Thirty years after the devastating famine that claimed the lives of over a million people, some have described Ethiopia’s transformation as ‘an economic miracle’.Footnote 1 Over the past decade, the country has seen impressive economic growth, a significant decrease in poverty, infrastructure development, and dramatic improvements in the health and education of the population.Footnote 2 Many attribute this transformation to a combination of the Ethiopian Government’s commitment to promoting development and the foreign aid provided by donor governments.Footnote 3
Nevertheless, human rights organizations, such as Human Rights Watch, have criticized donor governments for turning a blind eye to rights abuses committed by the Ethiopian Government, fuelling accusations of donors’ complicity in civil and political rights violations.Footnote 4 In 2012, lawyers representing an Ethiopian farmer named in court papers as ‘Mr O’ announced they would begin proceedings against the UK Government. The UK’s Department for International Development (DFID) funded an Ethiopian Government programme that, they claimed, had led to severe human rights abuses. Mr O stated that the Ethiopian Government’s ‘villagization’ programme led to the forced resettlement of thousands of families, including his own, to places that lacked basic amenities. He also claimed to have witnessed the beating and rape of villagers by soldiers.Footnote 5
The case demonstrates the tension between the benefits of development aid and concerns over rights violations committed by aid recipients, which we will hereafter refer to as ‘the donor’s dilemma’. How should donors decide whether to continue providing aid to such states? If donors do continue to provide aid to these governments, what other measures can be taken to ensure they are not complicit in rights violations?
In this article, we construct a framework on how donors can address this problem in a normatively justified way. We do this by developing a distinctively political theory of dilemmas – drawing on recent methodological advancements in normative political theory – which identifies three paradigmatic models: complicity, double effect and dirty hands. We consider this framework in the context of development aid to states accused of rights violations, and demonstrate how donors can identify the different types of dilemmas and respond accordingly. The article makes several contributions. First, it contributes to normative political theory by developing a distinctively political understanding of dilemmas, and uses the context of development aid to highlight the different features of the framework. Secondly, it informs the polemic debate in development policy and research on whether donors should provide aid to states involved in rights violations by demonstrating how a politically realistic approach to aid can address normative concerns related to civil and political rights. In doing so, the article also contributes to broader debates in international relations on foreign aid and human rights.
The article is structured as follows. First, we discuss the specific context of international aid to regimes that violate rights, which gives rise to the donor’s dilemma. We focus on official development assistance provided by bilateral donors. Secondly, we develop a general framework for analysing political dilemmas, in which we specify an analytical distinction between different models of dilemmas and consider the normative and political significance of each. Thirdly, we examine the general framework in the context of development aid and flesh out the relevant considerations in different cases, drawing on empirical examples. This approach, we argue, demonstrates that a political approach to aid, which is sensitive to political contexts and structural constraints, need not be normatively silent.
AID, POLITICS AND DEVELOPMENT
In the past two decades, there has been much debate about the effectiveness of aid in promoting development. The growing consensus is that the question of whether aid ‘works’ is misguided, as the answer depends on the type of aid and the outcome under consideration.Footnote 6 We do not focus on aid effectiveness here. Instead, we proceed on the basis of two assumptions: first, aid has some positive impact on socio-economic outcomes, such as income and health; and secondly, aid is most effective when delivered through national governments and institutions in developing countries, as is widely recognized by international development policy makers.Footnote 7 The dilemma arises because aid has additional and potentially undesired political effects, which have implications for the relationship between donor and recipient governments.
We focus on two political effects. First, aid provides donors with influence over recipient governments, which can reduce the accountability of recipient governments to their citizens. There has been considerable attention in international relations research on the extent to which donor governments should (and do) use this political influence to further their own interests in global politics.Footnote 8 Donors have also sought to use this influence to achieve more development-oriented goals by attaching conditions to the provision of aid based on recipient governments implementing economic and political reform.Footnote 9 Critics of aid, however, have long argued that providing donors with this influence reduces democratic accountability because it makes governments less dependent on their citizens for tax revenue, leading to lower accountability to citizens.Footnote 10
A second political effect of aid is strengthening the government’s position relative to other actors in a society, such as opposition political parties and civil society. Aid can provide a government access to greater material resources and often bolsters the legitimacy of a regime.Footnote 11 While strengthening the position of the government may be a positive outcome in some contexts, critics of foreign aid argue that providing aid to more autocratic regimes impedes democratization and the protection of human rights. This is because aid increases these governments’ ‘resources, patronage, and power in relation to the rest of society’, and thereby ‘enhances the hold of governments over their subjects’.Footnote 12
Therefore the donor’s dilemma arises because, in addition to helping improve socio-economic outcomes, aid can strengthen the position of recipient governments, including those that violate citizens’ political and civil rights, which may undermine the positive effects of aid. Since aid gives donors influence over recipient governments, some view donors as complicit in rights violations committed by these governments. This dilemma is further complicated in the context of developmental regimes – governments of poorer nations that actively promote socio-economic development through close involvement in the market – and have achieved some success in doing so.Footnote 13 Most developmental states are non-democratic and have frequently been implicated in the violation of political and civil rights. Indeed, in his analysis of developmental states, Leftwich argues ‘the combinations of their sometimes brutal suppression of civil rights, their apparently wide measure of legitimacy, and their generally sustained performance in delivering developmental goods […] are intimately connected’.Footnote 14
The success of developmental states has led many to question the focus on democratic governance and human rights in development assistance – particularly as donor efforts to promote more accountable governance are widely seen as having produced disappointing results.Footnote 15 This has prompted the recent ‘political turn’ in development research and policy.Footnote 16 This political turn – which the donor community has labelled ‘thinking and working politically’ (TWP) – emphasizes the need for donors to better engage with the political realities of local contexts,Footnote 17 and to accept that there are ‘tensions, conflicts [and] difficult trade-offs over time between the various goals of development – such as growth, democracy, stability, equity and autonomy’.Footnote 18
The emphasis on political realism, the acceptance of trade-offs and the importance placed on domestic leadership has meant that proponents of TWP tend to take a favourable view of developmental regimes, even if they exhibit authoritarian characteristics or are involved in rights abuses.Footnote 19 Indeed, proponents of TWP have criticized the donor emphasis on democratic governance and rights on the basis that ‘what poor developing countries really need are leaders who… can get things done’.Footnote 20 However, many have expressed concern about the TWP’s ‘top-down’ approach and the danger that ‘issues of power, rights and justice too easily get lost’ and ‘we end up helping governments that routinely kill or suppress their opponents [to] “deliver development”’.Footnote 21 Indeed, there has been growing criticism of donor governments’ willingness to overlook rights abuses in developing countries, as the Ethiopia example demonstrates.
As such, this dilemma has led to a polemic debate in development research and policy, which we look to move beyond. Effective development policy, we argue, requires recognizing the complex political contexts of, and structural constraints to, development. This does not mean, however, that normative challenges related to the donor’s dilemma ought to be sidelined. Since the various effects of aid are closely connected, the precise nature of these effects, and the normative significance of the dilemmas they generate, need to be better understood. We turn to the broader question of the nature of political dilemmas to explore the analytical complexity of the donor’s dilemma.
A POLITICAL THEORY OF DILEMMAS
Drawing on recent methodological developments in normative political theory, we develop a novel conceptual framework for analysing political dilemmas, which challenges the simplistic dichotomy characterizing the current debate between idealistic and abstract ‘normative theory’ on the one hand, and the kind of gritty approach to politics offered by TWP, on the other. In brief, we maintain that by analysing the predicament that donors face as a political dilemma, one can retain a normative perspective while being sensitive to the realities of politics. We suggest three possible models of political dilemmas. Rather than dissolving the donor’s dilemma by philosophical fiat, these models acknowledge its political significance and offer a perspective for practically engaging with it. With the conceptual distinction between different dilemmas in place, in the next section we consider this general conceptual framework in the context of the donor’s dilemma.
What is a dilemma? In moral philosophy, a dilemma arises when an agent is required to choose between two (or more) incompatible actions, and each choice she makes involves a moral wrong – either because she will act in a way that is morally prohibited, or because she will fail to act in a way that is morally obligatory. Subjectively, some cases of moral conflict may appear insoluble to the agent; importantly, however, moral dilemmas are an ontological, rather than an epistemic, concept. The debate in moral philosophy centres on a disagreement over whether ‘genuine’ moral dilemmas can exist; some philosophers deny that they can.Footnote 22 If a choice must be made between two incompatible actions, this is seen as equivalent to self-contradiction: one of the supposed moral requirements must be false.Footnote 23
Within moral philosophy, this extreme position has been heavily criticized for failing to acknowledge the importance of dilemmas. First, moral dilemmas explain the sense of remorse, guilt and distress that accompany choosing one action over another, without describing these attitudes as irrational. Secondly, moral dilemmas allow for the existence of what Williams calls ‘moral remainders’, meaning that we can acknowledge that in choosing one horn of the dilemma, the force of the other horn does not simply disappear. Even if, all things considered, the choice was justified, the unmet moral requirement generates further duties, for example of apology or reparation.Footnote 24 Thirdly, it may be the case that conceptually, in a morally perfect world, where all agents act as they should, moral dilemmas would not exist. Yet our world is not morally perfect in this sense; insofar as the role of moral philosophy is to provide guidance to actual moral agents, it must acknowledge the reality of moral dilemmas.Footnote 25 By acknowledging this reality, we are able to recognize two imperatives when facing a moral dilemma: choosing the lesser of two evils, and acting to avoid, as much as possible, additional moral dilemmas in the long term.
To be useful for a normative analysis of politics, however, the literature on moral dilemmas must be reconstructed. To do this, we build upon recent methodological developments in normative political theory, which highlight the distinct character of political conduct and the norms that ought to apply to it.Footnote 26 According to these approaches, politics is a distinctive normative context for two main reasons. First, politics is primarily a collective endeavour rather than an individual one. Therefore the norms that apply to it should take into account the specific social and institutional contexts, and the behaviour of other actors given these contexts. Norms such as democratic accountability, for example, are specific to political conduct. Secondly, while moral concerns play a role in political conduct, political actors also pursue ends that cannot be subsumed into moral ends, such as order and stability. Furthermore, there is significant disagreement and uncertainty among actors regarding the value and justification of these ends. For these reasons, it would be a mistake to assume that moral norms, specifically those that apply to individual actions, can simply be applied, mutatis mutandis, to political conduct.
Understanding dilemmas politically, therefore, means interpreting them as particular relationships between the actions of political actors situated in specific social and institutional structures, and the effects of these actions. More specifically, political dilemmas arise where the nature of this relationship means that the pursuit of a desirable end would have an undesired effect for the political actor.
The clash between ends arises from structural and institutional constraints on political agency, not from the incommensurability of the ends themselves, independent of this context. In some circumstances, where dilemmas are avoided, the agent’s actions achieve their desired effects and advance her goals in a harmonic fashion; in others, her actions would fail, as the pursuit of desired outcomes with regards to one end generates undesired effects with regards to the other end. Even then, some such failures are not political if they result from bad luck or simple misjudgement by the agent, which explain why the negative effect happened. Other cases, however, are political when the clash between ends results from the interactions between different agents within specific social, legal and institutional structures. Interpreting the dilemma as political, therefore, requires an analysis of the distinctive normative context in which it arises.
This interpretation of dilemmas is ‘political’, in the sense we have described above, in two additional ways. First, unlike the traditional understanding of dilemmas in moral philosophy, it does not focus on individual agents and their moral integrity, but rather on the institutional context of their actions and their effects. Secondly, it obviates the conceptual debate over whether ‘genuine’ dilemmas exist, from an objective point of view. As the focus is on political agents, the pursuit of their ends and the tension between those ends, dilemmas pose a normative political problem whether or not they are merely subjective. This is particularly important because political judgement also involves acknowledging the limits of knowledge imposed on the actor under conditions of uncertainty, often not knowing how to weigh the options she faces, or even if they are commensurable.
With these points in mind, we turn to three models of political dilemmas that are relevant to the case of the donor’s dilemma, reconstructing classical accounts of moral dilemmas: complicity, double effect and dirty hands.Footnote 27 In all of these models, the political agent undermines an important end to which she was committed; however, the different structures of the relationship between action and effects suggest that the short- and long-term judgments are different for each of these models.
COMPLICITY DILEMMAS
In complicity dilemmas, the actions required to achieve a desired effect lead the agent to become entangled in the wrongdoing of others. To avoid complicity entirely, the agent would have to sacrifice the positive effect of her intended actions; but due to the normative context created by others’ wrongdoing, achieving this positive effect would advance the undesired, negative effects of their actions.
In our use of the term, stating that the agent faces a dilemma of complicity in the wrongdoing implies two things. First, that she does not share the wrongdoer’s objectives (otherwise, this is not a dilemma). In this, we part ways with theorists who view shared intentions with the principal wrongdoer as constitutive of the concept of complicity.Footnote 28 Secondly, it implies that her acts do not, in themselves, constitute part of the wrongdoing – they are only indirectly linked to it. Crucially, the actions advancing the positive effect would not be connected to the negative effects were it not for the principal wrongdoer’s actions.
Not all instances of complicity are political in the sense we discuss here. A famous example of moral complicity is provided in Book I of Plato’s Republic. Cephalus, Socrates’ interlocutor, suggests that justice entails speaking the truth and paying one’s debts. Socrates immediately points out that in some instances, given another agent’s wrongdoing, repaying one’s debts will generate a negative effect. If a friend is not of a right mind, and is inclined to harm himself or others (therefore becoming a wrongdoer), returning a borrowed knife to him would be a bad idea. Socrates argues that ‘repaying one’s debts’ cannot be the right definition of justice; alternatively, we can suggest that an agent in this predicament faces a moral dilemma of complicity: they can only repay their debt by being complicit with harm, or avoid complicity by not repaying their debt.
Political complicity is more complicated, as the relationship between the actions of the principal wrongdoer and the complicit agent is mediated through institutions. For the sake of simplicity, and without committing our account to Lepora and Goodin’s influential map of complicity and its ‘conceptual cousins’, it suffices to say that the agent can become complicit with the wrongdoing in at least one of two ways.Footnote 29 First, the agent can become complicit by causally contributing to the wrongdoing. Secondly, the agent can become complicit by condoning or legitimizing the wrongdoing, even when she is not directly causally contributing to the wrong act itself. Thus, for example, Iris Marion Young argues that Western consumers are complicit in the harms of sweatshop labour in the developing world – even though their individual purchases cannot be seen as causing this harm – because their continued consumption legitimizes the practices of the apparel industry.Footnote 30
Beyond these two paths, there are more complicated (and more controversial) cases of complicity that can generate the dilemma. These include cases in which the agent becomes complicit in the wrongdoing of others by benefitting from it, which we may interpret as a way of legitimizing the wrongdoing. There are also cases in which the agent is just one part of a collective that is committing the wrongdoing, and therefore it is more difficult to say that she has made a direct causal contribution to the wrongdoing.Footnote 31 As these are not relevant to the case of the donor’s dilemma – in which donors do not typically benefit from rights violations, and are not part of a collective agent – we disregard these possibilities in what follows. It is important, however, to note that their general structure is similar: the agent facing the dilemma is not the principal wrongdoer, and the dilemma arises from the link between her actions (towards the positive effect) and the actions of the principal wrongdoer (towards the negative effect).
DOUBLE EFFECT DILEMMAS
In moral philosophy, the Doctrine of Double Effect (DDE) could be summarized as maintaining that ‘[…] sometimes it is permissible to cause a harm as a side effect of bringing about a good result even though it would not be permissible to cause such a harm as a means to bringing about the same good end’.Footnote 32 A classic example in moral philosophy involves what is known as ‘Trolley Problems’, in which the lives of five people could be saved by sacrificing the life of one other person. Moral intuitions differ between impermissible cases in which the death of the one person is a means of saving the five, and permissible cases in which this killing is an unintended side effect.Footnote 33 The conditions for permissible double effects are restrictive. They may require, for example, that the action itself must be morally good or at least indifferent; that the positive effect must be produced directly by the action, not by the negative effect; that the positive effect must be proportional to the negative effect; and, perhaps most importantly, that the agent does not will the negative effect.
From the perspective of political theory, however, this account is problematic. Focusing on a particular agent’s intention is not obviously relevant for assessing actions and effects.Footnote 34 Yet simply dropping the condition of intentionality would not do, since the evaluation of the act would then turn on a comparison of the positive and negative effects, collapsing into a consequentialist position, which merely dissolves the dilemma.
A further complication arises from the observation that political action in complex societies is very different from the actions described in traditional examples of double effect. In these examples, the direct action of the agent is the source of both the positive and negative effects. In Philippa Foot’s famous example, a doctor who saves the life of a pregnant woman with cancer, but in the process causes the unintentional but foreseeable death of her foetus, may morally be in the right even if killing a foetus is morally wrong.Footnote 35 For our purposes, it is important to note that the doctor directly causes both the negative and positive effects. Political action, however, rarely has this direct relationship to effects, as it is often mediated by other agents. Even the direct orders of a political leader are carried out by other people. Strictly speaking, the free reactions of an agent to another agent’s actions cannot count as effect; they are seen as a new action that breaks the chain of causation.
The general structure of double effect, however, can still be useful for political analysis if properly reconstructed. We argue that a political reading of double effect dilemmas would interpret them as cases in which the structural constraints imposed on the political agent produce a situation in which actions towards desired effects generate negative side effects. These structural constraints may be caused by the agent’s own previous failure, by the failure or wrong actions of other agents, or, more controversially, by conditions that are not the fault of any agent in particular. In other words, whether the trolley problem is a moral or political dilemma will depend on how the people tied to the tracks wound up there.
Unlike complicity dilemmas, the negative side effect in a double effect dilemma is not generated by the actions of an independent agent, with which the primary agent is complicit. The negative side effect is generated by the actions of the agent facing the dilemma. Even when the dilemma is forced on the agent by the structure of constraints, it is still her actions that generate the wrong.
This institutional interpretation allows us to answer the problem of socially mediated action. Effects are distinguished from free reactions not simply by being what follows naturally from action, but also by being what follows from the ordinary operation of social institutions. For example, when we mail a letter, it could be said that the letter’s arrival at its destination is the effect of our action even though it is the postman who actually delivers it.Footnote 36 This final point is crucial for distinguishing cases of complicity from cases of double effect since, as stated above, in double effect dilemmas the negative effects follow from the actions of the primary agent. Understood structurally, we posit that double effect dilemmas arise not only in cases in which the direct action of the agent causes the negative effect, but also when the foreseeable actions of other agents, given certain institutional and structural failings, generate the negative effect. Secondary agents, in cases of double effect, are viewed as part of the structure in which the action of the principal agent is performed.
DIRTY HANDS DILEMMAS
In dirty hands dilemmas, the agent acts in a way that would generate a negative effect as a means of achieving the positive effect. Unlike cases of complicity, the negative effects are the result of the agent’s own actions, and not an independent wrongdoer’s. Unlike double effect, the negative effect is not merely foreseeable. Rather, it is seen as a means – perhaps the only means – of achieving the desired positive effect. One thing may cause another thing without being a means to an end, and this distinction need not rely on the controversial distinction between intending an effect and merely foreseeing it.Footnote 37
Unlike Complicity and Double Effect, the problem of dirty hands is usually discussed in a political context. One may ask, therefore, what makes the political interpretation we offer here distinctive?Footnote 38 We argue that even within the context of political action, the problem of dirty hands is often interpreted as a moral dilemma – a clash between the politician’s personal ethical commitments and the requirements of political power.Footnote 39 A classic and notorious example of this moral dilemma is the ‘Ticking Bomb’ case: a person with knowledge of an imminent terrorist attack is held by the authorities, and will only share this information if he is tortured. A responsible political leader may have to authorize such measures, but given that torture is morally wrong she should still feel guilty about it.
From the point of view of political theory, however, this reading of the dirty hands problem is misleading. Viewing the dilemma as a clash between personal morality and the obligations of a politician’s role unduly focuses on the individual agent, rather than on the structure of competing political objectives. Understood politically, dirty hands dilemmas arise where different political goals – for example, stability, security, fairness, justice – are in tension with each other, and some goals must be sacrificed or compromised to maintain others.Footnote 40 We do not deny that the clash between politicians’ role obligations and their own moral commitments may have an effect on political institutions.Footnote 41 Importantly, however, the political reading of the dirty hands dilemma is primarily not about the tension between personal morality and political necessity.
In line with the general rejection of dilemmas, some ethicists reject the conceptual possibility of dirty hands as ‘doing wrong in order to do right’ and argue that there is no special problem of dirty hands.Footnote 42 However, as with other political dilemmas, dirty hands cases may arise from previous wrongdoings of the agent or from the wrongdoings of others, and as such are not conceptually confused. Nevertheless, dirty hands are unique, since they represent an exceptionally difficult case – one in which the agent herself has to do wrong, and not just contribute to it indirectly or as a side effect. This distinguishes the political understanding of dirty hands from the moral one, as it identifies these dilemmas as rare exceptions, or limiting cases for the normative framework – unlike the more common cases of clashes between the politician’s personal morality and the responsibilities of political office.
A political reading of the dirty hands dilemma, therefore, must be able to identify the exceptional circumstances under which such radical action may be unavoidable. One possible route is to consider cases in which certain normative commitments are suspended to secure the conditions that enable them in the first place. Several theorists, for example, have argued that the ‘first political question’ – that is, securing and maintaining peace, order and stability – is both prior to, and a necessary condition for, the fulfilment of further political ideals such as socio-economic justice, political representation and government transparency. This position sees the legitimacy of any government as hinging on its ability to address the first political question, especially in cases of catastrophe or emergency.Footnote 43
Complicity, double effect and dirty hands – understood as models of political dilemmas – arise in different contexts and create distinct normative problems for the political agent. It is therefore important for the agent not only to recognize that she is facing a dilemma, but also which dilemma she faces. Real politics is, of course, a messy business. We do not wish to suggest that every context will perfectly fit the mould of one of our proposed models. For instance, there may be more complex cases of different dilemmas arising simultaneously. For the moment, we propose this framework as a conceptual distinction between different ideal types of political dilemmas. With the general framework in place, we can turn to the context of aid to regimes that are implicated in rights violations. This will highlight the complexities of the normative framework, as well as inform the justified responses to each specific dilemma.
NAVIGATING THE DONOR’S DILEMMA
In this section, we consider our general framework of political dilemmas in the specific context of donor aid provisions to states that violate rights. In doing so, we provide examples of each type of dilemma in the development context. We also discuss how donors can recognize the type of dilemma they face, and their potential short- and long-term responses to navigate the specific dilemma.
In brief, we argue that the three dilemmas differ according to the relationship between the actions of political actors situated in specific social structures and the effects of these actions. To distinguish between the dilemmas that donors face, we first need to establish whether the rights violation is connected to positive development outcomes. If the wrongdoing is considered necessary for the government to bring about the stability and order needed to achieve positive development outcomes, the donor is faced with a dirty hands dilemma. If the wrongdoing is unrelated to the positive outcome, then the donor faces either a complicity or double effect dilemma. To distinguish between these latter two dilemmas, we need to consider whether the aid the donor provides has a direct causal role in the rights violation. If the provision of aid gives rise to the wrongdoing, then the donor faces a double effect dilemma. If, however, it is likely the rights violation would have occurred regardless of the aid, we argue the donor is faced with a complicity dilemma.
It is important to clarify that what follows is not intended to be an algorithm for resolving political dilemmas; nor do we provide practitioners with an answer to the perennial question of ‘what to do on Monday morning’.Footnote 44 Nevertheless, we argue that this framework can help highlight the relevant normative and political considerations in each context.
AID AND COMPLICITY
Complicity dilemmas are cases in which, given the wrongdoings of the recipient government, the donor government action towards development goals either directly contributes causally to rights violations or provides implicit legitimation for them. The principal wrongdoer is the recipient government, since its violation of rights generates the dilemma. This violation of rights is not generated or constituted by the aid, and is not a means of achieving positive development outcomes. Therefore the problem donors face is that the funding they provide solidifies the power of the recipient government and provides it with legitimacy.
Complicity is the most widely discussed normative context in the literature on the ethics of aid. A paradigmatic case is the abuse of international aid resources (provided by INGOs) in refugee camps in Zaire, Burundi and Tanzania during and immediately following the 1994 Rwandan genocide.Footnote 45 Members of the Hutu-dominated Rwandan Armed Forces, who sought to re-establish their military force, abused humanitarian aid and used civilians in the camps as human shields. As Rubenstein describes, ‘[S]hould humanitarian INGOs that are committed to acting consistently with democratic and egalitarian norms accept contributing knowingly but unintentionally to injustices committed primarily by other actors?’Footnote 46
These cases are structurally similar to the complicity dilemmas of state donors, although arguably, unlike INGOs that only serve quasi-governmental functions, state donors have more direct responsibility for the political effects of their actions. This is because they are both able to exert more influence on recipient governments and are accountable to their publics. Therefore the threshold for complicity is lower. One possible way of identifying complicity is to track the causal contribution to the wrong committed by recipient governments. For example, as Peter Uvin demonstrates in Aiding Violence, French development aid in Rwanda was used to acquire weapons later used in the Tutsi genocide.Footnote 47 There is no question here that the principal wrongdoer is the Hutu government, of course, but as Uvin’s narrative demonstrates, France and other Western states can be seen as complicit in this wrong since aid money was flowing in even when there were clear warning signs of the real possibility of mass violence, such as the escalation of racialized propaganda.
While providing external legitimacy to regimes that violate rights may not causally contribute to the rights violation directly, it can be considered a case of complicity. The most egregious recent examples were common during the Cold War, when Western aid – albeit with the ulterior intention of preventing the spread of communism – legitimized leaders who were on the ‘right’ side of the Iron Curtain, while at the same time strengthening their position vis-à-vis their internal opposition. The dictatorships of Mobutu Sese Seko in the Democratic Republic of the Congo and the Duvaliers in Haiti are glaring examples of Western failures to recognize this tension. For our purposes, the fact that the primary goal of aid during the Cold War was one of political influence rather than development is not significant; the structure of the dilemma remains the same.
As is evident from the example of French aid to Rwanda, ignorance of causal mechanisms is not an excuse, given the responsibility of donor states for the effects of their actions. To avoid blind complicity, donors need to invest resources in working with local stakeholders who may provide valuable contextual knowledge, identify reasons for concern and make it possible to monitor the situation on the ground to avoid culpable ignorance.
If and when complicity in rights violations is recognized, in most cases – at least in the short term – it is likely that the best course of action for donors is to detach; in other words, to withdraw aid and sever the causal or legitimizing contribution to the wrongdoing. This is clearly the case when the relative weight of the negative effect is considerably higher than any benefit generated by aid. Even in more nuanced cases, complicity in wrongdoing can undermine the positive effects of aid.Footnote 48 Nevertheless, conceptualizing the problem as a dilemma highlights the fact that withdrawing aid generates ‘moral remainders’, as the positive effect of development is now withdrawn. How best to address these remainders depends on the context – for example, whether aid was in place but then withdrawn, creating dependency. It would also not be enough simply to keep hands clean, especially in cases in which there was previous involvement; in these cases, there is a need to publicly denounce the recipient government’s wrongdoing, in addition to detaching.
We further consider donors’ short-term responses to complicity dilemmas by briefly discussing two recent examples. The first is the withdrawal of DFID’s and other development donors’ budget support to Malawi in 2011, which we argue is a positive example of how donors can respond to complicity dilemmas in the short term. The decision to stop providing aid directly to the Malawian Government was motivated by concerns about the country’s economic mismanagement and governance.Footnote 49 DFID issued a clear statement that explained the suspension of budget support was in large part due to deteriorating governance in the country, including the suppression of demonstrations, the intimidation of civil society organizations and the introduction of legislation that would make it easier for the Malawian Government to place restrictions on opponents without legal challenge. Andrew Mitchell, the UK secretary of state for international development at the time, added:
The UK provides development assistance in order to help communities lift themselves out of grinding poverty […] But poor people in Malawi and British taxpayers alike have been let down. In those circumstances I cannot justify the provision of general budget support for Malawi […] In the meantime we will use other means to ensure that programmes to protect poor Malawians […] are able to continue.Footnote 50
The decision to suspend budget support had a significant negative impact on the Malawian economy and on living standards in the country – demonstrating the dilemma the UK government faced.Footnote 51 However, as the statement makes clear, DFID and other donors continued to support initiatives to improve basic needs such as health, education, and water and sanitation through project aid.Footnote 52 In other words, while DFID’s decision had a negative effect, efforts were made to address the moral remainders created by DFID’s withdrawal of budget support. It is also worth noting that DFID explicitly linked the re-establishment of budget support to addressing issues of economic mismanagement and governance. Given the Malawian Government’s dependence on UK aid, the withdrawal of aid appears to have led to efforts to improve governance, which DFID has monitored and publicly encouraged.Footnote 53 DFID’s short-term response to the deterioration of political rights in Malawi, we argue, thus provides a good example of how donors might address complicity dilemmas.
In contrast, DFID’s response to accusations of complicity in rights violations committed by the Ethiopian Government demonstrates the problems with more typical donor responses. In late 2014, DFID suspended its project to support the Ethiopian police force at the same time Amnesty International produced a report documenting how Ethiopian security forces had tortured, mutilated, raped and murdered Ethiopians to suppress political opposition. Yet despite an internal assessment of the police support project that suggested it posed a high risk of being complicit in human rights violations, DFID denied that the cancellation of the project was related to concerns about human rights. The internal assessment document was also removed from DFID’s website.Footnote 54
There may be cases in which the benefits of aid, or the foreseeable negative effects of withdrawing aid, outweigh the negative effects of complicity.Footnote 55 In such cases, the best short-term response to the dilemma would not necessarily be detachment. However, it is important to note that this is merely the short-term, local response. A context in which complicity with the rights violations of a wrongdoer is the lesser evil is one that needs to be transformed, if and when such transformation is feasible. Donors may accept a certain level of complicity as a necessary evil but, given the dilemmatic structure of this case, they need to ensure they do not lose sight of their responsibility for the harmed parties, or of the imperative of reform.
AID AND DOUBLE EFFECT
Unlike complicity dilemmas, the negative effect of rights violations in double effect cases is seen as a consequence of the actions donors take towards positive development goals, rather than as the actions of independent wrongdoers. As we have pointed out, this need not mean that this is the direct effect of the donors’ actions, but it is a foreseeable side effect that is affected by structural and institutional constraints. It may be that the actions of the donor are mediated through the actions of participants in social institutions; however, the key point is that it is the donors’ actions that ultimately brought about the negative side effect.
Some examples of double effect dilemmas of aid are not directly related to rights abuses. For example, malaria prevention is clearly a beneficial effect of aid, and the free distribution of mosquito nets is considered a cost-effective and targeted measure towards achieving this goal. However, recent research in Zambia shows that the introduction of light, strong and free insecticide-treated nets has led to the unintended consequence of local communities using the nets for fishing rather than for their intended purpose. This has adverse effects not only on the wildlife of Lake Tanganyika, but also the health of local communities whose fish and drinking water are now contaminated by insecticide.Footnote 56
An example in the context of aid and rights is the case of the DFID-supported Local Government and Community Development Programme (LGCDP) in Nepal. The programme aimed to strengthen local governance by establishing local representative bodies that can influence government spending decisions. However, the Independent Commission for Aid Impact (ICAI) found that the project ‘increased the opportunities for corruption in society’.Footnote 57 Citizens had to pay bribes to government officials and forge documents to receive funding for projects supported by the LGCDP. The ICAI thus criticized the programme for pushing the poor towards corrupt practices, and increasing ‘the perception among the poor that corruption is a necessary and legitimate way of acting.’Footnote 58 Furthermore, the aid provided by DFID and other donors were captured by political elites, who directed the funds towards their preferred projects. The Economist’s investigation into the programme found that struggles among political elites to capture these funds led to violence in some districts. More broadly, the programme is seen as having led to a situation at the local government level where ‘politicians enjoy freedom to plunder with impunity’.Footnote 59
This case illustrates how double effect dilemmas differ from complicity dilemmas. This is particularly relevant when mediating agents cannot be seen as autonomous, but are instead part of the foreseeable political and institutional context. In states with weak institutions, donors’ actions can introduce new incentives that may have perverse effects on local government stability and accountability.Footnote 60 If the negative effect of rights violations can be traced back to the donor’s actions (and not, as in complicity cases, where donors only facilitate or legitimate the actions of the recipient government), then the normative context is one of double effect. The LGCDP case is an example of the double effect dilemma because donors’ aid provisions in Nepal’s weak institutional context created a situation in which politicians in local government could plunder from the local population. Furthermore, according to the ICAI, this effect of the aid was foreseeable given the political context.Footnote 61
It is important to emphasize that the difference between double effect and complicity is not one of moral responsibility. The fact that corruption was a foreseeable side effect given structural constraints does not relieve the corrupt officials in the recipient state of blame. However, unlike cases of complicity, since the negative side effect could be predicted on the basis of structural constraints rather than the agency of a particular leader or government, the political responsibility of donor states is of a different kind. In other words, the relevant distinction is not of moral responsibility (who is culpable), but of recognizing the link between actions and effects and best resolving the dilemma.Footnote 62
To recognize a double effect dilemma, it is therefore necessary to have both contextual knowledge and an understanding of political and social mechanisms. Donors need to ask whether the negative effects of aid would have occurred if a different local actor were in place. This would enable donors to understand whether the cause of the wrongdoing is structural or agential. In thinking about this, we need not say whether this other local actor actually exists at the present time, but whether the actor could feasibly exist given the history and institutional structures of a given context.
As the negative side effects arise from the donor’s own actions, the focus of short-term responses to double effect dilemmas needs to be on the proportionality of the positive effects of development outcomes and the negative side effects. Since the negative effects are side effects and not a means to an end, the aim would be to minimize them as much as possible and to engage with and compensate affected parties. Furthermore, as the relationship between actions and effects is structural and not agential, the short-term aim will be to strengthen institutions rather than address the wrongdoings of specific actors.
The types of measures to put in place when faced with double effect dilemmas include a greater emphasis on monitoring the impact of development programmes and more frequent and direct interactions between donors and stakeholders. Indeed, both recommendations were made by the ICAI in relation to the DFID-funded LGCDP programme in Nepal.Footnote 63 Additional measures include ensuring that accountability mechanisms are built into programme design. For example, the Asian Development Bank’s (ADB) Mekong Business Initiative project includes providing a forum ‘where people adversely affected by ADB-assisted projects can voice and seek solutions to their problems’ and register complaints.Footnote 64 In addition, donors may need to compensate those adversely affected by their programmes.
In the long term, however, proportionality is not enough. The institutional and structural failures that generate the dilemma in the first place need to be addressed and ameliorated in order to escape the dilemmatic context. The focus of long-term response, therefore, will be on reforming the institutional structures that generate the negative side effects as well as restoring agency to recipient governments. The latter type of response, of course, needs to take complicity concerns into account.
AID AND DIRTY HANDS
In dirty hands dilemmas, the violation or limitation of civic rights is seen as necessary for sustaining the basic order and stability of the state when these are threatened. Without these, the possibility of achieving socio-economic development and protecting political and civil liberties in the long term is undermined. This means that in the dirty hands scenario the negative outcome (rights violations) is seen as a means of achieving the positive outcome. In other words, the negative effect is not a side effect or unrelated to the positive effect as in the previous dilemmas; it is a necessary step towards achieving the positive outcome. The recipient government is faced with a dirty hands dilemma of the first order; the donors face a dirty hands dilemma of the second order, by providing aid to the recipient government.
There is evidence from the history of development to support the theory that democracy and rights protection requires a minimum level of economic development, and that autocratic developmental regimes may be best placed to achieve this minimum level. An important example is South Korea’s development process. The country’s transformation from one of the poorest nations in the world in 1960 to one of the wealthiest by the early 2000s is widely viewed as ‘one of the great development success stories, of the contemporary, or any other, era’.Footnote 65
The emergence of the developmental state in South Korea followed a military coup in 1961, which brought General Park Chung-Hee to power. The Korean developmental state experimented with different policies to promote industrialization and high rates of economic growth.Footnote 66 The government was able to implement its industrialization strategy because of ‘extensive access to American and international aid funds’.Footnote 67 Subsequently, South Korea is held up as a clear example of how ‘foreign assistance combined with sound institutions and policies […] produce good results’.Footnote 68
Importantly, South Korea’s development success occurred under an autocratic regime. Indeed, the country is used as an example of how autocratic regimes may be better placed to implement long-term measures to bring about economic transformation,Footnote 69 despite the substantial human rights abuses that took place under its autocratic regime.Footnote 70 It was only in the late 1980s, following mass pro-democracy protests, that Korea adopted a democratic system that led to the protection of civil and political rights. However, this came about once a minimum threshold of development was achieved. Hence, for many the Korean example demonstrates how the process of economic development can lead to democratic governance and the protection of civil and political rights. The short-term response from donors to a dirty hands dilemma, therefore, needs to focus on questions of political legitimacy, and for this task, donors will need evidence that the rights violations are being used as a means to a justifiable end. This evidence cannot merely rely on the claims of the recipient state but, as in the case of complicity, would need to build on stakeholders’ local knowledge. This is further complicated by the fact that even when some government policies might be legitimate, others are not, and donors should be sensitive to this fact. As the wrong committed is serious, the threshold for accepting this evidence must be high.
Legitimacy is a key factor in dirty hands dilemmas. As noted previously, developmental regimes gained legitimacy through their ability to deliver positive outcomes in the short term. It is crucial to acknowledge that – unlike in cases of complicity – the recipient government may need to be legitimized, domestically and internationally, in order to successfully maintain the level of stability required for development. Donors’ endorsement of the recipient government will be conditional, and both the donor and the recipient government will still need to acknowledge that a wrong is being committed while maintaining that this is necessary given the circumstances.
In the longer term, the donor needs to ensure the recipient government retains its legitimacy through setting clear timeframes and identifiable goals. These goals underlie the transitional nature of the dirty hands dilemma. Unlike traditional political conditionalities, these would have to be more flexible and context specific: they will differ from case to case according to the particular history, political constraints and feasibility of reform in each state. According to the rationale of dirty hands situations, current rights violations are transitional in that they are helping to pave the way towards a time when such rights violations do not take place and politics is ‘normalized’.
Donors would also need to ensure that they are not promoting a particular leader or political party, but are focusing on the function of governments in maintaining conditions of stability and order; donors have frequently failed to do so.Footnote 71 This failure restricts the move to politics as normal, and prevents donors from being flexible and adapting to changing circumstances. This is especially crucial when the normative context has moved from dirty hands to complicity, requiring a different kind of response.
RECOGNITION AND RESPONSE
Having discussed each of the dilemmas that can arise from providing aid to states that violate rights, we now consider the three dilemmas together. We outline the main differences between the dilemmas, further demonstrating how each can be recognized. This also provides further justification for the response to each of the dilemmas that we have identified.
As we have explained, the three dilemmas differ according to the relationship between the actions of political actors situated in specific social structures and the effects of these actions. This leads to two considerations when attempting to identify which of the three dilemmas an aid donor faces. The first is whether the rights violation is connected to positive development outcomes. If the wrongdoing is deemed necessary for the government to bring about the stability and order required to achieve positive development outcomes, then this would imply the donor is faced with a dirty hands dilemma. If the wrongdoing is unrelated to positive outcomes, then the donor is faced with either a complicity dilemma or a double effect dilemma. The second consideration, which enables us to distinguish between the latter two dilemmas, is whether aid provided by the donor has a direct causal role in the wrongdoing unmediated by an independent wrongdoer. If aid provisions give rise to the wrongdoing, then the donor faces a double effect dilemma. If, however, it would be reasonable to believe that the wrongdoing would have occurred regardless of the aid provision, then the donor is faced with a complicity dilemma.
The differences between the three dilemmas are related to the issues of agency and legitimacy. In both complicity and dirty hands dilemmas, the recipient government is seen as exercising its agency in committing the rights violations; the distinguishing factor is the background context. The required response to each of these dilemmas is very different, as a ‘false positive’ (wrongly identifying cases of dirty hands as complicity) risks undermining the conditions needed for rights to be secured, while a ‘false negative’ (wrongly identifying cases of complicity as dirty hands) risks the donors becoming complicit in the rights violations of the recipient government. In contrast, double effects dilemmas are those in which rights violations are seen as being generated by structural factors. Therefore, distinguishing between double effect and complicity dilemmas requires donors to assess whether agents or structures generate the negative effects.
The issue of legitimacy is also important in considering the differences between the three dilemmas. In the complicity and dirty hands dilemmas, aid provides the recipient government with de facto legitimacy. In the former, this is seen as unjustified and undermining the donor’s objectives. This is why the short-term response may be detachment. In the case of dirty hands dilemmas, this de facto legitimacy is seen in a more positive light, as it helps the recipient government address the first political question of providing stability, thereby contributing to achieving the donor’s goals. Hence, the response to dirty hands dilemmas is to ensure that there is reason to believe that the actions of the recipient government are providing political stability. In contrast, aid does not have a legitimizing effect in the case of double effect dilemmas, as the recipient government is not seen as an independent agent. Hence, the response to double effect dilemmas is to minimize the negative effects of aid, and ensure there is some proportionality between the positive and negative effects of aid.
CONCLUSION
In this article, we have considered the dilemmas donors face in providing aid to states that violate rights. By developing a distinctively political framework for analysing dilemmas, we have shown that the donor’s dilemma is, in fact, three different dilemmas. Complicity, double effect and dirty hands dilemmas in development aid arise in different contexts and require different responses. Using the context of development aid has also enabled us to highlight the differences between the models of dilemmas we have identified in our conceptual framework. It shows that the role our conceptual framework plays will depend on how the political agent perceives its goals, and the particular circumstances in which its actions give rise to both positive and negative effects.
Contrary to dominant approaches in development theory and practice, our argument demonstrates that a politically realistic approach to the donor’s dilemma, which is sensitive to context and political complexity, need not be normatively silent. Rather, it is important for donors to acknowledge that they face a dilemma, to recognize what kind of dilemma they face, and to respond accordingly. Acknowledging the dilemma is important because both recognizing the dilemma and deciding on the appropriate response require engaging with relevant actors and stakeholders.
A second insight of our research is that focusing on actions and effects would help donors address these dilemmas in a structural, rather than personal, manner. Acknowledging such dilemmas has frequently been hindered by donors’ tendency to conflate the choice between the two horns of the dilemma with the choice of whether to support or oppose the contingent leadership of the recipient state. As we have discussed, even when legitimizing the actions of the government may be important, an overemphasis on a particular leader or party is more likely to constrain the transition away from the context of the dilemma rather than facilitate it.
The third insight is that it is necessary to go beyond the dichotomies of aid provision, which tend to emphasize only two choices – fully endorsing financial support to developmental regimes, or withdrawing aid to preserve the moral integrity of the donors. The lack of attention to contextual constraints and to the dilemmatic nature of this problem leads to perverse outcomes for both of these well-intentioned approaches – for example, mistaking complicity dilemmas for dirty hands dilemmas. The framework we have proposed demonstrates the nuanced nature of these situations, and the different responses and measures available to donors based on better understanding the situation they face.
Finally, normative considerations aside, there are likely to be sound instrumental benefits to the framework developed in this article. The long-term success of development aid depends on consistent public support.Footnote 72 Some may see this explicit acknowledgement of the dilemmas donors face as risking support for development aid more broadly; we argue the opposite. As the media discussion of cases like that of Mr O and DFID’s support for the Ethiopian Government demonstrate, there is a real danger of a backlash against international aid where rights are seen to be violated. By explicitly acknowledging and addressing these dilemmas, donors would be able to make the case for responding differently to normatively distinct situations, thus strengthening the public legitimacy of development aid.
