Mitigating the costs of departure. Brain drain, disadvantage and fair burden-sharing

ABSTRACT In this article we aim to assess how the negative effects of brain drain can be mitigated in a fair way. We particularly focus on the policies of extraterritorial taxation and temporary compulsory service for highly skilled migrants in developing countries, which are most thoroughly defended by Gillian Brock. We argue that while Brock is right in pressing for policies seeking to combat the damaging effects of brain drain, she fails to properly characterize the complex strands of disadvantage that run through this phenomenon, placing an unfair redistributive burden on highly skilled migrants. By contrast, we maintain that any fair distribution of such burdens can only flow from a comprehensive account of existing comparative disadvantages, without regard for migratory status. The resulting policy implications are that a considerable part of the tax burden required for mitigating brain drain should be borne by citizens of developed countries, and that compulsory service in developing countries should be rejected.


Introduction
The migration of highly skilled individuals from developing to developed countries (commonly known as brain drain 1 ) came to be recognized as a discernible trend of an increasingly globalized world in the 1960s. The interest of economists, political scientists and, more recently, political philosophers for the issues raised by brain drain and the possible public responses to it only grew in the decades since. The foremost early contribution to the debate on brain drain is Bhagwati's proposal to tax the income earned by departing citizens in their host countries, in order to curb the negative effects on those left behind in source countries (Bhagwati 1972(Bhagwati , 1979Bhagwati and Dellalfar 1973). Although the initial formulation of the Bhagwati tax scheme relies on some key normative claims, the most influential work in analytical political philosophy dealing with policies which aim to counteract this form of emigration is Brock and Blake`s (2015) book-length treatment of the subject. Here and in subsequent articles, Brock (2016aBrock ( , 2016bBrock ( , 2016c defends the right of a state that is legitimate, responsible, but poor to constrain the migration of highly skilled citizens, if their skills were developed with the help of public funds. The main thrust of her argument resides in the idea that highly skilled emigrants from developing states impose, in certain conditions, an unfair burden on the citizens they leave behind when emigrating.
Our aim in the article is to offer a comprehensive account of the requirements of fairness in respect to brain drain. In doing this, we maintain that while Brock's argument has considerable appeal, it fails to properly characterize the complex strands of disadvantage that run through the brain drain phenomenon. As a result, the conception of fairness which she uses is only able to ground an impoverished view of the agents who should be held responsible for the mitigation of brain drain-induced disadvantages. On our alternative account, fairness considerations entail different policies than those advocated by Brock. Thus, while we do not purport to offer an allthings-considered policy answer to the full array of problems raised by brain drain, we try to elucidate the requirements of one, albeit crucial, moral value that bears on this phenomenon.
The paper is structured as follows: in the next section we discuss some background features of the brain drain debate and we outline Brock's argument for migration constraints. Subsequently, we examine the structure of her argument, claiming that even if we grant that highly skilled workers are causally responsible for disadvantaging individuals in source countries, it does not necessarily follow that they are the sole (or even primary) bearers of moral duties to compensate them. Finally, we outline a Dworkinian conception of how to identify disadvantages, recently articulated by Parr (2018), and we apply it to the case of brain drain, concluding that the asymmetrical treatment of highly skilled workers in developing and developed countries advocated by Brock is unjustifiable on grounds of fairness.

Brain drain, disadvantage, and migration constraints
The dynamic of skilled migration from developing to developed countries is sometimes captured by the phrase 'the perverse subsidy' (Labonté, Packer, and Klassen 2006;Kollar and Buyx 2013). The basic point here is simple and can be exemplified using the main domain in which discussions about brain drain arise: healthcare services. Poor countries in dire need of human resources for the provision of healthcare invest public money in the training of specialized personnel, but the citizens of these countries are often unable to enjoy the benefits, due to the departure of the latter. Rather, what happens is that developed countries end up benefiting from the supply of a large share of healthcare personnel, while the burdens of their training are supported by poor countries.
This seems at least prima facie morally problematic. It may ultimately turn out that the emigration of highly skilled individuals 2 brings about some other benefits which outweigh the burdens suffered by the developing country. It may also turn out that individuals have certain rights that prohibit any interventions seeking to counteract this perverse subsidy. But the motivating force of the inquiry remains: there is something morally wrong with the global poor paying -in a certain sense -for the benefits of the global rich. This raises the question of what types of policies may be pursued by either the host or the source country in order to respond to this phenomenon? Two distinctions are of particular importance here. The first one delineates public measures which seek to constrain immigration, and are applied in host countries, from public measures which seek to constrain emigration, and are applied in source countries. The second one delineates measures which only seek to instantiate a certain structure of incentives from measures which involve coercion (Dumitru 2012). This leaves us with four possible responses that can be pursued either by host or source countries: (1) the implementation of policies which seek to disincentivise the immigration of the highly skilled; (2) the implementation of coercive immigration restrictions targeting highly skilled workers coming from developing countries; (3) the implementation of policies which change the structure of incentives of the highly skilled by developing countries; and (4) the implementation of coercive emigration restrictions by developing countries. In what follows, we offer an overview of the latter three types of policies, which are the main objects of contention in the brain drain literature. 3 To be clear, in the case of (2), what is at stake is the justification of immigration restrictions on the grounds that they are required by duties of global justice owed by developed states. 4 Oberman (2013) offers the most thorough exploration of this argument. He begins by arguing that certain conditions must be fulfilled in order for rich countries to legitimately impose these restrictions. Besides the establishment of a clear duty from the part of skilled workers to provide assistance to those in need in their countries, one crucial aspect of this argument is that the developed state may impose these 'counter brain-drain immigration restrictions [. . .] if it has fulfilled its own duty of assistance to the global poor' (Oberman 2013, 453). Given the relevant empirical facts when it comes to the extent to which contemporary developed states fulfil their duties of assistance, Oberman concludes that 'cases in which immigration restrictions can be justified on brain drain grounds are thus likely to be those in which these two factors are conjoined: they will be cases involving skilled workers who are relatively well off yet live in states which suffer from problems that block effective assistance from outside' (Oberman 2013, 454). Thus, only a small number of exclusions can be justified on grounds pertaining to the harmful effects that brain drain has on the source countries.
Nevertheless, arguments seeking to establish the legitimacy of immigration restrictions on brain drain-related grounds are scarce. Especially after the publication of Brock and Blake's (2015) influential book, the main focus in the literature has been on permissible emigration restrictions. As Dumitru (2012) notes, most of the proposed policies can be grouped along two cross-cutting distinctions. On the one hand, there is the category of policies which seek to incentivize citizens to stay in a country by modifying the existing structure of benefits in those professional fields which may be vulnerable to a growing international demand for labour. The second category of policies doesn't seek to incentivize people to stay, but to compensate those left behind after a high number of individuals decide to move to other countries. One early proposal that justifies a policy largely along these lines is the Bhagwati tax scheme, mentioned above. The main idea behind it is that the state should tax the income of emigrant workers in the host country for a certain period of time, usually until they obtain citizenship rights. 5 With this clarification of policy alternatives on hand, we can place Brock's view in the proper context. In order to begin justifying these policies as permissible for a poor, but responsible and legitimate state, some empirical conditions must hold: (1) there must be evidence that 'skilled citizens can provide important services for which there are severe shortages, and their departures considerably undermine efforts to meet citizens' need' (Brock and Blake 2015, 101); (2) the government has invested in the training of the skilled workers; and (3) the needs of the other citizens can only be met through the continuing presence of the highly skilled in that particular country. 6 If these conditions hold, Brock argues, then some citizens acquire a particular set of duties, qua highly skilled workers. This is because we can tie the emigration of these workers to the deprivation of fellow citizens through what Brock calls the connective grounds for assigning remedial responsibility (Brock and Blake 2015, 69). Two of them are of particular importance for the value of fairness. First, those who emigrate stand in a causal or contributory connection with the deprivation suffered by those left behind, as their departure foreseeably brings about a situation in which the needs of the other citizens are not met. Second, those who emigrate have benefited from the public funds invested in their education, but have failed to reciprocate to those who shouldered the tax burden for the development of their skills. 7 If this account of moral responsibility is correct and the empirical conditions discussed above hold, then -Brock contendsdeveloping states should pursue policies of (1) compulsory public service after graduation and (2) taxation of highly skilled emigrants for a period of time after departure (or at the time of exit).
Brock's argument can be summarized, then, in the following form, which we term the Argument from Disadvantage for Emigration Constraints: P1: Highly skilled migrants from developing countries disadvantage the citizens who remain behind due to the fact that their departure increases their vulnerability. P2: Highly skilled migrants from developing countries disadvantage the citizens who remain behind due to the fact that while they have benefited from taxpayer-funded governmental investment in their education, they fail to reciprocate by benefiting these taxpayers in return. P3: Highly skilled migrants from developing countries have a duty to mitigate the disadvantages created by their departure. P4: The duties to mitigate disadvantages created by the departure of highly skilled migrants can be enforced through imposing constraining measures on them, such as extraterritorial taxation by source countries after departure, and temporary compulsory service after graduation in source countries.
Highly skilled migrants should be subjected to constraining measures, such as extraterritorial taxation by source countries after departure and temporary compulsory service after graduation in source countries.
One important point of contention is the status of the empirical premises employed in Brock's argument. The empirical literature on brain drain identifies both negative and positive effects experienced by source countries as a result of the migration of highly skilled workers. For example, an increase in the flow of remittances is a significant positive aspect, as is the influence it can have on the transfer of technology and the formation of human capital. Among the negative effects, the most important one is the impact suffered by those left behind due to the deterioration of public services provision. So should we really be concerned with Brock's type of argument if, on balance, the emigration of skilled workers from developing countries actually has net beneficial effects?
Two responses support the continuing interest of normative inquiries on the brain drain phenomenon, even if the impact is likely to have net beneficial effects. One response is that not all effects of highly skilled emigration are on par from a moral standpoint -or, to be more exact, not all effects have the same moral weight. The idea here is that meeting urgent needs weighs more heavily from a moral standpoint that the different long-term, incremental improvements which brain drain may bring to a particular society. If this is correct, then the fact that the flow of remittance or technological transfers would improve is not as important as the fact that, for example, the emigration of skilled medical workers and the resulting lack of personnel puts some 7 The third and final connective ground proposed by Brock is that of capacity, which refers to the agents who would be most effective and most able to bear the costs associated with rectifying the problematic situation. We do not analyse this feature further because: (1) it is not (at least as directly) related to the value of fairness, and (2) not only does it not damage our case against Brock and in favour of the policy proposals outlined in the fourth section, but considerably strengthens it since developed states are -on both grounds of effectiveness and ability to bear costsbetter positioned to rectify brain drain-related negative effects than developing ones.
of the citizens of a country in a particularly troubling situation of vulnerability. One objection may be that the distinction between short-term harmful effects and long-term beneficial effects doesn't hold because one positive component of emigration, i.e. remittances, might already bring benefits shortly after emigration. This brings us to the second (and decisive) response to the challenge that the existence of net beneficial effects makes Brock's argument a non-starter. Specifically, the fact that there are net positive effects and that remittances improve the conditions of those who receive them even in the short-term doesn't automatically imply that the distribution of these benefits in the developing country is also fair. As not all citizens in developing countries would receive the benefits of remittances, but all would experience deteriorated public services as a result of brain drain, it's still justified to inquire into what policies could fairly mitigate brain drain even if the benefits are positive on aggregate. The normative premises of Brock's argument, with which we are fundamentally concerned here, have been challenged on a number of grounds. Blake (2016) contends that her 'proposals unfairly distribute the burden of making the world a just place, by placing a disproportionate burden upon the skilled and educated residents of developing societies'. Specifically, he maintains that developed countries should contribute to the mitigation of brain drain primarily by stepping up their investment in the education of highly skilled workers, so that they no longer need to rely on importing highly skilled labour force from developing countries (Brock and Blake 2015, 223-224). The charge underlined above has been levelled at Brock by other authors as well, who maintain that developed countries are largely left off the hook by her proposal. Reglitz (2016, 74) favours Blake's suggestion, also adding that developed states have the duty to reimburse developing states for the economic resources that they invested in educating highly skilled workers. And Ypi (2016, 42) claims that to the extent that we accept a policy of temporary compulsory service, the costs which these contractual arrangements bring about should be borne by host institutions in rich societies, rather than by developing countries. While these alternative proposals are on the right track, we believe that they cut insufficiently deep and ultimately allow for too much residual unfairness to arise, since they do not thoroughly address the question of duty-bearers for brain drain effects in a systematic manner. As Ypi observes, though, there is a need to understand how 'the burdens between migrants, citizens of host states and citizens of source states should be distributed fairly' (Ypi 2016, 43). This is the task to which we now turn.

Does causal responsibility necessarily entail liability?
Let us first note that Brock's proposal has considerable intuitive appeal on any mainstream theory of distributive justice. On egalitarian grounds, Brock's policies are attractive because unrestricted migration would presumably exacerbate existing inequalities, as highly skilled migrants would become even better-off, while citizens who remain in source countries would become even worse-off. On prioritarian grounds, Brock's policies are attractive because they would presumably benefit the least advantaged, to a large enough extent so as to outweigh the costs imposed on highly skilled citizens who wish to migrate. On sufficientarian grounds, Brock's policies are attractive because they would presumably increase the likelihood that everyone reaches the threshold of advantage deemed adequate. It's plausible to think that the initial appeal of the Argument from Disadvantage is traceable to the fact that it correctly identifies the group of individuals whose prioritization in the distribution of benefits is most pressing, i.e. the most disadvantaged. Indeed, it would be accurate to say that an overwhelming amount of attention in the distributive justice literature has focused on the recipients of benefits and comparably less has been devoted to the bearers of distributive duties. 8 In the literature on brain drain, by contrast, the problem of duty-bearers is at the very heart of the discussion.
Thus, Brock's proposal should be assessed not only in light of the benefits brought by a presumed alleviation of the brain drain phenomenon, but also in light of its treatment of the burdens required for this alleviation. It is in respect to this latter question that Brock's position seems most vulnerable. To a large extent, the Argument from Disadvantage draws its normative force from the fact that if we accept P1 and P2, then the duty-assigning premise P3 automatically follows. But this is not necessarily the case. We can understand why by appealing to a distinction between two types of responsibility: the causal account and the liability account. According to Dworkin (2011, 103), 'a person is causally responsible for an event, we say, if some act of his figures (or figures substantially) in the best causal explanation of that event' and 'someone has liability responsibility for an event when he is required to repair, compensate for, or absorb any damage flowing from the event'. Thus, even if we agree that highly skilled migrants are causally responsible for compounding the disadvantage of source country residents, it does not mean that they automatically have liability responsibility for it, or at the very least it does not mean that they are necessarily the sole agents bearing this type of responsibility. Otherwise put, it is not necessarily the case that because agent A is causally responsible for creating a disadvantage for agent B, agent A should be solely (or even partly) responsible for compensating agent B. We illustrate this idea with a couple of examples.
Shopkeeper. Bako used to own and run the only musical shop in a small Nigerian town. Chinue has recently moved into the town and has opened up her own, better equipped and more reasonably priced, musical shop fairly close to Bako, drawing many of his former customers to her shop. In a few months, Bako is no longer able to make any profit and shuts down his business.
In Shopkeeper, Chinue's actions have the effect of making Bako worse-off. Had she not opened up a second musical shop in the vicinity, he would have had many more clients. Instead, he is forced to shut down his business, thereby incurring considerable disadvantage and becoming much more vulnerable, as a direct result of Chinue's actions. Structurally, this situation is similar to the one that brain drain creates, which we synthesize in P1. 9 Private security. Adeline is a home owner in a suburban neighbourhood. Due to a recent spike in burglaries, she purchases a moderately expensive security system for her place. A few days later, her neighbours decide to pool resources together and hire a private 8 The recent development of limitarianism as a partial theory of justice by Ingrid Robeyns (2017), which mainly seeks to map out the appropriate duty-bearers in redistributive processes (Volacu and Dumitru 2019, 250) could signal a shift in this regard, but there is still quite a long way to go before the problem of duty-bearers becomes more central in this literature. 9 And which is captured by Brock's Causal or contributory connection ground for assigning remedial responsibility. security company to monitor the streets in order to deter further burglaries. Adeline declines to contribute to the collective venture.
In Private security, Adeline benefits from the increased security which hiring the private company produces. However, she enjoys the public good free of charge, having contributed nothing in exchange for the benefits obtained. As a result, each neighbour has to pay an extra amount of money for the services rendered by the company. Furthermore, Adeline offers no equivalent reciprocal contribution to any other public good. Her neighbours are therefore disadvantaged in relation to Adeline, not because they are being made worse-off by her actions, but rather because she enjoys a benefit funded from their taxes without contributing something in return. Structurally, this situation is similar to the one that brain drain creates, which we synthesize in P2. 10 In Shopkeeper and Private security Chinue and Adeline are causally responsible for the disadvantage of others. But it does not follow that they also have liability responsibility and should appropriately compensate those agents who are disadvantaged. Note that we are not drawing on these cases in order to build an analogical argument, with the effect of undercutting the case for emigration restrictions described in the previous section. Rather, we only use them to defend the much more limited claim that P3 does not automatically follow from either P1 or P2. If harming third parties involuntarily through some of our actions (as in Shopkeeper) and if benefitting from a collective good without reciprocating (as in Private security) does not always generate a duty to mitigate the disadvantages created, we should therefore inquire when exactly those who are causally responsible for a state of affairs also have liability responsibility for the mitigation of disadvantages generated through their actions. Any reasonable answer to this question is bound to require some plausible background conception of how disadvantages are to be identified, from which a mapping of distributive duties should flow. In the following section we first outline such a conception, and then apply it to the problem of brain drain.

Identifying disadvantage in the case of brain drain
As Parr (2018) points out, there are two main ways of assessing whether comparative disadvantages exist. The first one, which has been standardly deployed in political philosophy, consists in the metric test, which can roughly be described as follows: 'we should treat an individual as disadvantaged if and only if her opportunities afford her an objectively lower level of well-being, a lower level of welfare, or fewer social primary goods' (Parr 2018, 306). In order to apply the metric test we first need to identify the most plausible currency of justice and then see how much of that currency individuals have compared to one another. But this account raises several problems, the most salient of which is that individuals can hold different views on whether opportunities are valuable for them (and the extent to which they are valuable), which will make any external assessments unjustifiable to these individuals since they are bound to wrongly treat at least some of them as (dis)advantaged (Parr 2018, 308-309).
An alternative way of proceeding is by appealing to the envy test proposed by Dworkin (1981Dworkin ( , 285-287, 2011. In its original version, Dworkin famously uses it as an instrument for determining when a distribution is fully egalitarian from a resourcist perspective. But it can also be reconstructed in a more general form as follows: 'we should treat an individual as disadvantaged if and only if that individual envies another's opportunities' (Parr 2018, 306). On this view, we are not strictly concerned with resource equality, but with any form of comparative disadvantage which may arise. Parr (2018, 311) further distinguishes between two possible interpretations of the envy test, previously unaccounted for by either Dworkin or his subsequent defenders: the choice-worthiness account, on which envying another's opportunities means that you would prefer to have her opportunities rather than your own, and the well-being value account, on which envying another's opportunities means that you regard her as having more or better opportunities for well-being than your own (Parr 2018, 311).
Parr begins his defence of the latter account by testing our intuitions in a case involving difficult parenting. But we can also appeal to emigration to illuminate the appeal of the well-being value account.
Engineer. Daphne is an American engineer. She is highly skilled in her profession and works at a big tech company in Miami, where she receives a high salary. Emmanuel is a Haitian engineer. He is just as skilled in that profession as Daphne, but works for a very low salary in Port-au-Prince. He lives close to his parents and other siblings, whom he frequently takes care of and whom he is very fond of.
In Engineer, deploying the two accounts of the envy test listed above can lead to diverging answers. On the choice-worthiness account, Emmanuel is not disadvantaged, because if it would be hypothetically possible to switch opportunities with Daphne, he would not do so, for example because he has a special relation with his siblings which makes him unwilling to emigrate. On the well-being value account, Emmanuel is disadvantaged, because he does see his own opportunities for well-being as less valuable than those of Daphne. If our intuitive assessment in Engineer is that there is some strand of disadvantage identifiable in this example, we have a good initial reason to favour the well-being value test, as Parr does. 11 So what does applying the well-being value account of the envy test to the problem of brain drain tell us about the disadvantages generated? 12 There are several relevant comparisons. The first one is between those in developing and developed countries, regardless of their skill level. Assume that no constraints aiming to mitigate brain drain are in place. Applying the envy test uncovers one important way in which citizens in developing countries are disadvantaged: such a citizen could rightfully complain that the opportunities for well-being associated with the appropriate functioning of public services are much more diminished for them, compared to those of citizens in 11 For a more robust defence of the well-being value account see Parr (2018, 312-313). 12 Surely, there are multiple accounts on which the envy test might be failed either for the average citizen of a developing country and for the low skilled worker of a developed country. While a comprehensive assessment of injustice should factor in all such disadvantages (e.g. in opportunities for gaining material wealth, opportunities for job satisfaction prospects, opportunities for gaining social status through occupational choice etc.), we cannot pursue such a broad exercise here. Rather, we will strictly be concerned with the disadvantages most intimately related to the brain drain problem and to those which would be brought about by measures aimed at mitigating it.
developed countries. This is because highly skilled workers have incentives to depart from developing countries in sufficiently large numbers so as to deplete the human resources pool required for the adequate functioning of many public services. Of course, this is merely a different characterization of the core concern we have in regard to the brain drain phenomenon, and does not add anything substantive so far. But there are two relevant upshots to this characterization. First, it shows that the central problem of brain drain is also captured by an interpersonal account of disadvantage. Second, and more importantly, it shows that highly skilled workers can only partly escape this disadvantage, due to their being in a favourable position to emigrate to countries where public services are adequately provided. In reality, many would still raise justified complaints, since in order to access well-being opportunities associated with public services on equal terms with citizens in developed countries, they would have to relinquish well-being opportunities associated with the special relations forged in their country of origin or with features of their own identity which are shaped by their existence in a particular national community (as is the case of Emmanuel in Engineer). By contrast, for most citizens of developed countries the two sets of opportunities are not mutually incompatible.
Having concluded that the brain drain phenomenon is comparatively disadvantageous for citizens of developing countries, we now move to the question of whether the disadvantage could be mitigated through the policies of extraterritorial taxation and temporary compulsory service. The assumptions which underlie these proposals are that public services would be improved in two ways: (1) more finances would be available for funding the respective public services and (2) more trained professionals would be available to deliver these public services.
Note, first, that both of these claims rely on key empirical assumptions which may not necessarily prove to be supportive. The former assumes that host countries, which are uniquely able to enforce this policy, would agree to a system in which their highly skilled non-citizen residents would be taxed disproportionately more than their colleagues. 13 This would make the respective country less apt at attracting the highly skilled migrant workforce desired, while gaining no economic benefits for itself -as all benefits would be directed towards source countries. It is also unclear how this system would fit into Brock's non-ideal project of spelling out 'what poor, developing states may do to solve their own problems' (Brock and Blake 2015, 95), since developing states would not have much to say on whether or to what extent these taxes will actually be collected. The second claim mentioned above may also turn out to be problematic, since imposing temporary compulsory service only for some professions may turn out to be discouraging for prospective students -and to such a degree that it may lead to less rather than more highly skilled workers available. So it is far from certain that the main disadvantage rightly identified by Brock will be even partly mitigated through her proposal.
But we leave these empirical difficulties aside for now and accept, for the sake of argument, that the policies will indeed manage to improve the delivery of public services in source countries. The question is whether they would manage to do it in a fair manner, or would themselves generate additional disadvantages. Appealing to the well-being value account of the envy test, we can first note that since the act of emigrating would be even more costly for a highly skilled worker (especially if exit taxes were introduced), the comparative disadvantage we've illustrated through Engineer not only remains in place, but is further compounded. Moreover, highly skilled migrants would be required to pay additional taxes relative to citizens of developed countries. This gives rise to a legitimate complaint that their opportunities for well-being would be unequal in an obvious way -they would either have fewer economic resources at their disposal, worse employment prospects, or both. Finally, they would be the only ones required to perform compulsory service in their source countries. This gives rise to a third legitimate complaint, since they would have worse or fewer opportunities for well-being than those in developed states, since -other things being equal -the latter would have more freedom of occupation and of movement than the former. Thus, even if Brock's proposal could manage to alleviate one source of disadvantage -that between citizens in developed and developing countries when it comes to public services -it does so only by severely deepening the already existing disadvantage between highly skilled migrants and citizens of developed countries.
So what can, therefore, be done in order to mitigate the negative effects of brain drain, without causing more unfairness by engendering other types of comparative disadvantages? The chief problem is surely the inability of developing countries to make the jobs which require highly skilled workers in public services sufficiently attractive. There are many reasons why this would be the case. For instance, political instability, lack of political freedoms, widespread corruption, poor general socio-economic conditions etc., are all reasons in favour of emigration, and some of these could be controlled by 'responsible, but poor states'. However, some reasons surely have to do with the particular situation of those jobs themselves. A proper infrastructure as well as wages comparable to those received by similar workers in developed countries, adjusted to living costs, would probably manage to go a long way in alleviating the problem. This is perfectly compatible with Brock's own view. But if we are truly interested in a fair distribution of benefits and burdens we cannot put the latter squarely on the shoulders of highly skilled migrants while taking no principled view on the duties of ordinary citizens in developed countries.
Rather, the scope of any tax scheme aiming to mitigate the disadvantages of brain drain should be extended to all citizens in developed countries. The exact details of how this scheme would look like go beyond the purpose of this paper. But in keeping with the Dworkinian framework we adhere to here, it could simply result from a well-being value account of the hypothetical insurance mechanism, extended so as to cast uncertainty over whether individuals would end up as citizens of developed or developing states. This would presumably put more of the tax burden on the more highly skilled and richer citizens in developed countries rather than on other fellow citizens, which is fully in accordance with our considered intuitions. Further, and more interesting, it would also presumably put more of the tax burden on the highly skilled citizens in developed countries rather than on highly skilled migrants. This is because the opportunity to gain benefits at the same level as highly skilled citizens in developed countries is only opened to highly skilled migrants by foregoing other opportunities of the type discussed in Engineer. Purchasing hypothetical insurance in order to compensate for these types of well-being opportunity deprivations in case of emigration would be rational to do in this situation.
To conclude then, a fair distribution of burdens for mitigating brain drain effects would indeed require taxing highly skilled migrants, but at a somewhat lower level than highly skilled citizens of host countries, and it would require appropriately taxing other citizens in host countries as well. Perhaps Brock would like to challenge this conclusion as being too ideal and paying insufficient attention to what developing states can do by themselves in order to solve the problem. But as we discussed earlier, her own proposal of extraterritorial taxation is faced with identical concerns regarding idealism, since only host countries can enforce effective taxation policies. The difference is that we simply map out how host countries can actually devise a fair system of burden allocation, assuming that they do wish to mitigate brain drain effects, either on their own initiative or under the pressure of international agreements.
Since the proposal of using taxation schemes in developed countries is ultimately one which should be endorsed, albeit in a much more expanded version than Brock suggests, it is worth asking whether the proposal of temporary compulsory service could also be adjusted in a such a way as to make it compatible with the value of fairness. One immediate reaction to this idea might be that the Dworkinian framework employed here would be inimical to the idea of compulsory service, since this would institute a kind of 'slavery of the talented' (Dworkin 1981, 312). But a closer look will reveal that this is not in fact the case, since the envy test is meant to apply diachronically. As long as prospective students are fully informed that a period of temporary compulsory service will be required after completing their studies, they cannot complain that their overall set of opportunities for well-being is reduced because they subsequently have less occupational freedom than other fellow citizens. However, as discussed before, they can complain that their set of opportunities for well-being is diminished by comparison with similarly positioned students in developed countries. This is because the former must accept a reduction in occupational freedom in order to pursue a highly skilled profession, while the latter must make no such concession.
The implication is that in order to satisfy fairness considerations, no distinction should be made from the perspective of compulsory service between highly skilled workers in developed and developing countries. In practice, this would mean either that (1) every highly skilled worker should undergo a period of compulsory service in developing countries at some period after completing their studies, or (2) no highly skilled worker should undergo such a period, regardless of the country of origin. Note that if (1) would be enacted, highly skilled workers in developed countries would also have to depart -even though only temporarily -from their own communities in order to take up positions abroad. While some would probably welcome such a move, others would legitimately complain that the opportunities for well-being of highly skilled workers in developing countries are now better, since they do not have to leave their countries in order to meet the requirements of the policy. Perhaps this could be mitigated by giving more external resources to highly skilled workers in developed countries, but even that might not be able to compensate for the other losses in wellbeing opportunities associated with having to depart their own countries. Furthermore, compelling individuals to temporarily work abroad could have the effect of depressing the pursuit of the respective profession, thereby affecting public services in developed countries as well. Finally, the idea that developed countries would accept to institute and legally enforce this practice is much more idealistic than requiring them to divert some of their economic resources towards developing countries. Therefore, since fairness considerations imply that temporary compulsory service cannot be instituted only for highly skilled workers in developing countries, and instituting this mechanism for highly skilled workers in developed countries is both normatively and empirically problematic, we would do well to reject it altogether. The need for qualified professionals in developing countries is a pressing matter, but it can only be reasonably solved through providing substantial positive incentives to highly skilled workers -in both developing and developed countries -for taking up the required jobs.
Before moving to the concluding part of the article we should briefly consider one immediate response which a defender of Brock's proposal could offer. Such a defender could claim that while it may be desirable for citizens in developed states to contribute to alleviating the problems of brain drain, highly skilled migrants have much more pressing duties to do so because of the reasons outlined in P1 and P2 of the Argument from Disadvantage. More specifically, since they are the ones who increase the vulnerability of fellow citizens by departing, and they are the ones who have benefited from the contributions of their fellow citizens without reciprocating, they are also the ones who should shoulder more of the burden.
In the third section we used the cases of Shopkeeper and Private security to show that even if A disadvantages B in these ways it does not automatically follow that A necessarily has a duty to mitigate this disadvantage. However, in many other cases, such a conclusion would indeed be warranted. Even so, we maintain that this not true for the case of brain drain. Benefit and causal contribution are compelling grounds for assigning primary remedial responsibility only in cases where no other sources of unfairness are salient. Otherwise, these connective grounds can only provide a partial and distorted picture of the duties binding the various actors involved. Brain drain, however, does take place in the context of deep background unfairness, and a fair allocation of these duties can only be provided by widening the scope of our normative analysis. Thus, challenging our proposal in the manner suggested here is flawed, since the objection fails to give proper consideration to both the historical injustices which are partly responsible for the situation of citizens in developed and developing countries and to the fact that duties to mitigate injustices may flow both from the moral requirement to redress harmful actions and from the fact that natural and social contingencies create unjust circumstances.

Conclusions
In this paper we sought to assess two influential policy proposals to combat the negative effects of brain drain, most forcefully defended by Gillian Brock. We maintained that although it is right to emphasize that brain drain unjustly compounds the disadvantages of those who are already in a vulnerable situation -and that this injustice is of urgent concern -, her proposed solutions fail to properly identify the corresponding duties to mitigate brain drain, since they are not based on a robust conception of unfair disadvantage. We further argued that one such conception is readily available, in the form of the well-being value account of the envy test, and that applying this account would necessarily lead to an expansion in the scope of policies to take stock of the duties of citizens in developed countries. Regarding the specific details of our alternative proposal, it differs from Brock's in three principal respects: (1) we are not agnostic about taxation of citizens in developed countries, but rather maintain that this is a central element of any policy which seeks to fairly mitigate the effects of brain drain, (2) we suggest that taxation policies should be somewhat more burdensome for highly skilled citizens in developed countries than for highly skilled migrants in the same countries, and (3) we maintain that temporary compulsory service in source countries should be rejected.