Nudging Charitable Giving: The ethics of Nudge in international poverty reduction

ABSTRACT The use of nudge theory by non-governmental organizations and governments as a means to encourage charitable donations in order to address global poverty has been the subject of much recent enthusiasm. Supporters argue that nudges aimed at altering behaviour are much less complex and normatively problematic than attempts to alter attitudes towards global poverty. This paper aims to challenge these assumptions. In order to do so, it identifies problems and shortcomings of the nudge approach, suggesting we should look again at alternative strategies. The argument consists of three parts. Section 1 outlines the nudge approach, dividing nudges into three distinct categories. Nudges in the first category, which operate through simple information provision, face no compelling objections as a means to motivate charitable giving. Section 2 examines nudges in the third category, typically considered the most pernicious form of nudge in the critical literature. As these nudges aim to bypass reflection, their use faces distinct normative problems in the context of global poverty. The third section, comprising the bulk of the discussion, examines the complex case presented by second-category nudges, which capitalize on prudential sources of motivation in order to motivate action to address global poverty. It argues that, although such cases may appear straightforward instances of favouring good outcomes over good motives, the reality is much more complex. Altering behaviour provides no easy substitute for the complex and vital task of altering attitudes towards global poverty.


Introduction
Global poverty represents a problem of pressing normative urgency. 1 Despite broad agreement that addressing this problem requires some action by ordinary citizens within the more affluent communities of the world, action is rarely forthcoming. In order to motivate this action, the following two broad strategies have been proposed: firstly, that we try to alter the underlying dispositions of individuals, increasing levels of concern for the global poor, with the hope that this change in attitudes will engender action to reduce global poverty 2 ; and secondly, that we attempt to influence actions directly, by prompting changes in unreflective behaviour or altering the conditions in which choices are made.
Nudge theory, 3 as advocated by the British Government's Behavioural Insights Team, and corresponding 'Nudge Units' in the United States and elsewhere, represents a version of this second strategy. Supporters argue that altering behaviour directly is a much less complex and normatively problematic affair than attempting to alter attitudes. 4 This paper aims to challenge these assumptions. In order to do so, it identifies problems and shortcomings with the nudge approach that suggest we ought to reconsider the former strategy.
The paper consists of three parts. Section 1 outlines and clarifies the nudge approach, offering a three-part typology of nudges appropriate for the context of charitable giving. It then proceeds to assess nudges in the first category, arguing that their use in this context is relatively benign. Section 2 examines nudges in the third category, typically considered the most pernicious form of nudge in the critical literature. It argues that as these nudges circumvent reflection on an issue, their use to motivate donations towards reducing global poverty faces significant normative objections. However, opportunities for their employment in this context are limited. The third section, comprising the bulk of the discussion, examines the complex case presented by second-category nudges. It argues that these pose a distinctively problematic means of altering other directed behaviour, as they are unlikely to engender a corresponding change in normative attitudes.

Part I: defining nudges
In the book in which they coined the term, Thaler and Sunstein describe a nudge as 'any aspect of the choice architecture that alters people's behavior in a predictable way without forbidding any options or significantly changing their economic incentives '. 5 This is unhelpfully broad, and encompasses some techniques they explicitly rule out, such as minor fines or other low-level sanctions. 6 However, Thaler and Sunstein's definition draws attention to a key difference between nudges and more traditional persuasion techniques. Distinct from traditional fundraising techniques, or government use of social advertising, nudges primarily aim to alter behaviour, rather than achieve attitudinal change. As Mols et al. put this, nudges attempt to achieve norm compliance, as opposed to norm internalization. 7 Similarly, McTernan differentiates between traditional interventions 'where the cause of the behavior is internalizing the norms of one's society and acting accordingly… [and] nudges, where behavior is influenced through minor situational factors.' 8 Nudges are typically illustrated through the use of examples. The paradigm cases are the etching of a fly in urinals at Schiphol Airport to 'improve aim', and the 'Don't Mess With Texas' anti-littering campaign, in which American football players were employed to help create associations in the minds of young men between disposing of rubbish responsibly and masculinity. 9 Judith Lichtenberg offers the following diversity of examples in her account of utilizing nudges to facilitate charitable giving: the use of defaults 'for example, an opt-out system on income tax forms in donating a certain amount to reducing poverty'; pre-commitment, capitalizing on the widespread bias to 'overvalue present satisfactions and undervalue future ones' to encourage people to commit now to substantial future donations; and framing donations so as to make them seem less burdensome, eschewing larger upfront figures in favour of a focus on 'pennies a day'. 10 The breadth of these examples demonstrates that any attempt to subject nudges to normative assessment requires drawing finer distinctions, as a number of different strategies of behavioural modification are encompassed under the general term. Accordingly, both the effectiveness and the normative status of these strategies as means to motivate donations to global poverty charities will differ.
It is important to note the two following factors, which are specific to the assessment of nudges in the context of charitable giving: firstly, nudges are being used here to alter other-directed behaviour, not self-directed behaviour such as lifestyle choices. Therefore, paternalism, as traditionally understood, is not the main concern in this context. Thaler and Sunstein define paternalism as 'making choosers better off, as judged by themselves.' 11 However, nudges aimed at promoting charitable giving neither require the nudging authority to know an individual's best interests, nor do they aim to make nudged individuals better off. Similarly, the use of nudges is often objected to on the basis that it is manipulative, or even coercive, 12 but there is a strong prima facie normative justification for the use of such techniques in this context, in the form of realizing the basic rights of the global poor.
Secondly, nudges are not being employed here to address bounded rationality leading to reasoning failure by the well intentioned. Instead, the issues nudges seek to address are, in significant part, attitudinal, with individuals not caring, or not caring enough, about the problem presented by global poverty. Accordingly, employing nudges faces distinct normative concerns in this context, as they do not simply operate to bring actions into line with considered attitudes, but to achieve behavioural changes that may not reflect an underlying normative commitment.

Three types of nudge
To effectively assess the use of nudges as a means to motivate donations towards reducing global poverty, it is necessary to separate nudges into three distinct categories. I broadly follow the three-part distinction drawn by Baldwin; however, whereas Baldwin's central concern is the extent to which nudges interfere with autonomy, the following categories differentiate between the different ways nudges interact with an agent's motivational structure. 13 First-category nudges: nudges in this category have been the source of some controversy in the literature. Many critics 14 do not consider them to constitute nudges at all, as they operate at a reflective level. First-category nudges aim to enhance decisionmaking through 'the supply of simple information…or the imparting of reminders.' 15 Examples include 'warning labels on cigarettes' and 'signs warning people on a hot day to drink more water.' 16 These interventions are characterized by operating at a deliberative level, and aiming to alter behaviour without attempting to alter an agent's existing motivations.
Second-category nudges: Baldwin's second category contains nudges that 'build on behavioral or volitional limitations so as to bias a decision in the desired direction…[Where] the target of the nudge would be capable, on reflection, of realizing that a nudge has been administered.' 17 I define second-category nudges as those that utilize non-altruistic motives to increase giving. Where this second category overlaps with Baldwin's is that, in both cases, the nudge is available to conscious reflection. The paradigm example of a second-category nudge on either account would be the 'Don't Mess with Texas' anti-littering campaign. Here, the decision not to litter is biased in the desired direction by the motive to live up to an ideal of masculinity. However, on reflection, the campaign's attempt to influence behaviour is consciously available.
Third-category nudges: rather than taking advantage of an alternative source of motivation, these nudges operate by circumventing motives altogether, with the agent making a donation without necessarily being aware that they have done so. A typical example would be switching from an opt-out to an opt-in default in a workplace giving scheme. Rather than capitalizing on prudential motives, the desired behaviour is primarily secured through inattention, with the success of the nudge relying on this lack of transparency. This has affinities with Hansen and Jespersen's definition of a non-transparent nudge, where 'the citizen in the situation cannot reconstruct either the intention or the means by which behavioral change is pursued.' 18 Thaler and Sunstein are likely aware that this form of nudge is normatively problematic, and, influenced by Rawls, they include a publicity condition in their account, where a government must be 'able or willing to defend [nudges] publically to its own citizens.' 19 Crucially, however, this is only a hypothetical requirement, because as Thaler and Sunstein acknowledge, pointing out a default would render the default ineffective. 20 In contrast, Rawls' publicity principle is not hypothetical, requiring 'that principles be known and understood by the public, not merely that they be publically defensible.' 21 Due to the deceptive manner in which they operate, third-category nudges are considered the most pernicious form of nudge in the critical literature. 22 14 Hausman and Welch, "To Nudge or Not to Nudge"; and Yeung, "Nudge as Fudge." 15 Baldwin, "From Regulation to Behaviour Change," 835. 16 Thaler and Sunstein, Nudge, 189-244. 17 Baldwin, "From Regulation to Behaviour Change," 836. 18 Hansen and Jespersen, "Nudge and the Manipulation of Choice," 18. 19 Thaler and Sunstein,Nudge,244. 20 As an anonymous reviewer pointed out, a hypothetical publicity condition is not a genuine publicity condition. 21  Having drawn these distinctions, I now examine the normative status of these three types of nudge as a means to motivate donations towards reducing global poverty. It is important to note that the categories are discussed out of numerical order. I begin with the more straightforward cases of firstand third-category nudges, and then move on to examine the complex case presented by the employment of second-category nudges in this context.

First-category nudges
First-category nudges are not typically thought to compromise autonomy in the decision-making process. 23 These nudges consist of the public display of factual information; therefore, the idea that they are intrinsically normatively problematic is unpersuasive. This would rule out public information signs and other everyday instances of straightforward information provision. Therefore, if first-category nudges face normative objections, these must derive from their employment within a charitable context. Firstly, I examine the objection that first-category nudges are inapplicable when motivating charitable giving, as instances of 'pure information provision' cannot occur in such an emotive context. Having established the applicability of first-category nudges in this context, l move on to examine their use as: (i) a means to facilitate donations; and (ii) a means to inspire deliberation on global poverty.

Information provision
Instances of 'pure' information provision may be thought unlikely to occur once we move from the heavily debated terrain of health nudges to the context of charitable giving. This is due to the emotive nature of the information typically featured in charitable campaigns. The information that '25,000 children die of poverty-related causes everyday,' 24 for example, is likely to provoke a significant affective reaction. This has led some critics to conclude that first-category nudges, understood as 'simple information provision', cannot exist in charitable campaigns. Hansen and Jespersen, for example, understand attempts to solicit donations that provoke an emotional response as attempts to interfere with 'rational deliberation', constituting manipulative cases of nudges proper. 25 I believe this conclusion is incorrect and demonstrates that these accounts of nudges are inadequate when assessing charitable giving. This is because they operate with accounts of rationality that are inappropriate in a normative context, failing to recognize the centrality of affect in moral deliberation. As Williams 26 has convincingly argued, and recent advances in neuroscience have shown, 27 moral deliberation necessarily involves a significant affective component. Therefore, it is unclear that limiting the emotive force of information will necessarily lead to a more 'rational judgement', as normative reasoning involves both affective and deliberative elements. Accordingly, materials that provoke sentimental concern cannot be assumed to be an 23 Hausman and Welch, "To Nudge or Not to Nudge." 24 Overland, "602 and One Dead," 281. 25 Hansen and Jespersen, "Nudge and the Manipulation of Choice." 26 Williams, Moral Luck. 27 Krause,Civil Passions,8. attempt to take advantage of our 'bounded rationality'. In the context of charitable campaigns, instances of information provision provoking affective concern are therefore still plausibly classified as first-category nudges. This still allows that highly emotive campaigns interfere with rational deliberation, as I have not claimed that moral deliberation is purely affective, but that it is implausible to view ideal moral deliberation as free from affect.
In the context of international poverty reduction, first-category nudges can be utilized in two ways. Firstly, they can facilitate donations by providing information concerning ways to donate, such as offering reminders or making donating easier. Secondly, they can provide factual information concerning a charitable cause, aiming to inspire deliberation leading to a donation.

Facilitating charitable giving
On the first method, instances of information provision draw attention to opportunities for charitable giving, making donating easier or more salient. Here, the motivational structure of potential donors remains unchanged, but the situation is framed such that a correlation between the behaviour and the attitudes of individuals who already care about the issue of global poverty becomes more likely. The use of first-category nudges to facilitate charitable giving in this manner appears uncontroversial. However, David Miller has raised the possibility that making altruism easier fails to sufficiently exercise our altruistic motives, undermining the capacity for altruism in the long run. As Miller puts this, 'the theory here being that altruism is a quality that is strengthened by being exercised and atrophies otherwise.' 28 This is not simply the argument that the correct motives must be present, either for intrinsic or consequential reasons, which would be unpersuasive, as first-category nudges do nothing to facilitate non-altruistic sources of motivation. Instead, it is the claim that if the correct motive is present, it ought to be 'stretched' to yield beneficial long-term consequences. However, this argument provides no way of knowing when the point is reached at which the beneficial consequences of the altruistic act outweigh the beneficial consequences of developing the altruistic capacity. The broad empirical claim underlying this argumentthat altruism develops when it is testedalso does not withstand scrutiny. Numerous social psychology studies testify to the common-sense conclusion that making altruism more difficult typically results in altruistic behaviour becoming less likely. 29 Inspiring moral deliberation Alternatively, information provision can focus on the charitable cause itself, with the aim of inspiring deliberation on global poverty. For example, information concerning the over 700,000 children under five years of age who die from malaria each year, when an insecticidal net costs $4, 30 leads to deliberation and ideally a donation. It is difficult to discern the utility of applying the nudge label to these interventions. However, Hansen and Jespersen define this as a 'type-two' nudge, which targets a deliberative, as opposed to an automatic, thought process. The rationale for labelling this a nudge is 28 Miller, Justice for Earthlings, 199. 29 Darley and Batson, "From Jerusalem to Jericho"; and Zimbardo, The Lucifer Effect. 30 Givewell, "Against Malaria Foundation." that it relies on 'framing effects', where the frame or context in which a decision problem is formulated is used to influence reflective choice. 31 As Thaler and Sunstein note, it is impossible to avoid framing effects altogether, as all information is presented in a context. 32 However, this fails to distinguish between active and passive senses of framing, with active framing consisting of attempts to influence behaviour by altering the context in which information is presented and passive framing constituting the acknowledgement that context influences how information is perceived. Attempts to 'frame' global poverty as a normative problem in order to encourage charitable donations can only constitute framing in the passive sense, as this is the context in which this information typically features in charitable campaigns. Defining passive framing as a type-two nudge fails to distinguish between attempts to alter behaviour through manipulating context, and the fact that all attempts to alter behaviour take place in a context that is thought to be conducive to this purpose. Instances of information provision presenting global poverty as a normative problem constitute the typical method through which charities appeal for donations. As these interventions operate at a reflective level and in a transparent manner, with the resulting donations typically supported by corresponding moral attitudes, these are straightforward cases of information provision. As the use of firstcategory nudges in campaigns to reduce global poverty faces no significant normative issues, they will be absent from the remaining discussion. I now move on to discuss nudges in the third category.

Part II: third-category nudges
Third-category nudges are considered the most normatively problematic form of nudge in the critical literature, as they operate by bypassing reflection altogether, and their effectiveness is dependent on this lack of transparency. As Bovens notes, these nudges 'work better in the dark' and their 'effects…are likely to disappear if they become transparent.' 33 The typical mechanism through which third-category nudges operate is switching from an opt-out to an opt-in default, where the nudge succeeds through capitalizing on the agent's inattention. An example of this is automatic enrolment in workplace charitable giving. 34 I do not want to deny that third-category nudges operate in a manner that is normatively problematic. However, in practice, opportunities for the use of thirdcategory nudges in the context of charitable giving are highly limited. In order for a thirdcategory nudge to operate, two things are necessary: there needs to be a transaction with a default option; and it needs to be possible that this default can be altered. In the context of charitable giving, we have reason to doubt that both will obtain. Transactions with a default option are limited in this context, as most attempts to solicit donations do not have this structure. Governments and non-governmental organizations (NGOs) are both in a position to potentially automatically enrol their own employees into workplace giving schemes. Beyond this example, any use of a third-category nudge would require a 31 Hansen and Jespersen, "Nudge and the Manipulation of Choice," 27. 32 Thaler and Sunstein, Nudge, 36. 33 Bovens, "The Ethics of Nudge." 34 Behavioural Insights Team, "Behavioural Insights to Charitable Giving." financial transaction with a third party where a donation was automatically included, such as a default donation to Oxfam included when purchasing a cinema ticket. It is not difficult to imagine instances where an opportunity to donate might be included in such transactions; however, it is difficult to see why a third party would agree for this to employ an opt-in default. This would make their partnership with the charity less obvious, and the potentially deceptive nature of the transaction might engender ill will towards the third party. Cases where third-category nudges are applicable in this context face a further practical constraint, as opt-in defaults on charitable giving have been legally prohibited in some countries, such as the United Kingdom. 35 Therefore, third-category nudges are not a primary concern for the normative assessment of nudges in motivating charitable giving. However, as this limited scope is in part due to legal prohibition, and as such subject to change, there is potential for their use in this context to become more widespread. Therefore, I want to draw attention to a novel feature of utilizing third-category nudges in this context, which raises significant normative concerns that are currently absent from the critical literature. I do so in order to demonstrate that, even where third-category nudges are applicable, we have reason to favour alternative strategies.

Circumventing reflection
Third-category nudges operate through circumventing reflection on an issue, the problematic nature of which has not been addressed in the critical literature. This presents a distinct problem in campaigns to reduce global poverty, due to the normative nature of the deliberation being prevented. In typical cases where third-category nudges are employed as a policy tool, prudential deliberation concerning the relative benefits that different courses of action will have for the deliberator is circumvented in order to improve the situation of the deliberator. The deliberation that the nudging authority has circumvented involves the individual reflecting on their own situation. However, in campaigns to secure donations to reduce international poverty, the deliberation that is circumvented involves the donor reflecting on the situation of the global poor. This reflection may be both instrumentally usefulinspiring further action to address global povertyand may have some intrinsic normative value.

Instrumental value
By circumventing reflection on global poverty, third-category nudges fail to provide any mechanism through which further action to address global poverty can occur. This act of reflection has instrumental value, as it may increase affective concern for the global poor and increase the conscious availability of global poverty, and any corresponding moral demands, to potential donors.
The claim that reflecting on global poverty may inspire sentimental concern for individuals in poverty takes its inspiration from the sentimental cosmopolitan project of fostering global empathy or solidarity. 36 However, here I am concerned with the more modest role that appeals to sentiment can play in motivating action to reduce global 35 Charities Trust, "Best Practice in Payroll Giving." 36 Nussbaum, Upheavals of Thought.
poverty. This finds some support in the thought that reflecting on an issue is more likely to lead to sentimental engagement with this issue than not reflecting on it; however, the strength of this effect is unclear. Whether reflection on the existence of global poverty leads to affective concern for the global poor will also depend to some extent on individual psychology and character. Therefore, it is probably unreasonable to expect the brief reflection on global poverty following a request for charitable giving to result in anything more than a modest affective reaction in most cases. Nevertheless, insofar as the absence of reflection cannot engender any degree of sentimental engagement, this provides a modest reason to favour alternatives.
Even where the degree of reflection that a third-category nudge bypasses is thought to provide insufficient stimulus for significant sentimental engagement, it is important to note that alternative attempts to motivate donations towards reducing global poverty typically attempt to engage sentiments directly. Such appeals to sentiment are a defining feature of charitable fundraising campaigns. Insofar as NGOs and governments have limited resources to devote to fundraising, the use of a third-category nudge will displace an alternative strategy offering a more significant sentimental stimulus.
The social psychology literature on salience also suggests that increased instances of reflecting on an issue cause this issue to be more consciously available. 37 Therefore, insofar as third-category nudges bypass reflection, they serve to reduce the availability of the phenomenon of global poverty to potential donors. Whether conscious availability will inspire action is a different matter, and this link cannot be guaranteed. However, as we are dealing with the reactions of autonomous agents, all that can be expected to hold is the modest conditional claim that an issue being consciously available increases the chances of the agent taking action to address this issue.
In support of this conditional claim, I offer three mechanisms through which the conscious availability of global poverty can serve to motivate remedial action. Firstly, for individuals who are already disposed to aid the global poor, increased instances of reflection may make this demand more salient, increasing the chances that they will act. Secondly, for individuals who are already altruistically disposed, the increased availability of global poverty may increase the chances that they will devote their altruistic attentions to this particular cause. Thirdly, reflection on global poverty may increase the likelihood of disinterested individuals becoming more disposed to aid the global poor. What these mechanisms suggest is that an issue being consciously available to an agent is more likely to inspire action to address the issue than inaction, all other things being equal. Therefore, third-category nudges undermining awareness of the issues they aim to address, coupled with this conditional link between conscious availability and action, provides a modest reason to favour alternative strategies.

Intrinsic value
There is also plausibly some inherent normative value in donors reflecting on the situation of the global poor, deriving directly from the act of refection itself. As Luc Boltanski argues, recognition of their suffering is something we fundamentally owe to 37 Fiske and Morling, "Salience." those in need. 38 I take this claim to be intuitively plausible, and it may account for some of the value placed on apologies and processes of restorative justice. 39 However, it is difficult to offer support for this position beyond this intuitive plausibility, as the normative claim that it makes is so basic. A weaker version of this argument obtains without relying on an appeal to potentially diverging intuitions. That it is normatively preferable for a campaign to foster reflection on the poverty it seeks to alleviate, and as third-category nudges fail to do this, they limit the normative value of the campaign. This argument offers a reason why third-category nudges are prima facie normatively inferior to methods of motivating donations that engender reflection on global poverty.
Part III: second-category nudges I now turn to the complex case of employing second-category nudges to motivate donations towards reducing global poverty. As discussed, the defining feature of second-category nudges in this context is that they utilize non-altruistic sources of motivation to facilitate giving. This is typically achieved through attempts to link charitable donation with high social status. This is exemplified in a recent study in which 'offering to publicize the names of everyone who donated a book to the local library increased donations compared with not offering that incentive.' 40 Secondcategory nudges present a complex case as a means to motivate donations towards reducing global poverty. Here, a good outcome is achieved in the form of the donation; however, this is not primarily motivated by concern for the global poor, or a corresponding normative commitment.

Motives and outcomes
Judith Lichtenberg has recently offered an argument in favour of utilizing nudges to encourage donations towards reducing global poverty. 41 Lichtenberg's argument is of particular relevance in assessing second-category nudges, as it is primarily concerned with justifying why we ought to be more concerned with consequences than with motives in poverty reduction campaigns. Therefore, it is helpful to examine Lichtenberg's argument in some detail.
Lichtenberg's argument in favour of valuing consequences over motives in cases where they conflict is characteristically nuanced. Starting from a commitment to pluralism over sources of value, she argues that criticism of nudged charitable giving rests on a confusion between two separate interests we have in morality: 'the outer realm of consequences and outcomes on the one hand, and the inner realm of character, motives, and reasons on the other.' She goes on to argue that 'we should care, about people's motives intrinsically' but that ultimately 'reducing suffering is more 38 Boltanski,Distant Suffering. 39 Responsibility for suffering might be thought to make these cases disanalogous. However, it is plausible that the affluent bear some responsibility for global poverty. 40 Behavioural Insights Team, "Behavioural Insights to Charitable Giving," 10. 41 Lichtenberg, Distant Strangers. important than the existence of good motives.' 42 It is worth quoting Lichtenberg's argument at length here: Suppose we have to choose between the following alternatives…we can eliminate A's malnutrition, disease, and ignorance without requiring that B act from good motives. Or, on the other hand, B has the best will in the world, exerting herself greatly to alleviate A's suffering, but…A's malnutrition, disease, and ignorance remain. If these are our only alternatives, we should choose the former. 43 Set out in this binary manner, Lichtenberg's account is very persuasive. Favouring motives over consequences when the consequence in question is the reduction of global poverty is prima facie wrong. Here, favouring consequences over motives seems not only justifiable, but even obligatory. However, this clear separation between motives and outcomes is only persuasive, and possible, in the context of a philosopher's hypothetical. This temporal snapshot is not the end of the matter; countless other starving individuals remain after A's 'malnutrition, disease, and ignorance' have been eliminated. Similarly, B's motives remain, altruistic or otherwise. Just as a one-shot prisoner's dilemma cannot tell us very much about social cooperation over time, this single instance cannot tell us very much about how to think about the normative significance of motives.
In order to challenge Lichtenberg's argument, I could defend one of two claims: the stronger claim of the virtue consequentialist position that we exclusively value motives insofar as certain motives typically produce good outcomes 44 ; or the weaker claim that we primarily value motives due to the outcomes they typically produce. This is consistent with the common-sense claim that we also value motives intrinsically to some degree. Due to reasons of space and doubts over strong virtue consequentialism, I will focus on the weaker claim: that we cannot easily separate motives from their longterm effects. Accordingly, it is mistaken to assess the normative permissibility of a motive in a given context by weighing it against the normative value of the consequences it achieves in this situation narrowly construed.
I offer two arguments in support of this claim: firstly, an explanatory argument, which attempts to demonstrate that some of the plausibility of separating motives and consequences derives from conflating backwards-looking and forwards-looking reasoning; and secondly, that separating motives from their longer-term consequences loses further plausibility when assessing nudges, due to the institutional context in which nudges typically operate. Rather than ascribing praise or blame to the actions of individuals, the primary concern here is whether institutions ought to capitalize on, or encourage, the development of certain motives.
The typical situation in which we make normative judgements concerning motives is in ascribing praise or blame to individuals over actions that have already taken place. This usually occurs in a legal context, or when judging everyday conduct. There is typically a single act we are assessing, and we consider motive insofar as it mitigates praise or blame for that act. 45 Assessment takes place after the fact and relevant 42 Ibid., 206-9. 43 Ibid., 210. 44 Driver, Uneasy Virtue. 45 Admittedly, in a legal context, certain motives might provide a reason to detain an individual due to their likely future consequences.
information is limited to a strictly defined temporal context, which ignores, or does not prioritize, long-term consequences. An example of this in a legal context would be the role motive plays in establishing the verdict of attempted murder, or in differentiating between murder and manslaughter. 46 Further support may be found in the law's focus on transgression. Legal assessment of motives is also primarily concerned with cases where good motives mitigate blame for a bad action, rather than where bad motives mitigate praise for a good action. This may lead us to underestimate the normative significance of future consequences in the latter case. To clarify, this is neither the claim that motives have future consequences, nor that consequentialist reasoning faces epistemic difficulties concerning future action. Instead, I am suggesting that the pervasiveness of backwards-looking reasoning in making normative judgements concerning motives may lend unwarranted plausibility to the idea that motives can be assessed in isolation from their potential future consequences. These problems are further compounded when assessing nudge theory due to the institutional context in which nudges typically operate. Nudges to motivate charitable giving take place in an institutional context, either directly through governments or through NGOs. Here, we need to be especially mindful of the link between motives and future consequences for the two following reasons: firstly, any negative consequences will be compounded by aggregative effects, as institutions affect the behaviour of large numbers of people; and secondly, these institutions are significant in shaping 'public morality', or cultural expectations of what is deemed acceptable within a given society. To some extent, people look to governments and NGOs for normative guidance. This encompasses both the perhaps slightly naïve assumption that values being propounded by these institutions increase the credibility of these values, and the expectation that these institutions in particular will act to promote normative values. Therefore, incorporating a motive into these institutions may have a didactic function, appearing to sanction or encourage this motive. This magnifies potential concerns over the longer-term consequences of encouraging non-ideal motives. For example, one might hold that acting from a level of self-interest is a legitimate motive for individuals, but deny that institutions ought to encourage the development of this motive, even to achieve short-term goods. There are two separate concerns here: firstly, that encouraging non-ideal motives may directly lead to negative consequences; and secondly, that sanctioning non-ideal motives may indirectly encourage morally objectionable attitudes. For example, donating to charity may become increasingly viewed as a means to increase social status, with normative reasons for donating, and the recipients themselves, being obscured from the picture. Such attitudes are both intrinsically normatively problematic and, more significantly, may have far-reaching pernicious consequences, such as a general coarsening of public culture and attitudes. 47 I have argued that the separation of motives from their long-term consequences, which underpins Lichtenberg's argument, is implausible. Instead, assessment of motives within a given context needs to take into account the longer-term consequences of cultivating these motives. Therefore, it mischaracterizes employing second-category nudges to motivate charitable giving to portray it as a case of favouring positive 46 Crown Prosecution Service, "Guidelines on Homicide." 47 This is not to deny that concerns over the credibility of both institutions exist in certain sections of the public. outcomes over positive motives. What this fails to take into account is the long-term effects of the prudential motives being employed and comparing these with alternative sources of motivation. Having demonstrated why I take Lichtenberg's account to be inadequate, I now offer an alternative assessment of the normative status of employing second-category nudges in the context of international poverty reduction.

Second-category nudges and motives
Second-category nudges can potentially secure desired changes in behaviour in the instances in which they are employed. However, as they rely on prudential sources of motivation, they do not attempt to alter attitudes. Therefore, second-category nudges faces two significant objections as a means by which to motivate donations towards reducing global poverty: firstly, there may be some intrinsic value in people doing the right thing for the right reasons; and secondly, altering attitudes may have positive consequences in the longer term, leading to further action to reduce global poverty. Although I take the first objection to have significant merit, I confine my focus to the second objection, as the consequences in question are of such normative significance. To support this second claim, I argue that by altering behaviour without altering attitudes, the effectiveness of second-category nudges is compromised as a means to motivate effective action to address global poverty.
Before doing so, it is necessary to answer two immediate objections that may occur to the reader at this point. The first objection is that this argument relies on the assumption that a single source of motivation is typically responsible for a given act, which presents an overly simplistic picture of human psychology and motivation. That this is overly simplistic is correct. However, I am operating on the assumption that some degree of altruistic motivation is present, but one that is insufficient to motivate a donation without the additional prudential motivation provided by the nudge. Charitable campaigns that attempt to alter attitudes, by fostering sentimental or moral concern, can potentially develop these more altruistic sources of motivation. However, in instances where secondcategory nudges are employed, this function is absent.
The second objection is that by criticizing the use of second-category nudges that aim to alter actions by securing donations without a corresponding change in attitudes, I am letting the perfect be the enemy of the good, operating from an idealized baseline where donations deriving from largely non-altruistic motives are wrongly criticized in comparison to donations stemming from altruistic motives. Here, it may be objected that I am failing to take into account that the donation in the former case would not take place without the nudge. 48 However, this objection implausibly assumes a baseline of inaction, which takes all charitable giving as supererogatory. Even if we grant that a baseline of inaction is appropriate in the moral assessment of individual donors, this does not apply when assessing the use of second-category nudges by NGOs and governments as a means to facilitate charitable donations. In these institutional cases, the adoption of nudge strategies is a zero-sum issue, as NGOs and governments only have finite resources. Therefore, employing second-category nudges will necessarily be at the expense of alternative 48 Lichtenberg, Distant Strangers, 243.
approaches. As such, the correct baseline for assessing second-category nudges in these circumstances is in comparison to alternative methods of motivating donations.

Effectiveness
Fundraising campaigns to reduce global poverty can be effective in two different ways. In the more obvious sense, effectiveness can be understood as a function of the amount of money raised and the success of the scheme to which this is directed. However, fundraising campaigns can also be effective by motivating changes in attitudes: encouraging normative reflection on global poverty or increasing levels of affective concern for the global poor. These changes in attitude may in turn motivate further action to reduce global poverty, such as future donations, lifestyle changes or commitments to political action. Second-category nudges that rely on prudential sources of motivation, aiming to alter actions not attitudes, cannot contribute towards this second type of effectiveness.
In contrast, more traditional methods of motivating donations, which aim to provoke moral or sentimental concern for the global poor, may lead to the development, or strengthening, of broader motivating commitments to reducing global poverty. There is significant empirical support for the claim that the traditional fundraising strategies employed by NGOs working to reduce global poverty are effective in altering normative attitudes towards global poverty 49 and increasing levels of affective concern for the global poor. 50 Social psychological studies also lend weight to the further claim that altering normative attitudes 51 and affective responses 52 towards instances of global injustice leads individuals to engage in action to address these injustices. Broader research in the psychological sub-field of social influence further supports the general claim that altering normative attitudes is necessary to achieve lasting behaviour change. 53 It might be objected that second-category nudges are sufficiently effective at achieving the desired behavioural changes in the instances in which they are employed so as to outweigh any advantages traditional interventions can offer via this second type of effectiveness. The effectiveness of second-category nudges in motivating donations in the specific instances in which they are employed will depend on the specific nudge in question, and likely differ between agents. However, it is by no means clear that prudential motives offer a more effective means of motivating charitable donations in the short term than appeals to moral or sentimental concern. Studies by Heyman and Ariely 54 suggest that appealing to incentives undermines the effectiveness of fundraising campaigns, as this encourages individuals to perceive the transaction in market terms, causing potential donors to focus on instrumental losses and gains and to ignore the normative element. The same result is replicated in separate studies by Kamenica, Kosters and van Heijden (2012). 55 Although answering this empirical question definitively is beyond the scope of this paper, the three following reasons suggest that the burden of proof is on the advocate of second-category nudges. Firstly, in order for a second-category nudge to be preferable to an alternative means of motivating a donation, its effectiveness in motivating a donation would need to be sufficiently great so as to outweigh the combined benefits of the alternative approach, both as a means to secure donations and as a means to alter attitudes. Absent strong empirical evidence in favour of the effectiveness of secondcategory nudges, this supports a presumption in favour of alternatives.
Secondly, the empirical literature on poverty reduction suggests that many charitable schemes prove ineffective, or even counterproductive. 56 Although this is not typical, the outcome is by no means unusual. Determining the effectiveness of a given charitable scheme also faces epistemic limitations and is vulnerable to brute luck. Employing secondcategory nudges limits possible positive consequences to this specific context. In contrast, fundraising techniques that attempt to alter normative attitudes, or increase concern for the global poor can potentially contribute towards reducing global poverty, even when the specific campaign they are part of proves ineffective. Therefore, modesty over our ability to determine the effectiveness of a given charitable scheme in advance offers a further reason to place the burden of proof on the advocate of second-category nudges.
Thirdly, charitable contributions alone are unlikely to provide a comprehensive solution to global poverty. Alongside financial support for the work of NGOs, political reform of global and national institutions is likely to be required. Recognition of this need for a combined approach is commonplace in both international political theory and international development. 57 Comprehensive solutions are also likely to require changes in lifestyle by individuals in developed countries. As second-category nudges aim to alter behaviour in a given instance without provoking a corresponding change in attitudes, they are unlikely to contribute towards motivating comprehensive solutions.

Second-category nudges and changes in attitude
I have argued that although second-category nudges may secure desired behavioural changes in the short term, these are not supported by corresponding changes in normative attitudes. However, Luc Bovens has suggested that by altering an individual's behaviour, second-category nudges may lead to changes in attitudes over time. 58 I now examine the plausibility of this claim. First, I assess the three mechanisms that Bovens outlines through which nudges may prompt attitude changes in individuals who are subject to nudges. I then consider a final possibility: that second-category nudges may alter attitudes indirectly through prompting reflection on the part of non-donors. I argue that, on either method, the role of second-category nudges in altering attitudes by altering behaviour is likely to be minimal. 55 Kosters and van der Heijden, "From Mechanism to Virtue."; Kamenica, "Behavioral Economics and Psychology of Incentives." 56 Easterly, The White Man 's Burden;and Lichtenberg,Distant Strangers, Pogge, World Poverty and Human Rights, 36; Singer, The Life You Can Save, 20; and Oxfam, "What We Work On." 58 Bovens, "The Ethics of Nudge."

Direct mechanisms
Bovens offers three mechanisms by which, through altering behaviour, nudges might lead to corresponding changes in attitudes: (i) changes in behaviour may serve an 'educative function'; (ii) by performing certain actions, 'feelings may simply shift'; and (iii) 'one may come to self-identify as a person who acts [in a certain manner] on grounds of cognitive dissonance.' 59 I will address each in turn.
The educative function of actions. Unlike health nudges, which are the focus of Bovens' account, the experiential content of the action performed by donating to charity is fairly minimal. Plausibly, an agent learns more about running by going running than they learn about charity by setting up a direct debit or physically handing over money. More significantly, the act of donating has little educative function concerning the charitable cause itself. However, in altering patterns of behaviour, the nudge may inspire reflection on the issues the charity seeks to address, and in doing so potentially inspire a genuine moral commitment. As John et al. put this, 'a nudge may lead to a demand to think.' 60 For example, a second-category nudge leading an individual to donate to charity may result in this individual reflecting on the value of the work the charity does, leading to the development of a moral commitment.
However, in this instance, the educative function is primarily performed by the attempt to solicit a donation in the form of media developed by the charity or an encounter with a fundraiser. Accordingly, a nudged act of charitable giving does not offer a greater opportunity for reflection on the charitable cause than an attempt to solicit a donation that fails. The prudential motivation provided by the second-category nudge may help direct the attention of the nudged individual to this information, but unlike nudges aimed at prompting changes in lifestyle, the action itself has little educative function.
Shifts in feelings. The experience of donating to charity may have little educative value, but it does have potential positive content. On the 'warm glow theory of altruism', the positive feelings resulting from performing altruistic acts serve to make altruism attractive. 61 Therefore, a second-category nudge may have the potential to inspire further commitments, as these 'hedonistic' benefits might not be obvious prior to giving.
However, the simplicity of the mechanism suggested by this argument does not offer a plausible basis for the development of the complex phenomenon of a normative commitment to reducing global poverty. The hedonistic benefits of giving provide a reason to think that additional donations may follow when the nudge is no longer in place, and potential support for the thought that the nudge may inspire further acts of charity. But it presents an implausibly simplistic and deterministic picture of human psychology in suggesting that the development of complex normative beliefs concerning global poverty can be primarily attributed to this mechanism. This simplistic picture relies on a very strong reading of the warm glow theory of altruism, rather than a more 59 Ibid., 214. 60 John et al.,"Nudge Nudge,Think Think," 369. 61 Andreoni, "Altruism and Donations to Public Goods." nuanced account. Experiments by Batson et al. that attempted to isolate the desire to behave altruistically from causal responsibly for the altruistic act have cast serious doubt on the plausibility of a strong reading of the warm glow theory. In these experiments, altruistic choices did not decrease, despite the warm glow theory suggesting otherwise. 62 This does not rule out the more sophisticated psychological egoism of the 'empathic joy hypothesis', which suggests that altruism results from the hedonistic benefit we derive from the pleasure of others. 63 However, the empathic joy hypothesis suggests that hedonistic benefits result regardless of an agent's own causal responsibility for another's pleasure. Therefore, it is unclear that changes in attitudes can be directly attributed to the experiential content of the donation itself, and that the nudge is primarily responsible for altering attitudes.
Cognitive dissonance. Bovens' most promising suggestion as to how second-category nudges might lead to changes in attitudes is through the phenomenon of cognitive dissonance. 'Cognitive dissonance' is used here to refer to the unpleasant feeling when our beliefs and actions do not cohere. The significance of this for second-category nudges is that, in an attempt to achieve consistency, individuals often alter their beliefs rather than their actions. As Stoker notes, 'psychologists suggest that people seek consistency between their beliefs and their behavior. However, when beliefs and behavior clash, we frequently alter our beliefs instead of adjusting our behavior.' 64 Therefore, where a second-category nudge leads an individual to donate towards reducing global poverty, they may come to identify with this end in order to achieve cognitive consistency. This suggests a potential mechanism by which nudged charitable giving could lead to the development of a corresponding normative commitment.
However, instances where this mechanism results in a change in attitudes will be limited for the following two reasons: firstly, some people can live with a high degree of inconsistency between their attitudes and actionspolitical philosophers are likely more troubled by dissonance than other people; and secondly, it is an assumption of this argument that it offers a fairly novel method of belief formation, and instead, actions typically reflect attitudes. Despite these qualifications, this strikes me as a plausible account of how a second-category nudge might play a limited role in altering attitudes. However, this argument offers no reason to think that any changes in attitude that occur will be sufficiently robust to motivate further actions. On this model, the dissonance is solved by the thought that one cares enough about global poverty to make the donation resulting from the nudge. This brings attitudes into line with actions rather than motivating further action. Therefore, this mechanism offers little reason to think that any changes in belief that result will be significant enough to motivate further action towards reducing global poverty.

Indirect mechanisms
However, the role of second-category nudges in altering attitudes may be more indirect. Rather than altering the attitudes of nudged individuals, second-category nudges may 62 Stich et al.,"Altruism," 197. 63 Batson, The Altruism Question. 64 Stoker, "The Politics of Nudge," 227.
alter the attitudes of non-donors. Through increasing instances of charitable giving, second-category nudges may lead to increased instances of non-donors reflecting on global poverty, which may in turn lead to normative commitments to donate. In the final section of this paper, I suggest two tentative mechanisms through which this process could occur, and offer some reasons for thinking that their role will be limited. Although neither mechanism accords second-category nudges a central role in attitude change, they may suggest a modest role for second-category nudges in this process.
Valuing charity. Where second-category nudges operate by capitalizing on a desire for esteem or social status, this relies on the existence of a culture in which making charitable contributions is socially valued. Here, the use of second-category nudges may cause non-donors to reflect on why charitable giving is socially valued, which may in turn lead to a normative commitment to donate. 65 The first concern here is the highly abstract nature of the deliberation, which sits uneasily with the assumption of bounded rationality underpinning the nudge approach. This is less of a concern for deliberation prompted by traditional fundraising campaigns, which operate at a more affective level, employing emotive portrayals of distant others to engage sentimental concern. Secondly, this mechanism relies on donating to charity in general being socially valued, rather than the specific concern of addressing global poverty. 66 Charity being held in high regard may encourage reflective commitments to donating, but this offers no reason to motivate commitments to addressing global poverty more broadly, such as through political action or lifestyle changes. Finally, this mechanism operates on the basis that donating to charity is socially valued, not the recognition of an obligation to donate. Where social norms reflect the belief that donations towards global poverty are supererogatory, any moral commitments arising as a result are likely to be relatively weak in nature.
Social pressure. Where second-category nudges lead to increased instances of charitable giving, this behaviour may alter perceptions of the norms that are operative in society, with nudged charitable giving being taken by non-donors to reflect a corresponding normative belief on the part of donors. Here, social pressure may provide an additional incentive for non-donors to reconsider their own views, potentially leading to a change in normative attitudes. This is supported by recent work in moral psychology on social norms suggesting 'that injunctive normswhat people ought to do -… can be overwhelmed by descriptive normswhat people are actually doing.' 67 However, the role for second-category nudges in altering perceptions of social norms is likely to be relatively modest for the following reasons. Firstly, many acts of charity are not publically observable, or are minimally observable; therefore, it is hard to see how they can play a significant part in altering perceptions of social norms. 68 Secondly, this mechanism is too localized to play a substantial role in altering social norms.
Available information would differ too substantially from person to person to plausibly account for the development of norms that existed beyond the level of the individual. Finally, according a primary causal role to deductive inferences from personally observed behaviour offers an implausibly atomistic account of social norms. This ignores the social nature of these norms and fails to take into account the role of the media, and communication more generally, in their development and propagation.
I have outlined two tentative mechanisms through which second-category nudges might play a very limited role in attitude change in the long term. At present, I do not think these mechanisms offer a convincing rebuttal of the normative objections facing the use of second-category nudges to motivate donations towards reducing global poverty. However, in order to draw more definitive normative conclusions, further empirical research on these mechanisms is needed. I hope that the arguments made within this paper highlight the need for future collaborative research in this area between political philosophers and social scientists, and can offer a useful starting point.

Conclusion
This paper has examined the normative status of the use of nudges by NGOs and governments to motivate charitable donations towards reducing global poverty. I began by arguing that the definitions of nudges provided in the literature are unsuitable for this context, failing to distinguish between very different methods of behavioural modification. To address this, I offered a three-part typology of nudges, based on how a nudge interacts with an agent's motivational structure. I then assessed the use of nudges in each category as a means to motivate donations towards reducing global poverty, drawing the following three conclusions: (i) the use of first-category nudges faces no significant normative objections in this context; (ii) the use of third-category nudges in the context of charitable giving faces distinct normative concerns, as they circumvent reflection on the poverty they seek to addresshowever, scope for their employment in this context is limited, as they face both legal and practical obstacles; and (iii) second-category nudges, which aim to secure charitable donations by capitalizing on prudential sources of motivation, present a complex case, but where possible, NGOs and governments ought to favour alternative methods of encouraging donations. To support this conclusion, I advanced the following two claims: firstly, that it oversimplifies the complexity of using second-category nudges to motivate charitable giving to portray this as a case of choosing good outcomes over good motivesinstead, as second-category nudges aim to achieve changes in behaviour without altering attitudes, the positive consequences of their employment in charitable campaigns are limited; and secondly, although it has been suggested that, by altering behaviour-, second-category nudges might themselves engender a change in normative attitudes over time, their causal role in this process appears to be limited. However, there is a need for further collaborative research in this area between political philosophers and social scientists. In sum, nudges provide no easy substitute for the complex and vital task of altering attitudes towards global poverty.