justification suffers from other problems. Consider a case in which A is the victim of an unjust act that benefited no one. Subsequently, B becomes the victim of a similar unjust act that benefits A. Should A then restore justice by compensating B, while the injustice of which she herself was a victim goes uncorrected? This too sounds implausible.

The most severe problem, however, is one that is common both to the BPP and the GBBP: There are quite a few ways of benefiting from injustice, and many of them appear not to imply that one is not entitled to one’s holdings or benefits. To be sure, there are many cases in which the BPP has some intuitive plausibility. Suppose that A takes B’s wallet, and that the wallet comes into C’s possession (through, presumably, a different act). C should clearly give back the wallet. This, I think, can be explained rather straightforwardly in terms of some justified system of property rights. But the BPP has a wider application. It also encompasses situations in which the benefit that accrues to some individual is not material, and does not in any clear sense belong to the victim. Suppose B has a giant tree in his backyard, and that he is within his right to have it there. The tree provides the shadow necessary to grow some delicate vegetables and flowers that he is fond of. Shadow-intensive gardening, moreover, is B’s favourite hobby. A and C, B’s neighbours on each side, both despise the tree, as it blocks the sun from their backyards, and prevents them from enjoying fully the pools they have both put in. (A prefers bathing in the morning, while B prefers bathing in the evening. That is why the tree bothers both of them.) After a while, A’s patience runs out, and one weekend, while both B and C are away, A cuts the tree down. Subsequently, A dies from the effort. B, of course, is devastated (over the loss of the tree rather than A’s death). C too is appalled by A’s act, and sorry for B’s loss, but also quite happy that the tree is finally gone. He can now enjoy swimming in the baking sun. B is clearly harmed by A’s unjust act, and C has clearly benefited from it. It is very unclear, however, that C owes B compensation, or that C is not entitled to either his holdings or his level of benefits or welfare.

Conclusion

It seems that the BPP does not only have limited application to climate justice, but that it is altogether an indefensible principle. It has consequences that are counterintuitive, and the two most plausible justifications that can be appealed to in its behalf, the normative relevance of causality, and the aim of restoring justice, do not stand up to scrutiny. Moreover, the generalized version of the principle, the GBBP, seems indefensible as well.