The morality of substitution intervention: The case of Yemen

Throughout the Yemeni Civil War, western states have supplied weapons used in the indiscriminate bombing campaign conducted by the Saudis. In defence of their actions, British politicians have argued that they are exchanging weapons for influence, and using the influence obtained to encourage compliance with humanitarian law. An additional premise in the argument is that Britain is using its influence more benignly than alternative suppliers would use theirs if Britain were not on the scene. The idea is that Britain is substituting itself for other, less scrupulous, interveners. I argue that, regardless of whether British substitution intervention could be justified in this way…, it is not in fact justified, because Britain has not plausibly used its influence to secure an amount of good sufficient to offset the various harms that its actions have created (or to discharge the expanded duties of rescue that greater influence entails). In addition, the article identifies the various forms that substitution intervention can take, and shows how the concept reveals hitherto neglected reasons to both support and oppose intervention in foreign conflicts.

ousted during the Arab Spring, and who many liberal observers were glad to see the back of. Saleh had served continuously since 1978 (first as president of the Yemen Arab Republic and then of Yemen), presiding over an endemically corrupt regime that placed few constraints on the exercise of executive power. Saleh's government exerted a stifling grip on the media, while security forces imprisoned and tortured with impunity.
On the other side of the conflict are the Houthi rebels. The Houthis participated in the 2011 uprising that toppled Saleh's government, and are nominally committed to fighting poverty and corruption. However, Houthi actions derailed UN-sponsored efforts to organize parliamentary elections and a referendum on constitutional reform. Moreover, in the territories that they control, the Houthis maintain a harsh rule, imprisoning, 'disappearing', and torturing their political opponents (Basu, 2019).
Another important party to the conflict is a coalition of states led by Saudi Arabia, which is fighting on behalf of the Hadi government. According to the UN, the vast majority of civilian casualties are attributable to coalition acts (UN Office of the High Commissioner for Human Rights (OHCHR), 2018). The coalition has conducted numerous airstrikes, which have hit residential areas, marketplaces, mosques, schools, and hospitals (Human Rights Watch 2019).
Notoriously, the Saudis have been supplied with arms by a number of western states, including Britain, France, Italy, and Spain. While several western states -including Denmark, Finland, Germany, the Netherlands, and, more recently, the United Stateshave now ceased supplying weapons, others -including Britain and France -have continued. British sales were briefly suspended -after being ruled unlawful by the court of appeal -but have now resumed.
The sale of weapons to Saudi Arabia looks manifestly unjust. After all, the weapons that the Saudis are receiving are being used in acts that clearly violate the rules of jus in bello. Politicians and government lawyers have offered various arguments in support of selling weapons, and most of these are not even superficially plausible. However, when confronted with a practice that seems manifestly unjust, and which we are intuitively inclined to oppose and try to end, we should endeavour to identify the strongest arguments that could be made in its favour. It is important to do this for at least two reasons. First, we want to be confident that the practice in question is in fact unjust. A seemingly unjust practice might in fact be justified, and we do not want to oppose a practice that should be allowed to continue. Second, if a practice is unjust, we can challenge it more effectively if we take seriously the strongest arguments available to its supporters. If we simply ignore or dismiss those arguments, and make no attempt to assess them or to demonstrate their shortcomings, then we neglect an opportunity to convert supporters of the practice to the correct view.
This article develops and evaluates a provocative and potentially powerful argument that British politicians have offered in defence of their arms sales to the Saudis. 1 According to this argument, arming the Saudis can be morally permissible, even laudable, and this judgement can be sustained even if we concede that the war being waged by the Saudis is unjust. This is because selling weapons puts Britain in a position to limit the degree of unjust harm that the Saudis inflict.
Throughout the article, I shall focus on the conduct of Britain, but I do this merely for the sake of expositional ease. The considerations adduced can also be used to assess the behaviour of other western states. Moreover, while there is always an exigent need for clear moral thinking in times of crisis, it is not the mere topicality of the Yemen conflict that singles it out for attention. On the contrary, the conflict has been chosen as a subject of inquiry because it raises distinctive normative issues, issues that invite us to expand our conceptual vocabulary and to reflect on hitherto neglected arguments both for and against outside intervention in military disputes. As we shall see, the particular defence of British actions to be considered prompts us to conceive of Britain's involvement in novel terms, to wit, as a form of what I shall call substitution intervention. Stated briefly, the suggestion is that Britain is mitigating the degree of unjust harm that occurs in Yemen by substituting itself for other, less scrupulous, interveners.
The first section of this article introduces the idea of substitution intervention and contrasts it with its more familiar analogues. The first section also distinguishes among the various forms that substitution intervention can take, and explains how a substitution intervention can be humanitarian in nature even if it contributes to an unjust war. The second section explains how British involvement in Yemen can be framed as a form of substitution intervention. The third and fourth sections assess the particular form of substitution intervention that the British have undertaken, thereby identifying criteria with which to evaluate other instances of substitution intervention in the future. The final section concludes. The conclusion is that British substitution intervention in Yemen is unjust, but that substitution intervention in different contexts need not be impugned by the various considerations that discredit the actions undertaken on this particular occasion.

Traditional intervention versus substitution intervention
Traditionally conceived, international intervention (1) introduces to a conflict or crisis a new element that was previously absent, or (2) supplements or (3) augments elements that were already present. An intervener supplements by adding more of a particular kind of element (such as military force), while an intervener augments by enhancing the quality of an element. These three forms of traditional intervention are often pursued simultaneously.
By contrast to traditional intervention, substitution intervention (4) substitutes or exchanges one (set of) element(s) for another. For example, a state attempts a substitution intervention if it provides rebels with weapons, and intends those weapons to be used instead of (and not in addition to) the weapons that the rebels already possess. The substitution intervention is successful if the weapons are in fact used in the intended manner.
An intervention can have a substitutive effect without being motivated by a substitutive rationale. For example, troops that were intended to be supplementary may end up substituting for (a subset of) the fighters who were already involved in hostilities. Conversely, an intervention can be motivated by a substitutive rationale but not achieve a substitutive effect. When this happens, we have a failed substitution intervention.
Different instances of substitution intervention (henceforth, SI) can vary in six general respects. First, different instances of SI vary according to the nature of the elements being exchanged. These elements might be weapons, or troops, or something else. Another exchangeable element, and one that will have particular relevance in subsequent sections, is influence; an intervening state can seek to substitute its own influence for that of another intervening state. One fact that complicates matters is that a state might substitute one element as a means to substitute another element. As we shall see, in the case of British arms sales to Saudi Arabia, the British government can be understood to be substituting weapons as a means to substituting influence.
When an intervener substitutes its own weapons, troops, and/or influence for analogues that would otherwise have been provided by a different intervener, it can be appropriate to say that the elements being exchanged are the interveners themselves. (This should become clearer in the next section.) Second, the elements that an SI replaces may have originated within the relevant territory, or they may have been provided by another intervener. We can refer to an SI that replaces elements that originated within the relevant territory as an internal SI, and to an SI that replaces elements provided by another intervener as an external SI.
Third, SI can be either synchronic or subjunctive. Synchronic SI exchanges a new element for one that is already present. For example, a state provides weapons to replace those currently in use. By contrast, subjunctive SI exchanges one element for another that it is reasonable to believe would otherwise have been introduced. For example, a state sends troops believing that, if it did not, a different state would send its troops.
Fourth, SI can be complete or partial. SI is complete if an intervening state, X, undertakes all of the functions that might otherwise have been performed by another actor, Y, with the result that Y performs none of these functions. By contrast, SI is partial if X undertakes some subset of the functions that would otherwise have been performed exclusively by Y, with the result that Y also performs some subset of these functions, rather than performing them exclusively. 2,3 Fifth, like more traditional forms of intervention, SI can be self-interested or humanitarian. Self-interested SI is pursued exclusively for the role it plays in promoting the interests of the intervener. By contrast, humanitarian SI is motivated (at least in part) by a desire to protect the interests of others, and actually achieves some humanitarian end (such as the protection of civilians).
Finally, SI can contribute to wars and acts of war that either satisfy or fail to satisfy the conditions of jus ad bellum and jus in bello. This sixth aspect does not map neatly onto the fifth. An SI can be humanitarian even if it contributes to wars or acts of war that are unjust, as I shall now explain.
One way in which a humanitarian SI might contribute to a war that is both initiated and fought unjustly is by enabling combatants to inflict less (or less severe) collateral damage than they otherwise would have done. To spell out this idea in more detail, consider the fact that irregular forces often make use of improvised weaponry. Improvised weapons are often less accurate, and therefore more productive of collateral harm, than their more sophisticated counterparts. For example, Syrian militants fighting to topple the government of Bashar al-Assad have used makeshift catapults to hurl explosive projectiles at government forces. Such weapons are considerably less accurate than modern artillery. Since combatants often use less accurate makeshift weapons simply because they are unable to get their hands on more conventional items, and not because the former are thought to possess valuable attributes absent from the latter, there can be a reason to support arms transfers that would substitute more accurate weapons for less accurate improvised alternatives, even if these transfers are made to combatants engaged in a war that fails to satisfy the rules of jus ad bellum and/or jus in bello. The reason is that such transfers could facilitate a reduction in the degree of harm unintentionally inflicted upon non-combatants. 4 Now, many contemporary just war theorists argue that harms to certain combatants -namely, those combatants who are fighting a just war -can affect a war's morality in the same kind of ways that harms to civilians can affect a war's morality. We can accept this claim and still believe that a war's badness can be reduced by the kind of substitution intervention envisioned in the previous paragraph. This is for two reasons. First, harms inflicted on civilians are often thought to be worse than harms inflicted on just combatants (Lazar, 2015). Second, some conflicts are fought exclusively by unjust combatants. 5 Of course, there will often be both moral and prudential reasons not to provide unjust combatants with more accurate weaponry. This is because while more accurate weaponry may reduce the degree of collateral harm that unjust combatants inflict, it is likely also to bolster the efficiency with which these combatants pursue their ends, and we will often have good reason to frustrate, rather than facilitate, the realization of those ends. However, these considerations will not necessarily be decisive in all contexts.
Here is a different way in which a humanitarian SI might contribute to a war that is both initiated and fought unjustly. As will be discussed in detail in subsequent sections, a state might intervene in an unjust war by substituting its influence for that of another. It may then use its influence to alter the way in which the war is fought. For example, it might pressure military leaders to refrain from deliberately targeting civilians, and to focus their attacks on combatants.
This second way in which a humanitarian SI might contribute to a war that is both initiated and fought unjustly will be scrutinized in the remainder of this article. As we shall see, a potentially powerful argument for British intervention in Yemen suggests that the British are using their influence to reduce the badness of the Saudis' unjust conduct.

The mitigation argument
Having introduced the idea of SI, let us now focus on the example of the Yemen war. As we shall see, there is a potentially powerful argument in favour of British conduct in this war that invites us to view Britain's contribution as a form of SI.
I said in the introduction that the Yemen war involves no obvious 'good guys'. Before proceeding, we should get clearer about the moral status of the wars being waged by each belligerent. While the Yemeni government is led by a man who played a prominent role in years of misrule, it is plausible to hold that the government is fighting with a just cause. Prior to the outbreak of hostilities, Yemen was in the process of an important transition, a transition which could potentially have put the country on a road to a brighter, more democratic, future. The government and its supporters were guiding Yemen through this process, and, in doing so, did nothing to make themselves liable to violent attack. The Houthi rebellion, which upended the government's proposed political and legislative reforms, was initiated unjustly. Repelling Houthi aggression and restoring the status quo ante plausibly constitutes a just cause for war. This is not to say that the Saudi intervention on behalf of the transitional government was initiated justly. The Saudi resort to war may in fact have been unjust, in virtue of failing the ad bellum principle of necessity. Plausibly, there were reasonable alternatives available to the Saudis, which were unjustly neglected in favour of war. However, it does not follow from this that an SI which contributes to the Saudi campaign should be opposed. As I suggested in the previous section, SI can be humanitarian even if it contributes to a war that was initiated unjustly.
But let us not get ahead of ourselves. We first need to acquire a clearer picture of the argument for British intervention that is the primary subject of this article. We also need to comprehend how this argument frames British intervention as a form of SI. The argument runs as follows: Britain is providing weapons as a means of acquiring influence, and influence in turn enables Britain to ensure that the Saudis fight less unjustly than would otherwise be the case. Because it provides weapons, Britain is able to station personnel with the Saudi military. British personnel can then supervise Saudi operations and encourage them to refrain from deliberately attacking civilians. In the words of government officials, British personnel are able to 'lobby' the Saudis 'on priority issues' (Crompton, 2016: 17) and to 'help support Saudi compliance with International Humanitarian Law' (Lancaster, 2019). We can call this the mitigation argument, because it claims that, by supporting an unjustly fought (and perhaps unjustly initiated) campaign, an intervener can mitigate the degree of unjust harm that combatants are inflicting. (These considerations do not yet reveal why Britain's involvement can be understood as a form of SI. I will get to that.) A version of the mitigation argument has been advanced by Jeremy Hunt, the former British foreign secretary. Hunt (2019) wrote, Our strategic relationship with Saudi Arabia . . . allows us the opportunity to influence [its] leaders . . . We could, of course, decide to condemn them instead. We could halt our military exports and sever the ties that British governments of all parties have carefully preserved for decades, as critics are urging.
But in doing so we would also surrender our influence and make ourselves irrelevant to the course of events in Yemen. Our policy would be simply to leave the parties to fight it out, while denouncing them impotently from the sidelines.
That would be morally bankrupt and the people of Yemen would be the biggest losers.
Throughout this discussion, I am going to assume that, if British arms sales to the Saudis can be permissible, they can be permissible only if they really do have a good chance of mitigating the harms that the Saudis inflict. More specifically, I am going to assume that British arms sales could not be permissible simply in virtue of not exacerbating the harms that the Saudis inflict. A common defence of arms transfers to aggressive and oppressive states claims that they leave the victims of those states no worse off than they would otherwise have been. This argument merits independent treatment (see Christensen, 2021), and, for present purposes, I am setting it aside.
Let us return to the mitigation argument. This argument draws a counterfactual comparison. It says that, if Britain provides weapons, less unjust harm occurs than would otherwise occur in the most relevant possible world in which Britain does not provide weapons. In order to assess this claim, we have to identify the correct description of the most relevant possible world in which Britain does not provide weapons.
Here are three candidates for this description. First, Britain does not provide weapons and, therefore, does not acquire any influence. However, Britain is substituted by a Backup Supplier, that is, a state that provides weapons comparable to those that could have been provided by Britain. The Backup Supplier acquires influence over the Saudis that is comparable in kind and degree to the influence that could have been acquired by Britain.
Second, Britain does not provide weapons, nor is it substituted by a Backup Supplier. However, this does not prevent the Saudis from fighting. The Saudis continue to fight, and because no outsider is in a position to exert much influence, the Saudis are free to fight in a manner of their choosing.
Third, Britain does not provide weapons, nor is it substituted by a Backup Supplier. Moreover, this does prevent the Saudis from fighting, or at least from fighting as brutally as they had done previously. Some opponents of British arms sales seem to imagine that the third description provides the best account of the most relevant possible world in which Britain does not provide weapons. That is, they seem to believe that if Britain did not provide weapons, the Saudis would be unable to keep fighting, or to keep fighting as brutally as they have been. From this, opponents of arms sales infer that the mitigation argument fails. They make this inference because they assume that in a world where the Saudis' fighting capacity is significantly diminished, there would be fewer (or less severe) unjust harms than in a world where Britain provides weapons and thereby maintains the Saudis' fighting capacity at its present level.
By contrast, some supporters of British arms sales seem to imagine that the second description provides the best account of the most relevant possible world in which Britain does not provide weapons. They believe that if Britain did not provide weapons, it would not be substituted by a Backup Supplier, but the Saudis would nevertheless be able to continue fighting. Moreover, the Saudis would fight in a manner of their own choosing, unconstrained by external influence.
From their belief that the second description provides the best account of the most relevant possible world in which Britain does not provide weapons, supporters of arms sales infer that the mitigation argument succeeds. They make this inference because they assume that in a world where the Saudis fight in a manner of their own choosing, there would be more (or more severe) unjust harm than in a world where Britain provides weapons and thereby acquires influence.
However, while both the second and third descriptions seem to enjoy some degree of endorsement, it is plausible to maintain that it is in fact the first description that provides the best account of the most relevant possible world in which Britain does not provide weapons. That is, it seems plausible to maintain that if Britain did not provide weapons, it would most likely be replaced by a Backup Supplier, which would provide weaponsand thereby acquire (further) influence -instead of Britain. Indeed, this is the view apparently endorsed by Michael Fallon, who, while acting as the British defence secretary, notoriously advised his fellow lawmakers to cease criticizing Saudi Arabia, on the grounds that criticism would jeopardize efforts to sell the Saudis more fighter planes (MacAskill and Elgot, 2017). Presumably, Fallon reasoned that if British politicians offended the Saudis, the Saudis would take their business elsewhere.
Suppose that the first description does in fact provide the best account of the most relevant possible world in which Britain does not provide weapons. If this is the case, a proponent of the mitigation argument must defend the following comparative claim: if Britain provides weapons, less unjust harm occurs than would otherwise occur in a world where the Backup Supplier provides weapons instead. This is because the Backup Supplier would place fewer (or less severe) constraints than Britain does on Saudi conduct. We can call any argument that defends an intervention on the grounds that it mitigates unjust harm by substituting one intervener for another a substitution argument. It is the particular substitution argument just described (which I will refer to as the substitution argument, or SA) that frames British involvement in Yemen as a form of SI. More specifically, this argument casts British involvement as a form of SI that is external, subjunctive, partial, and humanitarian, despite contributing to a war that is being unjustly fought and was arguably unjustly initiated.
Substitution arguments look plausible in certain contexts. Consider the following case. (Note that, in this case, it is only weapons that are exchanged, not weapons plus influence.) Terrorist A terrorist is running toward the centre of a crowded marketplace, where he intends to blow himself up. Neither Anne nor Ben is in a position to shoot the terrorist, but they can throw their guns to Chris, who is better placed. Because the guns are inaccurate, and the marketplace very crowded, shooting the terrorist would be a disproportionate act. The collateral harm that it would likely cause is excessive in relation to the good that it would likely achieve. Anne's gun is more accurate, and, if used, would cause less collateral harm than Ben's (though using it would still be disproportionate). Anne has good reason to think (i) that if she does not throw her gun to Chris, Ben will throw his, (ii) that if she does throw her gun, Ben will not throw his, (iii) that Chris will use whichever gun he receives to shoot the terrorist, and (iv) throwing her gun to Chris will be less harmful in the long run than shooting Chris or Ben. Anne throws her gun to Chris.
It seems intuitively plausible that Anne acts permissibly. While Anne contributes to the infliction of unjust harm, her contribution has the expected effect of reducing the degree of unjust harm that is inflicted, and it is because of this expected effect that Anne acts as she does. Anne seems to have a lesser evil justification for throwing her gun to Chris. Moreover, the magnitude of the evil committed is reduced by the fact that all of the innocent people harmed by Anne's act are expected beneficiaries of her act. Consider, by way of comparison, the classic trolley case, where a bystander saves five innocent people by killing one innocent person (Thomson, 1985). In the trolley case, the innocent person who is killed in the process of saving the five is in no danger until the bystander acts. By contrast, those harmed when Chris shoots at the terrorist with Anne's gun would also have been at risk had Chris used Ben's gun. To be sure, Anne's act, like the conduct of the bystander in the trolley case, may have the effect of transferring harm from one set of individuals to another. Perhaps the exact time at which Chris opens fire is affected by whose gun he receives, and perhaps this affects who is caught in the crossfire. Nevertheless, by acting to reduce the total number of innocent victims, Anne thereby reduces the probability of any one innocent person being harmed.
Anne's conduct can seemingly be justified by a substitution argument. Moreover, since Anne's conduct appears to resemble Britain's in several important respects, it seems that a defence of British intervention built around a substitution argument cannot be easily dismissed. 6 However, Britain's intervention in Yemen is marked by a variety of complexities absent from the simple hypothetical case above. To assess British intervention in Yemen, we shall have to take these complexities into account.

The substitution argument
For the SA to vindicate British involvement in Yemen, it must be reasonable to believe that the Backup Supplier really would impose fewer constraints on Saudi conduct than Britain does. To establish whether this belief is reasonable, we would first have to identify which state is likely to fill the role of Backup Supplier, and then establish how that state is likely to act in that role. With regard to the likely identity of the Backup Supplier, we are far from being completely in the dark. On the contrary, it would be reasonably easy to draw up a shortlist of candidates. We know who the largest arms exporters are; we know who manufactures the relevant weapons; we know who has armed Saudi Arabia in the past; and we know who has declined to arm Saudi Arabia under present circumstances. Nor are we completely in the dark with regard to how any potential Backup Supplier is likely to act. We can study how potential Backup Suppliers have acted in the past in other contexts, and, given that several potential Backup Suppliers are already currently arming the Saudis in some capacity, we can examine their conduct in the present.
Nevertheless, there are undoubtedly unknowns. We cannot know with confidence who would substitute for Britain if Britain refused to provide arms, nor can we know with confidence how the actual Backup Supplier would behave. This uncertainty might be regarded as a virtue by British politicians keen to exploit the SA for their own ends. These politicians might maintain that, since we do not know the identity of the Backup Supplier, nor exactly how the Backup Supplier would behave, their own conduct should be given the benefit of the doubt.
But this hardly seems plausible. By arming the Saudis under current conditions, Britain is doing something that is presumptively wrongful. It is providing weapons to a vicious regime that is using those weapons to fight in an unjust manner. To vindicate this act, the presumption of wrongdoing must be overturned. Moreover, while Britain might be imposing some constraints that the Saudis would not have adopted unilaterally, it is also blocking the interventive acts of others. If, by substituting itself for the Backup Supplier, Britain thereby prevents a more benign exporter from imposing greater constraints on Saudi conduct, Britain has made matters worse. It has ensured that more injustice occurs than would have occurred had Britain not provided weapons. The onus is thus on British officials to demonstrate that they are not making matters worse, or at least that they have not acted recklessly in running an unduly large risk of making matters worse. Importantly, if the identity of the Backup Supplier is unknown, or we cannot predict with confidence how the Backup Supplier would behave, Britain might be morally required to do more good than it would be required to do if these epistemic limitations were absent.
To illustrate, suppose that we could predict with confidence that France would be the Backup Supplier, and that France would impose extremely minimal constraints on Saudi conduct. In order to demonstrate that it has improved the situation, Britain might have only to impose constraints that are marginally stricter than the extremely minimal constraints that would otherwise be imposed by France. But now suppose that we cannot predict with confidence that France would be the Backup Supplier. Suppose there is a non-negligible possibility that Australia might be the Backup Supplier. Moreover, there is a non-negligible possibility that Australia might impose moderately strict constraints on Saudi conduct. To demonstrate that it has improved the situation, Britain would have to impose constraints that are greater in severity than those that would otherwise have been imposed by Australia. Perhaps it is unreasonable to expect Britain to prove that it has not made matters worse, but at the very least, and as suggested earlier, it is required to demonstrate that it has not run an unduly large risk of making matters worse.
When attempting to establish whether Britain has improved matters or made them worse, we cannot focus exclusively on the comparison between the way Britain uses its influence and how the Backup Supplier would use its influence. We must also identify any costs that British intervention entails, and which would have to be deducted from any good that Britain does. In this context, the following four considerations are especially pertinent: 1. First, it is important to note that if the Saudis have chosen to purchase certain weapons from Britain, rather tha from a rival supplier, this might be because British weapons have some virtue that weapons from rival suppliers' lack. One virtue that could potentially distinguish British weapons is price. Britain might be providing weapons more cheaply than its rivals. (This might be true with regard to some weapons, and in relation to some rivals, and not others.) If Britain is providing weapons more cheaply, this increases the risk that, by substituting itself for the Backup Supplier, Britain is making matters worse. This is because if Britain is providing weapons more cheaply, the Saudis might be able to purchase more (or more destructive) weapons than would otherwise have been possible, and thereby inflict more unjust harm. 2. Second, it is important to consider the opportunity costs of Britain's strategy. If Britain terminated arms sales to the Saudis, it could then encourage other states to follow suit. Moreover, Britain would not have to act in isolation. Rather, it could act in concert with Germany, Denmark, and others, and thereby exert greater pressure on those states that are continuing to provide weapons.
In response, it might be said that while some states could perhaps be persuaded to terminate their arms sales, others will not buckle to external pressure. Prior to the inauguration of Joe Biden, this could confidently have been claimed of the United States. When Congress attempted to block arms sales to the Saudis, President Trump simply vetoed its resolutions. Moreover, Biden's subsequent suspension of arms sales notwithstanding, it would be a mistake to regard the status quo ante as an anomalous state of affairs attributable to the idiosyncrasies of the Trump Administration. Notice that the Obama government oversaw more arms sales than any other since the Second World War, and US$94 billion of these sales went to the Saudis (Farid, 2017).
However, even if we concede that some states will stubbornly resist calls to terminate sales, and that there will inevitably be at least one state willing to play the role of Backup Supplier, the ability to influence some potential suppliers should not be dismissed as unimportant. Reducing the number of potential Backup Suppliers can be a positive contribution. This is because reducing the number of states competing for Saudi business is likely to drive up prices. And, as we have already noted, higher prices might lead to the Saudis purchasing fewer (or less destructive) weapons.
A further point to make in this context is that it is misleading to think of each arms exporter as engaging in discrete activities unrelated to those of its counterparts. The sales of one exporter often complement the sales of another. To illustrate, among Britain's exports to the Saudis is the Eurofighter Typhoon, a jet manufactured by a consortium of British, German, Italian, and Spanish companies. When Germany announced that it would sell no more weapons to the Saudis, British politicians complained that the moratorium would prevent BAE Systems (the British partner in the Eurofighter consortium) from fulfilling its contracts with Riyadh (Wintour, 2019). The takeaway point here is that, by terminating their arms exports, states can often do good by hobbling the efforts of other governments that wish to continue making sales. By refusing to terminate its exports, Britain has potentially squandered opportunities to promote morally desirable states of affairs.
It is worth pausing to reflect on this point. Globalization is typically associated with a neutering of the state. When commerce went global, capital gained the agility with which to outmanoeuvre the political forces that we had traditionally relied upon to keep it in check. But the foregoing observations reveal that the internationalization of production also renders the operations of certain arms firms -already inordinately dependent upon their respective governments -vulnerable. Policies introduced by one state can have dramatic knock-on effects that impact upon firms beyond their jurisdiction. In this respect, globalization actually seems to magnify the state's power. With the will to do so, governments can take on the more malign manifestations of the international political economy -including those that they are often expected to sponsor -and topple them like dominoes. Governments with genuinely humanitarian ambitions should be on the lookout for opportunities to do just that.
3. Third, we must take a broad view of the 'matters' that might be made worse by British intervention. In particular, we should not limit our attention to Saudi conduct in the Yemen war, for the effects of British intervention might extend beyond this. Of relevance here is the fact that British arms sales to Saudi Arabia appear to contravene Britain's obligations under the Arms Trade Treaty (ATT), which entered into force in 2014. British arms sales are, therefore, most likely illegal, and are challenging a fledgling legal framework that aims to curb the worst excesses of the existing arms regime. While imperfect, the ATT is a first step towards demolishing an unjust system responsible for massive quantities of unjust harm. This system is currently protected by a number of powerful arms exporters -most notably China, Russia, Israel, and the United States -who have either refused to sign the ATT, or, in the case of the United States, withdrawn their signature. Britain has both signed and ratified the ATT, but by continuing to arm terrorist states like Saudi Arabia, Britain seriously jeopardizes the emergence of a new, less harmful, arms regime. 4. Fourth, and in keeping with my suggestion that we take a broad view of the matters that British intervention might make worse, we should stress the obvious fact that the Saudis' unjust behaviour in Yemen is in no way anomalous. Rather, it is consistent with the Kingdom's general status as an internally oppressive and outwardly aggressive regime. Notice that when Germany and Denmark announced that they were halting arms sales to the Saudis, they did so not directly in response to Saudi bombing in Yemen, but in response to the brutal slaying of Jamal Khashoggi, the dissident journalist who, in October 2018, was assassinated at the Saudi consulate in Istanbul. The German and Danish governments terminated arms sales, it seems, in order to express, in the clearest possible terms, their contempt for this vicious act, and for the Saudis' ongoing disregard for the most basic legal and moral norms. By contrast, the British decision to continue arms sales expresses something very different, namely, a willingness to pursue businessas-usual with a ruthless regime disdainful of ethical imperatives. It communicates to the Saudis (and to the other outlaw states watching) that they may act as they please, and expect little in the way of material costs (cf. Christensen, 2019: 125). For the sake of brevity, we can say that British conduct emboldens the Saudis.
In response, it might be said that Britain's emboldening act simply substitutes for a comparable emboldening act that would otherwise have been performed by the Backup Supplier. Therefore, when calculating how much good Britain achieves, the cost of Britain's emboldening does not need to be subtracted.
However, this response overlooks two important considerations. First, it overlooks the fact that Britain's SI in Yemen is partial, not complete. The Backup Supplier is almost certainly one of the states that is already arming the Saudis in addition to Britain. Britain is substituting for some of the Backup Supplier's sales, but not for all of them. Because any sales constitute an emboldening act, the particular emboldening act performed by Britain does not substitute, but rather supplements, the emboldening act of the Backup Supplier.
Second, sales are unnecessary to embolden. Simply offering weapons in the market is sufficient for emboldening. Thus, even if Britain's intervention in Yemen were complete rather than partial, Britain's emboldening act would still be supplementary rather than substitutive, because the Backup Supplier would be competing for the Saudis' business in the marketplace.
An important point that emerges from the foregoing observations is that, in order for British SI to improve matters all-things-considered, Britain must use its influence to do an amount of good that is sufficiently large not just to exceed any good that the Backup Supplier would use its influence to produce, but also to offset any bad its intervention engenders. We can call the amount of good that is sufficiently large to exceed any good that the Backup Supplier would create and to offset any bad that the intervener creates the sufficient good.
It remains an open question how much good the British must do for their intervention to be justified (if it can be justified). Are the British merely required to do the sufficient good, or are they required to do more?
Of relevance here is the fact that interveners are burdened by a duty of rescue. If they (or, rather, their citizens, acting collectively through the apparatus of the state) have the capacity to rescue someone from a threat of severe harm at reasonable cost to themselves, they are morally required to do so. Now, the SA suggests that, by selling weapons and acquiring influence, the British enhance their capacity to rescue. It follows from the nature of the duty to rescue that the British must exercise this capacity up to the point at which the cost of doing so becomes unreasonable. Call this point the cost threshold. 7 If, in doing the sufficient good, Britain incurs a cost that happens to match the cost threshold, then it is plausible to hold that the British are required to do no more than the sufficient good. However, if the cost threshold lies beyond the cost that Britain incurs while achieving the sufficient good, then Britain must continue exercising its influence to achieve more than the sufficient good. Britain must continue to act until it reaches the cost threshold. This is not because the British, in their capacity as substitution interveners, acquire special duties that others lack. It is simply because the implications of the general duty to rescue change as one's capacity to rescue shrinks or grows.

The verdict
Does the substitution argument vindicate British intervention in Yemen? I think it does not. As we have seen, British arms sales to the Saudis have several significant costs. First, they threaten reformative efforts embodied in the Arms Trade Treaty, and thereby contribute to the maintenance of a deeply unjust arms regime. Second, they communicate the message that ruthless regimes may stray as far as they please from the imperatives of morality and law while continuing to enjoy opportunities for valuable international cooperation. Moreover, if Britain is selling weapons more cheaply than its competitors, it is most likely enabling the infliction of greater harm than would otherwise be possible. In these ways, Britain has created a serious 'good deficit', which needs to be paid-off before intervention could make any positive difference.
In addition, Britain's strategy has important opportunity costs. In providing arms, Britain foregoes an opportunity to persuade other states to terminate their exports. As we have seen, persuading some states to cease competing for Saudi business is valuable even if other states will inevitably continue to trade. Furthermore, by providing arms, Britain may be supplying components that are used in conjunction with weapons sold by other exporters, thereby squandering an opportunity to hobble a particularly toxic market.
Meanwhile, the benefits of British intervention are elusive at best. The Saudis have continued to wage a vicious war, creating the impression that the influence claimed by British politicians is either illusory or has otherwise been squandered. If Britain really has acquired meaningful influence over the Saudis, and if, contrary to appearances, its intervention really has somehow produced enough good to ensure that Britain's moral balance is no longer in the red, it is highly unlikely that Britain has fully exercised the enhanced capacities for rescue that real influence would bring, and thereby discharged the duties that such capacities entail.
I have not argued that British intervention in Yemen is unjustified all-things-considered. I have accepted that the transitional Yemeni government has a just cause in repelling Houthi aggression and in endeavouring to restore the status quo ante, and it might be said that the Saudi intervention can play an important role in achieving these aims (at least if it receives the kind of support that Britain and others are providing). I do not find this argument persuasive, but I have not sought to show that it fails. Relatedly, it is often said that the Saudi intervention plays an important role in checking the expansion of Iranian influence in the region. Again, I do not find this argument compelling, but I have not addressed it in this article.
Instead, I have focused my attention on one specific defence of Britain's conduct, a defence that seeks to appropriate the humanitarian concerns of those critics moved by the shocking depravity of Saudi airstrikes on civilian sites. According to this defence, British arms sales are justified by the opportunities they create to benignly influence Saudi behaviour. I have argued that, regardless of whether British intervention could be justified in this way, it is not in fact justified, because the British have not plausibly made adequate use of any influence that they may have acquired.
More positively, I have identified a previously overlooked type of intervention, as well as the various forms that this type of intervention might assume. In doing so, I have shed light on previously neglected reasons for intervening in foreign conflicts. While I have criticized British substitution intervention in Yemen, I have not claimed that substitution intervention is inevitably wrongful. When the precise constellation of circumstances surrounding an intervention differ from those that provide the backdrop to British involvement in Yemen, or when interveners use their influence much more effectively than the British have used theirs, our assessment of the intervention could potentially be quite different. In any case, the arguments developed earlier have hopefully ensured that we are now better equipped both to understand substitution intervention, and to morally evaluate such intervention when it occurs.