War and poverty

Abstract

Because the poorest people tend to die from easily preventable diseases, addressing poverty is a relatively cheap way to save lives. War, by contrast, is extremely expensive. This article argues that, since states that wage war could alleviate poverty instead, poverty can render war unjust. Two just war theory conditions prove relevant: proportionality and last resort. Proportionality requires that war does not yield excessive costs in relation to the benefits. Standardly, just war theorists count only the direct costs: the death and destruction wrought by war. This article argues that it can sometimes be appropriate to add the opportunity costs of a failure to alleviate poverty. Last resort is the condition that there must be no better alternative means of achieving the same just end for which the war is waged. This article argues that there are some cases in which alleviating poverty may constitute a better alternative. These are cases in which the most fitting description of the just end for war is sufficiently general that poverty alleviation offers a means to pursue it. The idea that poverty can sometimes render war unjust has, to date, been largely overlooked. It is, nevertheless, an idea with profound implications since, once taken seriously, war becomes much harder to justify. Wars that, in every other respect, seem just may prove disproportionate or unnecessary given the alternative of alleviating poverty.

This article argues that, since states that wage war could alleviate poverty instead, poverty can render war unjust. Poverty is arguably the world’s greatest evil. Addressing it requires, among other things, that greater resources be spent on programs benefitting the poor. Because the poorest people tend to die from easily preventable diseases, addressing poverty is a cheap way to save lives. The most cost-effective programs can save lives for as little as $3000 to $4000 (Weller 2015). War, by contrast, is expensive. Consider three recent wars: Afghanistan, Iraq and Libya. The US government estimates that from 2001 to 2009 it spent $444 billion on the war in Afghanistan and $806 billion on the war in Iraq (Belasco 2009, 2). Others contend that the US will end up spending three trillion dollars on the Iraq war alone (Stiglitz and Bilmes 2008). By comparison, intervention in Libya was cheap: the US spent $1.1 billion (Daalder and Stavridis 2012); the UK £212 million (The Huffington Post 2011). But intervention in Libya occurred during recession, when many of the governments that sanctioned intervention were cutting anti-poverty programs. If one accepts what these governments claimed, that the cuts were made because of insufficient funds, one must conclude that Libya diverted limited funds away from the poor.

The idea that a war could be wrong because states could do better addressing poverty is, I hope, an intuitive one. It is, nevertheless, an idea with profound implications since, once taken seriously, war becomes much harder to justify. Wars that, in every other respect, are just may prove unjust given the alternative of alleviating poverty. Recent wars that many ethically concerned people supported, such as Kosovo, Afghanistan and the Ivory Coast, might be of this kind.

Moreover, the idea that poverty can render war unjust is almost entirely neglected in modern just war theory. Michael Walzer’s Just and Unjust Wars makes no mention of it (Walzer 2006). In later work, Walzer is happy to declare the Afghanistan war just without first considering whether the sums spent waging it might have been better spent relieving poverty (Walzer 2004, 137). The philosophers who have briefly considered such thoughts have rejected them: their arguments are examined below.

While just war theorists have tended to neglect poverty as a possible objection to war, that objection can nevertheless be made by reference to two just war theory conditions: proportionality and last resort. Proportionality is the condition that the costs of war must not be excessive. Last resort is the condition that there must be no better alternative means of achieving the same ends. Poverty can render a war disproportionate (thus violating proportionality) because it is sometimes appropriate to include the costs of a failure to relieve poverty among the costs of war. Poverty can render a war unnecessary (thus violating last resort) because poverty alleviation sometimes offers a better means than war to achieve the same ends.

It is worth stressing from the start what the article does not argue for: the application of a utility maximising principle to the morality of war. Such an argument would contend that states must always pursue the policy that produces the greatest good and condemn states that wage war in preference to that policy. As we shall see, the idea that poverty can render war unjust has sometimes been discounted on the assumption that it must take a utilitarian form. One aim of this article is to show that one can object to war on poverty grounds without endorsing utilitarianism. The article takes seriously complaints against five features utilitarianism is commonly thought to possess. Since these features prove relevant below, it is worth recalling them here.

First, utilitarianism is said to be highly demanding, making no distinction between duty and supererogation. Utilitarianism may require agents to produce the greatest good even in cases in which they seem free to pursue their own ends. Second, utilitarianism is foundationally impartial. Partiality towards particular others can only be justified instrumentally. Utilitarianism would condemn a state for showing special concern for its own citizens, for instance, unless that concern was the best means to maximise utility worldwide. Third, utilitarianism endorses a monist theory of the good. It recognises just one ultimate good: utility (or, as on some variants, happiness or pleasure). Fourth, utilitarianism is indifferent to distribution. It sees no more reason to help people in need than well-off people if either option increases utility by the same degree. Fifth, famously, utilitarianism recognises no deontic constraints on the pursuit of the good. As such, utilitarianism does not share the same concern for distinguishing, as just war theorists typically do, between doing harm and allowing harm to occur, intentional and unintentional harm, innocence and liability.

Avoiding criticisms of utilitarianism, this article leaves space for supererogation, partiality, value pluralism, distribution and deontic constraints. Just war theorists cannot then dismiss the arguments presented here as relying on some alien moral schema. What the arguments rely upon, rather, is an interpretation of just war theory conditions. A war may be just even if an alternative policy of poverty alleviation produces a greater good, but it could not be just if that alternative renders war disproportionate or unnecessary.¹

One final preliminary. This article concerns objections to war from poverty, but one might wonder whether the arguments set forth have wider implications. If poverty can render war unjust, why not climate change, health care or education policy? My answer is that I am confident that other social issues could also provide objections to war, but proving this requires further work. Since poverty is arguably the world’s greatest evil, it seems a good issue with which to start.

1 Proportionality

The argument that poverty can render war disproportionate proceeds from the recognition of a problem concerning the proportionality condition. To apply the condition, we must make a proportionality calculation comparing the costs and benefits of war. But which costs warrant inclusion? Only the direct costs of war (the death and destruction wrought by military action) or opportunity costs as well (the costs of a failure to do what could be done in the absence of war)? Standardly, only direct costs are counted, but this can yield absurd results.

Consider the following example. State A is facing two crises in quick succession. First, the crops failed. To avert famine, the government needs to import food at significant expense. Then, a neighbouring state invaded a worthless strip of State A’s territory.

State A’s president meets her ministers to discuss the invasion. The ministers inform her that the military is in bad repair, relying on outdated equipment. A war could be won but would drag on, killing one thousand State A citizens. Call this proposed war the Long War.

The Long War, let us assume, is disproportionate. Reclaiming the territory is not worth a thousand lives. But the ministers have an alternative plan. They propose refitting the military with fancy new equipment. Once refitted, the military could recover the territory swiftly, with minimal casualties. Call this proposed war the Short War. The Short War has a downside. Refitting the military will be expensive; so expensive, in fact, it will leave State A without funds to avert famine. A million people will starve.

It seems clear that if the Long War is disproportionate, so is the Short War. The proportionality condition is there to prevent wars yielding excessive costs. If a thousand deaths are excessive, a million certainly are. State A’s ministers might defend their plan by noting that, standardly, the proportionality condition counts only the direct costs of war, but all this shows is that the ministers should be fired and the standard approach revised.

Opportunity costs should sometimes be included in a proportionality calculation. Should they always? Consider another example. Like State A, State B has lost territory to an invader. State B, however, is wealthy. It has funds to meet its people’s needs and recover its territory. But here is an alternative plan: rather than wage war it could use the funds to buy everyone in the world a new TV.

If State B wages war to reclaim its territory, the never gifted TVs represent an opportunity cost of a kind. Yet this opportunity cost should not be included in a proportionality calculation. Why not? The answer can be found in the first anti-utilitarian idea considered above: the distinction between duty and supererogation. Since State B is not obliged to buy everyone a new TV, that opportunity cost should not count as a cost of war.

We can see then that only some opportunity costs should be included in a proportionality calculation and we have a way to determine which. An opportunity cost should be included when it represents a duty violation (Mellow 2006). State A has a duty to prevent famine; State B has no duty to buy TVs. Why is it duty that matters here? Because unjust wars are not merely suboptimal; they are unjust. They violate moral requirements. A failure to do what is supererogatory violates no moral requirements.

Consider another route to the same point. Distributive justice tells us what, morally speaking, belongs to whom. State A’s citizens are entitled to subsistence; the Short War would take what is theirs. The would-be recipients of State B’s TVs, by contrast, are not entitled to the TVs; state B’s war takes nothing of theirs. No doubt, it can sometimes be permissible to take what belongs to others, but the costs of doing so must be included in a proportionality calculation. The failure to include them involves a kind of moral accountancy dodge; hiding the full costs of war by treating resources owed to others as ‘free money’ to be used without consequence. The opportunity costs of war to those denied their entitlements are as real and important as the direct costs wrought by bullets and bombs.

From this, we see that poverty will render war disproportionate under three conditions:

• (1) The warring state has a duty to alleviate poverty.

• (2) The warring state cannot fulfil its duty to alleviate poverty if it wages war (at least not without violating some further duty).

• (3) The costs of failing to fulfil the duty to alleviate poverty are so great that, when added to the direct costs, they outweigh the benefits of waging war.

A comment on each condition. Condition (1) demonstrates the difference between the proportionality argument and the utility maximising argument. The proportionality argument does not assume that states must pursue the greatest good, but merely that states have a duty to alleviate poverty. The proportionality argument thus leaves space for supererogation.

That states do have duties to alleviate poverty is the predominant view in the distributive justice literature. Disagreement arises as to how demanding the duty is. On some accounts, it is highly demanding (Singer 2009; Unger 1996); on other accounts, only moderately so (Cullity 2004; Miller 2004; Murphy 2000). I shall assume here that state do have duties to alleviate poverty and will briefly defend this view in Sect. 3. On the question of demandingness, I remain neutral between the rival accounts.

It is worth noting that the duty to alleviate poverty is not necessarily the same as a duty to send aid. Two points brings this out. First, there are other, possibly more effective, means to reduce global poverty, such as changing international rules governing trade, investment and migration or assisting domestic reforms (Moyo 2009; Risse 2005; Wenar 2008; Easterly 2006). Second, global poverty is not our only focus. States have duties to alleviate poverty at home as well as abroad. Indeed, it might be that the former takes priority.

Whatever methods we adopt and whatever our focus, alleviating poverty is likely to carry costs. Changing rules and restructuring institutions is costly for those states and individuals who benefit from the status quo. Where there are costs, it is possible that funding poverty alleviation will prove incompatible with waging war.²

(2) If a state can afford to wage war and fulfil its duty to alleviate poverty, without violating some further duty, then the costs of unalleviated poverty should not count as costs of war. To count, the two must be incompatible. But when are the two incompatible? It might be thought that incompatibility only arises if we interpret the duty to alleviate poverty as highly demanding. If we interpret the duty as moderately demanding, then one might think that states can do both. Section 4 demonstrates that this is not necessarily the case.

(3) Even if war prevents a state from fulfilling its duty to alleviate poverty, war would still be proportionate if the benefits were so great as to outweigh the costs, including the costs of unalleviated poverty. A state does nothing wrong in failing to alleviate poverty if some greater good is thereby achieved. But how should the benefits of war be weighed against the costs of unalleviated poverty?

2 Weighing the goods of war and poverty alleviation

Proportionality calculations are difficult irrespective of whether one includes the costs of unalleviated poverty among the costs of war. One central problem is determining how factors of different kinds should be weighed against one another. For instance, if the defence of territory counts for something, how much does it count against the death and destruction wrought by war? This article is silent on such issues not because they are unimportant—they are crucial—but because they are not unique to the topic raised here. What we need to consider is whether there is any reason to weigh the same type of benefit (or cost) differently depending on whether it is generated by war or poverty alleviation. I shall take as my example the benefit of saving lives. Human lives might not be the only thing worth caring about, but, clearly, they are of great importance.

Suppose a war saves lives by, say, preventing terrorism or ending despotism. How should those lives be weighed against lives lost due to unalleviated poverty? A straightforward answer is one-for-one. Since alleviating poverty is typically more cost-effective at saving lives than waging wars, this answer makes it more likely that war will prove disproportionate. But is this answer right?

Some might argue that lives saved by war should be awarded additional weight since war prevents intentional harm rather than harms that are natural or unintentional. Such an argument appeals to at least one of two moral doctrines: the doctrine of doing and allowing (DDA) and the doctrine of double effect (DDE). DDA draws a distinction between doing harming and merely allowing harm to occur, deeming it worse to do harm. DDE draws a distinction between intentionally harming and merely harming as a side effect, deeming it worse to intentionally harm. From DDA and DDE, one might seek to deduce the philosophical claim that it is more important to save people from intentional harm than natural misfortune or unintentional harm. Combine this philosophical claim with the empirical claim that poverty is something natural and/or unintentional, while war prevents intentional harm and one arrives at the conclusion that lives saved by war deserve additional weight.

This argument goes wrong in several places. The empirical claim is problematic. While there is fierce disagreement among economists as to the causes of poverty, there is wide agreement that human injustices, such as violence, discrimination and misgovernment, play a crucial role (Rodrik et al. 2004; Collier 2007; Sen 1981). Poverty is a harm caused by humans, not a natural misfortune. The claim that it is an unintentional harm may seem more plausible, but even this does not always hold. Sometimes people are poor because the powerful intended them to be so. In extreme cases, such as Stalin’s USSR, leaders have intentionally starved the inhabitants of disobedient regions. In other cases, such as the US, minorities have been impoverished by a history of discrimination intended to keep them down.

The philosophical claim is also controversial. It is not obvious that we have greater reason to save people from intentional harm than natural misfortune or unintentional harm. Consider a case in which you can save 100 people from an arson attack or a 101 from a fire caused by lightning. It seems clear you should save the 101. We might conclude from this, as several philosophers have, that the type of threat victims are subject to is an entirely irrelevant factor when deciding whom to rescue (McMahan 2010, 60–62; Singer 2010; Parfit 1984, 46; Murphy 2000, 126–127). Alternatively, we might conclude that threat-type is relevant but not especially important. Perhaps we should always save the greater number but prioritise those threatened with intentional harm when the numbers are equal; save 100 arson victims over 100 natural fire victims, for instance (Tadros 2011, 105–106). It does not really matter which conclusion we draw. When it comes to proportionality, they deliver near identical results.

Note that in making this argument we are not opposing DDA or DDE. These doctrines are primarily there to guide us in decisions over when we may permissibly harm. They do not have obvious implications for rescuers. One can consistently hold that there is something especially wrong about harming and intentional harming, without believing that, in rescue cases, those who have been wronged in these ways should be prioritised over the victims of other misfortunes (Murphy 2000, 126–127).

Let us turn to another reason to question a one-for-one trade-off between lives saved by war against lives lost due to unalleviated poverty. Some argue that states should be partial towards their own citizens. This view has implications for how we weigh lives saved by war against lives saved by poverty alleviation when the nationality of the victims differs. It will not, however, always direct us to award lives saved by war greater weight. To see this, consider the four following choices a government may face:

• (1) Saving citizens by waging war versus saving foreigners by alleviating poverty abroad.

• (2) Saving citizens by waging war versus saving citizens by alleviating poverty at home.

• (3) Saving foreigners by waging war versus saving foreigners by alleviating poverty abroad.

• (4) Saving foreigners by waging war versus saving citizens by alleviating poverty at home.

Only in case (1) does the idea of partiality towards citizens direct us to award the lives saved by war greater weight than lives saved by poverty alleviation. In cases (2) and (3), there is no effect, and in case (4), partiality would add additional weight to lives saved by poverty alleviation.

This section has not sought to rule out the possibility that war could prove proportionate even if it saves fewer lives than it costs. Saving lives is not the only thing that matters. There could be other benefits of war that warrant inclusion within a proportionately calculation (e.g. defending territory or deterring aggression). There are also, let us note, often other benefits of poverty alleviation (e.g. improved health and education). A full proportionality calculation would weigh all relevant benefits and costs. What this section has argued is that a given benefit, saving lives, is of equivalent importance whether it is achieved by poverty alleviation or by war. This makes it more likely that the costs of unalleviated poverty will render war disproportionate.

3 Defending a duty to alleviate poverty

The idea that poverty can render war unjust has not been recognised by just war theorists. The theorists that have considered it have swiftly rejected it. Thomas Hurka, for instance, discounts the idea that the Gulf War would have been disproportionate even if “the money it cost would have done more good if it had spent on development aid to Africa” (Hurka 2005, 41). McKim and McMahan come to a similar conclusion: for them the Gulf War would not have been rendered unjust even if aid would have saved more lives than the total population of Kuwait (McMahan and McKim 1993, 526).

Why do these theorists reject a poverty-based objection to war? One reason seems to be a failure to take the idea of a duty to alleviate poverty seriously. These theorists seem to have assumed that a poverty-based objection must take the form of the utility maximizing argument I bracketed at the start (Hurka 2008, 134; McMahan and McKim 1993, 526). That argument holds that states should always adopt the policy that produces the greater good. These theorists reject this argument as “excessively restrictive” (McMahan and McKim 1993). But as we have seen, the utility maximising argument is not the only relevant argument here. The argument presented above, based on a duty to alleviate poverty, leaves space for supererogation.

Why have just war theorists overlooked the idea of a duty to alleviate poverty? Do they reject it? Interestingly, the one philosopher that comes closest to making the proportionality argument, Mellow (2006, 447–451), does air skepticism. Mellow accepts that if a state has a duty to alleviate poverty and war prevents it from fulfilling that duty, then the costs of unalleviated poverty should be included within a proportionality calculation. Nevertheless, he claims that the “commonly held view” is that an elected leader has no duty to alleviate poverty (Mellow 2006, 450).

[A]ssuming she is duly elected or appointed, and is acting within the scope of her role, [she] generally has the permission to use (on behalf of her fellow citizens) public resources in a wide range of ways. Spending decisions by political leaders are often rightly criticized, even on moral grounds, but it is generally accepted that a leader has broad discretion in spending public resources, and, in particular, does not generally violate her duties if she spends those resources in ways that do not maximize the benefit for the poor (Mellow 2006, 450).

Mellow does not say whether he supports this “commonly held view”, but he does not reject it. Below, I shall explain why the view is misguided. First, let me query whether the view is even commonly held. As I have noted, the view is not commonly held among theorists of distributive justice. But nor does it seem to be commonly held among the public. The evidence suggests that people do think that the government is obligated to take some action to alleviate poverty (Miller 1992, 573).

There is, however, a view that is commonly held that is suggested in the above quotation: the idea that a leader has political authority to decide how revenue is spent. In other words, Mellow’s mistake is to muddle authority with distributive justice. It is a question of authority whether a leader has the right to rule on a given issue, e.g. public spending. It is a question of distributive justice whether a leader has a duty to spend revenue in a particular way, e.g. alleviate poverty. The two are distinct. A leader can have a right to rule and a duty to alleviate poverty. Authority does not then entail moral permissibility. What precisely authority does entail I shall leave aside although good candidates include (i) others have a duty to enact the leader’s decisions, (ii) no one should interfere with their enactment, and (iii) the leader should not be deposed or punished for her decisions.

Once authority and distributive justice are distinguished, the “commonly held view” loses its plausibility. A leader’s right to rule how revenue is spent does not entail the absence of a duty to alleviate poverty. A leader can make an authoritative decision to wage war and violate her duty to alleviate poverty.³

Mellow's argument against a duty to alleviate poverty fails. Perhaps this is unsurprising since, had it succeeded, a widely-held view about distributive justice would have been prone to easy refutation. So why then have just war theorists not taken the idea of a duty to alleviate poverty seriously? Perhaps the answer is that, for too long, the philosophical debate regarding justice in war took place in isolation from the philosophical debate regarding distributive justice. Two literatures developed upon two related topics with little attempt at integration. This was surely a mistake.⁴

4 Incompatibility

While states have a duty to alleviate poverty, this duty could only render war disproportionate if waging war and fulfilling the duty were incompatible. Incompatibility is likely to arise if the duty to alleviate poverty is interpreted as highly demanding. If the duty to alleviate poverty requires a state to devote most of its resources to that end, then the state will struggle to find resources to wage war as well. Many people, however, do not interpret the duty as this demanding. In their view, the duty to alleviate poverty only requires people to make a moderate sacrifice for others. This moderate duty to alleviate poverty could be fulfilled while leaving surplus revenue for other projects, such as waging war.

In this section, I will argue that, contrary to what one might initially expect, incompatibility can arise even if we interpret the duty to alleviate poverty as only moderately demanding.⁵ Let us start by considering what I take to be the most plausible argument for a moderate duty. This argument holds that there are limits to the sacrifices people can be expected to make to help others. Each person has her own projects; her own life to lead. We should not expect people to act as saints or heroes, surpassing ordinary levels of human generosity (Wolf 1982; Scheffler 1982; Cullity 2004; Miller 2004).

If individuals cannot be expected to make grave sacrifices to help others, individual taxpayers cannot be expected to make grave sacrifices to help the poor. So clearly this argument, if accepted, would support the idea of a moderate duty to alleviate poverty. But notice that the argument applies generally, to all forms of altruism. People cannot be expected to make grave sacrifices for others by way of any altruistic activity, not just poverty alleviation.

Another implication of the argument is that governments should not force people to make grave sacrifices to help others. For if people cannot be expected to help others at a certain cost to themselves, it would seem wrong for a government to force them to provide this help (Haydar 2002; Murphy 2000, 82–84). On the moderate duty view then, governments are not only permitted but also required not to force their citizens to make grave sacrifices to help others. Governments can tax citizens to enforce moderate duties, but they must stop there.⁶

Now consider war. When states wage wars, taxpayers are forced to fund them. How can governments justify imposing these costs? Some wars are arguably in the interests of all taxpayers and can be justified as such. Many wars, however, require some or all taxpayers to make sacrifices to help others. If these wars can be justified at all, they are justified as a means to assist a different group of people to the taxpayers who pay for it. In this sense, these wars may be referred to as “altruistic wars”. (The term should not mislead. An altruistic war is defined by the strongest justification for war not what motivates people to wage it. The leaders of a country may declare war for self-interested reasons, but still the war could be altruistic in the defined sense).

Wars of humanitarian intervention, such as Kosovo, Libya and the Ivory Coast, are perhaps the most obvious examples of altruistic wars. Less obvious examples are wars of national self-defence that, in actuality, only constitute a net-benefit for a minority of citizens. Suppose, for instance, that a state goes to war against terrorists targeting a subgroup of its citizens, such as an ethnic minority. The strongest justification for such a war is to defend the subgroup, not all taxpayers. A slightly different example is a war like Afghanistan that aims to confront terrorists who do constitute a threat to all taxpayers but a threat so minimal that it seems unlikely that a self-interested taxpayer, who understood the risks, would voluntarily contribute. The strongest argument for forcing US taxpayers to contribute to a war as expensive as Afghanistan is not “this is in your interests” but rather “you have a duty to protect those people who would otherwise be killed”. The Afghanistan war is thus also an “altruistic war” in the defined sense.⁷

Let us now bring these points together. We have seen that the most plausible argument for a moderate duty to alleviate poverty implies that there are limits upon the level of altruistic sacrifice that governments can force taxpayers to make. We have also seen that when governments wage altruistic wars, they force some or all taxpayers to make altruistic sacrifices. From this, it is clear that a moderate duty to alleviate poverty can prove incompatible with war. If the argument for a moderate duty to alleviate poverty is right, the revenue that governments can justly collect for altruistic projects is limited. Altruistic wars draw from this limited revenue. If a government spends what it is permitted to raise for altruistic causes on altruistic war, it cannot raise additional revenue to fulfil its citizens’ duty to alleviate poverty, since doing so would involve forcing taxpayers to make a greater sacrifice than is required of them. Waging altruistic wars consumes revenue that could be spent on poverty alleviation and since this is so, when we make a proportionality calculation, the costs of unalleviated poverty should be included among the costs of altruistic war.

It might be objected that often when a government wages war, its citizens support it. If the citizens support the war, the government cannot be said to be forcing them to make altruistic sacrifices. If there are limits to the revenue governments can force from taxpayers to fund altruistic purposes, such wars do not draw from it. Popular wars and a moderate duty to alleviate poverty are compatible.

Two replies. First, even when a state does have the support of most of its citizens, there will always be dissenters. These dissenters are forced to pay for a war they do not support. If there are limits to what states can force their citizens to do for others, then, by taxing dissenters to pay for an altruistic war, governments deplete the resources they may spend on poverty alleviation. The existence of dissenting minorities shows that incompatibility remains.

Second, even in the case of the pro-war majority, altruistic war is incompatible with, what we might be term, a “conditional duty to alleviate poverty”. By a conditional duty to alleviate poverty, I mean a duty to alleviate poverty that is conditional on a willingness to make an altruistic sacrifice. If citizens are willing to spend X dollars on an altruistic war, but spending X dollars alleviating poverty would produce a greater good, then citizens may have a conditional duty to alleviate poverty. This is true even if the citizens have no duty to spend X dollars on either project were they unwilling to make an altruistic sacrifice.

To understand this point, consider the following example. Suppose you are escaping from a burning building. While escaping, you encounter a baby and a caged parrot. You hesitate. Escaping the building is not easy. There are fallen timbers in your path and to clamber past them quickly you need free hands. You are sure to escape if unencumbered but risk death if you attempt to carry either with you. It seems plausible that you have no duty to rescue. Suppose that, despite the risks, you act heroically. You rescue the parrot. Once outside, you face shocked onlookers demanding an explanation. Have you acted wrongly? Absolutely. It was the risks of carrying something that enabled you to reject a duty to rescue the baby. In choosing to run those risks, you became obliged to rescue the baby.⁸

The lesson of the example is that people can have a duty to do something conditional on them being willing to expend a cost, even if they otherwise had no duty to expend that cost (Fried and Parfit 1979, 295–296; Kagan 1989, 240; Pummer 2016, 92–93). In this way, citizens may have no duty to alleviate poverty but still have a conditional duty to alleviate poverty rather than wage an altruistic war.

Now in making this argument, I do not mean to suggest that people always have a conditional duty to do the most good. Just as people can sometimes escape ordinary duties on grounds of cost, so people can sometimes escape conditional duties on grounds of cost (Pummer 2016, 92–93). Nevertheless, one cannot escape conditional duties on grounds of mere preference. In the fire example, you have a conditional duty to save the baby even if you prefer parrots to babies. Likewise, citizens cannot escape a conditional duty to alleviate poverty by merely preferring war to poverty alleviation.

To summarise: incompatibility between a state’s duty to alleviate poverty and waging war can arise even if we interpret a state’s duty to alleviate poverty as only moderately demanding. Some wars are altruistic wars in the sense that, if justified at all, they are justified as a means to assist a different set of people than those who fund them. We have also seen that the argument for a moderate duty limits the amount a government can spend on any altruistic project, including altruistic war. A dollar spent on altruistic war is thus a dollar less for poverty alleviation. Even if most citizens of a country support an altruistic war, this changes nothing as long as (1) a minority oppose it and/or (2) the majority have a conditional duty to alleviate poverty instead.

5 Last resort

Having presented the proportionality argument, let us turn to last resort. To see how poverty can render war unnecessary we need to appreciate both the role that last resort plays within just war theory and an often-overlooked problem concerning this condition.

To see the role that last resort plays, it is helpful to imagine that the other just war conditions have first been fulfilled. Imagine, then, that a state has a just cause for war. This means that it has a just end: some goal that is so important it could be worth waging war to achieve. It also means that there is some state or sub-state party that has done something to make itself liable to attack, such as engaging in internal or external aggression. Imagine also that the war is proportionate: the costs are not excessive. When these, and all other,⁹ relevant conditions have been fulfilled, we know the state has reason to pursue the just end for war ahead of its ordinary business.¹⁰ What is left to determine is whether there is some better means to achieve that end other than war. That is the role of last resort.

The process of applying last resort is then a process of exploring alternatives. If we find some alternative means of achieving the same just end that is preferable to war—for instance, one that does not involve killing—then war is unnecessary and should not be waged. But here we come to the problem. To decide whether an alternative really does achieve the just end for war, we need a particular description of that end and there are many possible descriptions. The choice of description is crucial. The more specific the description, the harder it is to find an alternative to war; the less specific, the easier. We need some method to choose. But what is that method? To help find it, consider some easy cases.

Imagine a state has invaded a part of its neighbour’s territory. The victim state has a possible just end for war. How should that end be described? Here are two candidates:

• (1) To recover its occupied territory.

• (2) To recover its occupied territory courageously.

The difference between these two descriptions could prove pivotal. Suppose that diplomacy could recover the occupied territory but would not, as a war would, allow the victim state to demonstrate courage. In that case, a war would fail last resort on the first description but not the second.

Clearly, the first description is the relevant one; the second is not. Why not? Because the extra specification the second description introduces is not something for which any war should be waged. Here this is obvious, but in harder cases it is helpful to consider two tests. In both tests, we focus on the difference between the descriptions. In the first test, we ask whether the difference represents anything the warring party has reason to care about. Call this the “Test of Moral Salience”. In the second test, we ask whether the difference is something the warring state has such strong reason to care about that it could be worth waging war to achieve. Call this the “Test of Sufficient Moral Salience”. It is this second test which really matters. The first test is just a helpful aid. If a difference does not past the first test it will not pass the second.

In this case, the difference between descriptions (1) and (2) fails the second test, whether or not it passes the first. The demonstration of courage is not something worth waging war to achieve. Since the difference fails the second test, description (2) should be rejected.

It is worth emphasising that when we apply either test, we do so from the standpoint of the warring state. To bring this out, consider an example involving two victim states: V1 and V2. Both have lost territory to the same invader. There are at least two ways of describing V1’s just end for war:

• (1) To recover its occupied territory.

• (2) To recover occupied territory.

Imagine V1 can only recover its own territory by means of war but could recover V2’s territory by means of diplomacy (the invader is more willing to cede V2’s territory than V1’s). In that case, given description (2), a war to recover V1’s territory would fail last resort. But (1) remains a plausible description. The difference between (1) and (2) arguably passes both tests. The opportunity to recover its territory is something V1 has reason to care about. No state need regard the recovery of another state’s territory as a satisfactory alternative to recovering its own. So, the Test of Moral Salience is passed. And the opportunity to recover its own territory is arguably something of such moral importance to V1 that it could be worth waging war to achieve. So, the Test of Sufficient Moral Salience might also be passed.

These are easy cases because the two tests deliver descriptions that we would have intuitively accepted. But there are harder cases. Suppose, for instance, that state S is considering whether to intervene against foreign regime R to prevent the violent repression of R’s citizens, the Cs. The Cs, let us imagine, number one million. There are several ways to describe the possible just end for war. Let me list four, in order of specificity, from the narrowest to the broadest:

• (1) To protect the lives and basic wellbeing of those Cs who, in the absence of intervention, would be subject to violent repression by R.

• (2) To protect the lives and basic wellbeing of Cs against violent repression by R.

• (3) To protect the lives and basic wellbeing of Cs.

• (4) To protect the lives and basic wellbeing of up to one million people.

What are the differences between these descriptions? Compare (1) and (2). Description (2), or something like it, is standard. It is less specific than (1) since it does not specify the particular Cs that S is seeking to protect. On description (2), war would be unnecessary if there were a diplomatic solution that leaves Cs in general better protected, with fewer falling victim to R. Description (1), by contrast, insists that the same people who are protected by means of war be protected by war’s alternatives. On that description, a diplomatic solution that left fewer victims overall would not be a relevant alternative if the people who would be protected if war were waged would be killed or injured while a diplomatic solution is reached. War would satisfy last resort despite the possibility of diplomacy.

The difference between description (2) and descriptions (3) and (4) is more readily apparent. The latter descriptions do not specify a kind of threat. On these descriptions, intervention would be unnecessary if the intervening state could better protect people’s lives and basic wellbeing by other means, whether or not those other means addressed the threat posed by R. Description (4) goes further by not specifying that the people protected must be Cs.

It can be tempting to assume that (2) must be right because it is the kind of description that is standardly given. But clearly the matter should not be settled by mere tradition. We should instead apply the two tests described. If the standard description includes what is morally irrelevant or insufficiently salient to be worth waging war over, we should reject the standard description.

Adopting this approach, we see that just war theory has been correct to reject (1). The difference between the broader description (2) and the narrower description (1) is the individual identity of the victims. This is not morally relevant to the warring state. It has no more reason to protect those Cs who, in the absence of intervention, would be victimised by R over any other group of Cs. Of course, the victims themselves have reason to worry who precisely is protected. But, when we compare descriptions, we must do so from the standpoint of the warring party. It is questions concerning what the warring party should do, not what victims should do, that we are seeking to answer. So the difference here—victim identity—fails the weaker Test of Moral Salience. Having failed that test, it will certainly fail the stronger Test of Sufficient Moral Salience. The standard view is right to regard a diplomatic solution that protected more Cs as an effective alternative to war, even if it did not protect the same Cs as would be protected by war.

Moral salience can explain why we should choose description (2) over (1). But why stop at (2)? In Sect. 2, we found that the type of threat is either morally irrelevant to rescue cases or, if relevant, is of minimal importance. If the former is true, description (2) fails the Test of Moral Salience; if the latter, it fails the Test of Sufficient Moral Salience. Either way, our description of the just end for war here should not refer to a particular threat. Nor is the nationality of victims morally salient in this case. As we also saw in Sect. 2, the nationality of the victims is only salient if the warring state has reason to be partial towards a particular nationality, as it does in the case of its own citizens. Assuming that S has no special tie to the Cs, their nationality is irrelevant.

Putting these points together, we see that description (4) is the right choice in this case. Description (4) covers what is salient and excludes what is irrelevant. That a war may protect people’s lives and basic wellbeing is morally salient, as is the numbers that could be protected (up to one million). The nationality of the victims and the threat they are under, however, is either completely irrelevant or insufficiently important to warrant inclusion in the description of the just end.

Having presented the case for description (4), we see how poverty can render war unnecessary. Suppose S can protect the lives and basic wellbeing of up to one million people by funding poverty alleviation. If so, it does not have to wage war to obtain this end; poverty alleviation constitutes a better alternative.

Description (4) is radically different from how we would standardly describe the just end for war. Is there nothing that could be said in defence of the standard description? It might be objected that (4) fails to take account of the identity of the victims, but as we have seen, this is also true of the standard description (description 2). Description 2 is correct not to prize the lives of one group of Cs over another group of Cs, but then, from the perspective of S, there is similarly no reason to prize the lives of Cs over other foreigners.

Another objection might hold that (4) is too permissive. If “to protect the lives and basic wellbeing of up to one million people” were a just end for war, then, it might be argued, states could launch a war against anyone to achieve these ends, including against states that are not internally or externally aggressive. This seems unacceptable.

This objection fails, however, once we recall that there are other conditions a just war fulfils besides pursuing a just end. As indicated above, just cause involves more than just ends. A state will only have a just cause if the state or sub-state party it is targeting has done something to make itself liable to attack. On most accounts, the target must have engaged in internal or external aggression. So, we need not worry that in adopting a broad description of the just end for war we permit attacks on non-liable targets.

To this it might be objected that the distinction between just ends and just cause is unnecessary if we keep to a standard description like (2). Standard descriptions, this objection holds, are so narrow they can only be achieved by attacking a liable target. But this is simply not true. Consider description (2): “To protect the lives and basic wellbeing of Cs against violent repression by R”. There could be other ways to achieve these ends than by attacking R. Suppose R’s neighbouring state is a back route into R’s territory or the trading partner upon which R depends. If so, attacking that neighbouring state may be the easiest means to prevent R’s repression.

Consider another example. Early in the Second World War, Winston Churchill demanded an attack on Norway, a neutral power, to cut off Germany’s supply of iron ore. Many at the time, and many since, deemed Churchill’s plan unjust. Why? A standard description of Churchill’s objective would be “to stop Nazi aggression”. That seems like a just end for war. If there was a problem with Churchill’s plan it was not his ends but his means. Norway, being neutral, was not liable to attack.¹¹

We can conclude that standard descriptions of just ends do not make a liability condition redundant. A further liability condition is required and, once added, will prevent wars against non-liable targets whatever description we select.

The underlying point here is that what is relevant to just cause need not be relevant to last resort. When we are thinking about just cause, our focus is on war. Since war involves intentional attack, we want to ensure that the target is liable to attack. When we are thinking about last resort, by contrast, we are exploring alternatives to war. What we want to know is whether these alternatives can achieve what we ultimately care about. What we ultimately care about need not be thwarting the liable party. It might be, as in our case of the Cs, something much broader, such as protecting people’s lives and basic wellbeing. We should not then muddy our conception of what we ultimately care about—our just end—with questions of liability just because questions of liability are crucial to just cause. Ensuring a just cause and exploring alternatives are separate activities demanding separate criteria.

6 Last resort, national self-defence and utility maximisation

How far does the last resort argument extend? So far, I have presented it using an example of humanitarian intervention. What about wars of national self-defence? Whether the argument applies to cases of national self-defence depends on how exactly we specify the ends of war. Consider the war in Afghanistan. One of the primary justifications given for that war was defence against a terrorist threat. If states have special reason to care about the lives of their citizens but not the kind of threat posed to their citizens, the best description of the just end for the Afghanistan war would actually be (something like) “the protection of the lives and basic wellbeing of US citizens from a certain level of threat”. If we describe the just end for war in this way, then war was not a last resort. The US government had an alternative means to protect the lives and basic wellbeing of US citizens: it could have done more to alleviate poverty at home.

The threat that poverty poses to US citizens is many times larger than the threat posed by terrorism. Poverty is one of America’s biggest killers, accounting for 4.5% of American deaths (Galea et al. 2011). Terrorism, by contrast, is an extremely low risk. The chance of an American falling victim to terrorism is around one in 3.5 million per year (Mueller and Stewart 2012, 103). How many additional American lives might have been saved had the US government chosen to focus its resources on tackling deprivation at home rather than terrorism abroad?¹²

Although the last resort argument could extend to wars of national self-defence, this is not always the case. There are some ends which cannot be achieved by poverty alleviation. Poverty alleviation cannot deter aggression, secure territory or uphold political independence. If such ends are themselves just ends for war, then there are some just ends that cannot be achieved by poverty alleviation. It might even turn out that the war in Afghanistan satisfied last resort. For there may be other just ends the war pursued (law enforcement or punishment, perhaps), which could not be achieved by alleviating poverty at home.

The fact that there are some just ends which cannot be achieved by poverty alleviation helps to emphasise the difference between the last resort argument and a utilitarian argument. A utilitarian description of the just end for war would be something extremely broad: something like, “to increase overall utility” or “to increase overall utility by X amount (where X is the amount of net utility the war would achieve)”. Under such descriptions, poverty alleviation could always constitute an alternative to war.

A final challenge therefore remains: to explain why I have not adopted a utilitarian description. The answer lies in the anti-utilitarian ideas reviewed in the introduction. Three ideas prove critical. First, partiality. If states are sometimes right to award significant weight to their own concerns, then we need not always accept impartial descriptions of the just ends for war. We saw this in the case of V1 and V2 from the previous section. Second, value-pluralism. If utility is not the only thing that ultimately matters, there is no reason to insist that just ends should only refer to utility. Sovereignty, territory, deterrence, punishment or the rule of law might also be goods of ultimate value and of such importance as to be worth waging war to achieve.¹³ Third, distribution. It may be extremely important to help people in need. It may be comparatively unimportant, even trivial, to help people who are well-off increase their utility by the same degree.

This last point helps to explain the refusal to adopt a utilitarian description even in the case of humanitarian intervention considered in the last section. There, I argued for a broad description of the just end for war, but stopped short of something even broader. Compare (4) with (5):

• (4) To protect the lives and basic wellbeing of up to one million people.

• (5) To increase overall utility by X amount (where X is the amount of net utility the war would achieve).

Description (4) focusses on people in need. Description (5) does not. If it is extremely important to help people in need, the difference between (4) and (5) will pass the Test of Sufficient Moral Salience.

7 Conclusion

The article has made two arguments. In these concluding remarks, let me bring them together to explore their similarities and differences. Differences first. Each argument applies to a different range of cases. The proportionality argument only applies to cases in which a state has a duty to alleviate poverty that it cannot fulfil if it wages war. The last resort argument only applies to cases in which alleviating poverty offers an alternative means to achieve the (appropriately described) just end for war.

The proportionality argument can apply to cases to which the last resort argument does not: cases in which alleviating poverty offers no alternative means to achieve the just end for war. Consider, for instance, the case of State A we considered in Sect. 1. I argued there that a war to regain its territory at the cost of famine would fail proportionality. Note here that it would not fail last resort. After all, averting famine is no means to regain territory.

Similarly, the last resort argument can apply to cases to which the proportionality argument does not: cases in which a state has no duty to alleviate poverty or cases in which fulfilling such a duty and waging war are compatible. To take the latter possibility, suppose state S is extremely rich and intervening against regime R is unusually cheap. S could both wage war to protect the lives and basic well-being of the million Cs and fulfil its duty to alleviate poverty, thereby protecting the lives and basic wellbeing of millions of people elsewhere in the world. Suppose S chooses to only wage war. The war is not then disproportionate: the war has not prevented S from fulfilling its duty to alleviate poverty. But the war is unnecessary. S could protect the lives and basic wellbeing of a million people by alleviating poverty. Unless and until it has exhausted that option, it does not need to wage a war to achieve this end.

Why do the arguments apply to a different range of cases? Because each argument represents a distinct concern. The concern that the costs of war are excessive is different to the concern that there is some better means to achieve the same end. One concern can arise even when the other does not.

What the two arguments have in common is that they both seek to broaden the scope of the two conditions. The proportionality argument proposes that we sometimes include the opportunity costs of war in a proportionality calculation. The last resort argument proposes that we sometimes adopt wider descriptions of just ends than the standard descriptions usually assumed. By broadening the two conditions, the arguments take us a step towards pacifism, making war much harder to justify. They do not, however, take us all the way. As I have made clear throughout, the arguments only apply under certain conditions. In the absence of these conditions, poverty offers no objection to war.

One last thought is worth considering. Someone might accept that poverty can render a war unjust, but suggest that this concern is better placed outside of just war theory. On this view, we should conceive of just war theory as merely one test a war must pass. The test determines whether a war is, what we might call, “JWT just”. But there can be further tests applying further criteria. A war can be JWT just without being all-things-considered just.

This suggestion is worth considering because it allows us to confront one way the arguments made in this article can be misconceived. These arguments do not assume that poverty must be able to render war unjust and then look for ways to express that thought within just war theory. Instead, the arguments arise out of solutions to more general problems besetting two just war theory conditions. The problem besetting proportionality is the problem of determining which kind of costs count: only direct costs or also (some) opportunity costs. The problem besetting last resort is the problem of choosing among multiple descriptions of the just end for war. The most plausible solutions to these problems have the effect that poverty can render war disproportionate and unnecessary.

The suggestion, then, that we leave just war theory alone and deal with poverty-based objections to war externally will not work. Some of the components of the just war theory machine require adjustment. Once those adjustments are made, just war theory itself generates poverty-based objections to war.