The ethics of tradable refugee quotas

Abstract

Thousands of refugees die each year fleeing prosecution in their home state. But receiving states are often reluctant to admit, process and house refugees. This is in part because refugee protection is a public good, and so subject to free-riding. A promising, but controversial, solution is to set up markets in tradable refugee quotas (e.g., in the European Union). One of the main objections to such proposals is that they lead to the commodification and objectification of refugees. Another objection, less often discussed, is that such markets seem to legitimate negative and prejudicial attitudes towards refugees. In this article, I defend markets in refugee quotas against such criticisms. Given how many die each year, we are more than ever in need of creative solutions. Tradable quotas provide a promising mechanism to increase the number of refugees processed and protected, especially in regional schemes like the EU. Ethical objections such as the ones canvassed here should not stand in our way.

Refugee protection is a public good. Receiving states would rather that a greater rather than fewer number of refugees are protected, but would prefer if other states take up the burden. This is why refugee protection is subject to classic collective action problems, which result in fewer refugees being protected globally than would be optimal (given state preferences). ¹ A promising solution – especially when compared with fixed quotas or insurance schemes that compensate states that receive large influxes – is markets in tradable refugee quotas. One of their main advantages, as we will see, is that they do not require an estimation of the costs and benefits of reception and protection for any given allocation or compensation. ² Such costs and benefits are determined by the market. But markets in refugee quotas are controversial. Most of the controversy ³ turns on whether they objectionably commodify or objectify those they are designed to protect. In this article, I will first argue that both objections fail. I will then raise another, new objection to TRQs, namely that a market in quotas would legitimate demeaning attitudes towards refugees. This is a more promising objection than the two ones just mentioned, but it, too, fails. TRQs, I conclude, have more going for them than is usually thought. Given the high likelihood of increasingly severe refugee crises in the future (e.g., think of climate change), and the large numbers of refugees that die trying to reach places that do not have easy-to-access resettlement schemes, we need creative solutions more than ever. ⁴ TRQs deserve more attention than they have received thus far.

The most well-worked and influential proposal for TRQs comes from two economists, Rapoport and Moraga (which I will refer to as the RM proposal). To fix ideas it is best to have in mind a smaller subset of states rather than the international system as a whole. This is, in part, because such schemes are most likely to work in regions that already have significant institutional capacity at an international level but face significant burden-sharing dilemmas when refugee influxes suddenly increase, such as the European Union. (Consider also that the EU has tried fixed allocation schemes with an insurance component, so far without success.) The scheme is straightforward. ⁵ First, member states must choose an initial allocation of quotas – on which more below – after which they are permitted to pay other member states to take over some or all of their share. ⁶ Once shares have traded, a central administrator (e.g., the European Commission) uses a matching algorithm to assign refugees or asylum seekers to member states based on an ordered ranking: states give an ordered ranking of ‘types’ of refugee/asylum seeker (by skill level, nationality, etc.) and refugees and asylum seekers give an ordered ranking of destination states. The matching algorithm ensures that the ultimate allocation is more sensitive to state and refugee/asylum seeker preferences – and does not (ex ante) favor any particular set of preferences (with the exception of family reunification [see below]) – than a random allocation. ⁷

From an ethical perspective, the matching algorithm used to rank refugee and asylum seeker preferences is a large improvement over proposals that lack any such mechanism. ⁸ Giving refugees and asylum seekers the possibility of ranking, among other things, increases the chances that their temporary/permanent integration into the host community will be successful. (Consider, e.g., that refugees and asylum seekers are likely to rank countries highly in which there are already members from their home community, or where they speak the language.) This is sure to be preferred, especially considering the personal turmoil and upheaval that refugees and asylum seekers have experienced both in their country of origin and en route. Note that the RM proposal also allows for family reunification to play a ‘trumping’ role: if a refugee or asylum seeker has family members in the proposed destination country, then they go to the beginning of the queue (i.e., they receive a visa independently of the algorithm's sorting mechanism).

The matching mechanism for states fares less well under scrutiny. The rationale for the mechanism is, as we will discuss in more detail below, that there are cost savings in allowing states to rank refugee types. If states can rank refugees without upsetting the efficiency of the market mechanism then more preferences will be satisfied than would be the case without such a mechanism. This is easy to see: if we imagine a post-traded allocation of visas by country, then a matching mechanism just determines who in particular will take up those visas: it does not, on its own, either change the cost or quantity of visas. ⁹ Indeed, if states know that they will be able to rank, then they may be more willing, all else equal, to accept either a greater number of total quotas under the scheme or be more willing to join the scheme in the first place.

What's not to like about country-based matching? The trouble is that states are likely to use discriminatory criteria. ¹⁰ One might respond that it would be easy to limit the kinds of criteria states may use when ranking. ¹¹ For example, one might bar rankings on the basis of race, religion or nationality, but allow rankings for language, skill (level or type) and education. But what guarantee is there that states wouldn’t use such criteria to discriminate by proxy? Indeed, skill level or type can be used to discriminate by gender, and education and language to discriminate by country of origin and religion. I will return to the issue of discrimination below. As we will see, similar considerations also implicate the use of a market to allocate visas as such.

I now want to address the two most common ethical critiques of TRQs. The first is the argument from corruption; the second is the argument from commodification. Neither hits the mark. This will open up a distinct but related critique that does better.

Imagine there was a market that allowed one to pay someone else to attend jury duty in one's place. Or, imagine, as was the case in the American Civil War period, that, once conscripted, conscripts could pay someone else to enlist in the armed forces in their place. We may object in both cases on the grounds that introducing a market into duty-based practices of citizenship corrupts the value and significance of those duties. ¹² The market changes the norms governing the practices, and hence their social meaning, in the same way as awarding the Nobel Prize to the highest bidder would change the norms of that practice, and hence its social meaning. (Would the Nobel Prize continue to have the broader social value it currently does if it were offered to the highest bidder?) Suppose, as would be likely, that a market in jury duty (where people could pay others to do their jury duty) would end up creating a professional class of jurors. Suppose further, as would be much less likely, that professional jurors would be no better, or worse, or less representative of the general population, than non-professionals. Would we still have grounds to object? In the case of citizenship-based duties, a market, it is argued, would devalue their social meaning in ways that we would have reason to regret. This is because citizenship-based duties are intended to reflect the insulation of citizenship from market values. In the domain of citizenship, we are all one another's equals – we all, for example, possess the same rights and duties – whatever one's social class, profession, religion, race, sexual orientation and so on. The market, on the other hand, is designed to respond to one's relative preferences for different things, to reflect and embody our differences in background, disposition, talents and beliefs. At the same time, as citizens, we are meant to realize that our ability to make use of our talents in the market – indeed, the very existence of a market itself – depends on the existence of an underlying structure of law and regulation. Without this background structure, our talents would be worthless. Against this background, citizenship duties should be understood as a responsibility we bear that reflects the fair return we owe others, and the state itself, for making a flourishing life possible. They should not reflect those differential abilities, talents, income, wealth or preferences that would allow us either to avoid citizenship duties or to gain unequally from advantageous market trades.

Is an analogous argument available for a market in refugee quotas? One might think there is. States have moral (and legal) duties to offer protection to refugees. These duties are, in the relevant senses, just like an individual's duties to discharge jury duty. Just as it corrupts the meaning of citizenship in the market jury duty case, it corrupts the meaning of global citizenship in the TRQ case. To pay another state to take over one's protection/processing duties, then, is wrong for the same reasons as paying someone to do your jury duty. ¹³

There are two problems with the argument. The first is that it is not clear that there is a global citizenship possessed by states that is equivalent in social meaning to the state-based citizenship possessed by individuals. The equal status that is at the heart of citizenship at the domestic level serves an important function: to shield against the corrosive effects on self-respect of cultural, social and economic inequality. It also serves, of course, to protect individuals against the abuse of power and privilege that would ensue if and when offices are for sale, or the law is no longer impartial. The duties of citizenship, in turn, serve to reinforce that underlying equality: no matter what someone's cultural, social and economic background, one is expected to serve in this war, to maintain the justice system in this state. It is unclear whether state duties to protect refugees serve the same function. Although it is of course true that state equality in international bodies is important both instrumentally and non-instrumentally (along with corresponding rights to collective self-determination enshrined in both of the main human rights covenants), that equality is not supported and reinforced in the same way by state duties to protect refugees. On one popular account (with which I agree), duties of non-refoulement and protection flow from membership in the state system: states owe refugees protection because the nature of the state system (from which all states benefit) makes refugee flows inevitable. ¹⁴ From the point of view of the international system what matters is, therefore, that all states contribute in a fair way to the provision of protection. This is why it makes sense to cast the question of allocation in terms of burden-sharing. But might the burden-sharing afforded by a TRQ scheme undermine the equal standing of states in the international system vis-à-vis one another? Could one say: ‘The point of a fixed (rather than market) quota system is to reinforce the idea that, under international law, state duties do not vary depending on one's GDP, level of development, history or institutions’? I’m not sure. It is surely relevant, one might think, that international law does not have the authority and status-influencing character that domestic law does.

But even if we believe that a market in quotas would undermine the equal status of states under international law in some analogous way, there is another problem that lies further upstream. The problem is the following: how is it that jury duty reinforces equality? Surely the main objection to a market in jury service is that it would undermine the unique role of juries in common law systems. When one is convicted of a crime in a common law system, the rationale for a jury is to ensure that one's case be heard by a jury of one's peers. The point is to have a range of people who are representative of the wider community. Paying someone else to take our place would corrupt our standing as equals before the law by biasing the system. A market in jury duty service, that is, undermines citizenship by undermining the uniquely impartial and representative role that jurors are expected to play within a common law system, not by corrupting the social meaning of citizenship as such. If creating a market in jury service didn’t have the effect of undermining the impartiality, representativeness or effectiveness of juries, it would be unproblematic. After all, there is already a market in adjudication: judges are paid competitive wages based on market pressures (and let us not forget that jurors are paid, too). To draw the analogy, imagine, in addition, that there was an official system whereby judges could sell cases on their docket (that have not yet begun) to other judges by paying them a fee (or by forgoing some amount in wages sufficient to cover their time). We may object to the bias, partiality and inefficiency such a system would introduce. But suppose there were no such bias, partiality or inefficiency. What reasons to object would be left? After all, judges would get the cases they most wanted and were most interested in, as well as enable them to manage their overall workload. If a market in jury duty corrupts citizenship, then why wouldn’t the system just mention corrupt citizenship, too? ¹⁵

To be sure, reasons for fairness are also important. Allowing some to pay others to do something considered onerous can be unfair, for example, if the marginal costs of paying someone the market rate are higher for one party than for the other. This is typically the case if one party is rich and the other poor. But notice that this is not an argument from corruption. According to this objection, there is nothing wrong with paying someone to do one's duty as such. Were marginal costs to be equal, there would be no unfairness and so no objection: each would have the same opportunity as everyone else to pay their way out (or in). ¹⁶

Similar arguments could be made when evaluating whether it is legitimate to hire substitutes for military duty. If there is any objection to paying substitutes, it lies either (a) in the possibility that paying substitutes would make the army less efficient or effective, or (b) in the possibility of exploitation (i.e., the idea that anyone who is willing to put their life at risk in the way a soldier does must have been in some way forced to do so, by, e.g., having no reasonable alternative). But if either of these constitutes grounds for objection, then the argument has nothing to do with the corruption of social meanings, and it would be much more difficult to translate to the TRQ case since states are not putting the equivalent of their lives at risk by accepting payment to accept additional refugees.

If this is right, then the analogy to TRQ schemes dissolves: Paying other states to take over one's allocation does not undermine the function or role of refugee protection in the way paying jurors would undermine the function or role of a jury system. Indeed, if refugees are allocated, as a result of a market, to states that are more willing and able to offer protection (as reflected in their preferences for accepting additional refugees), then it could enhance that function (especially if states, given the presence of a market, would be both more willing, all else equal, to accept a higher total quota of refugees or asylum seekers and more willing to enter such schemes in the first place). I return to this point below.

The argument from commodification is more straightforwardly problematic. ¹⁷ The argument says that what makes a TRQ scheme unacceptable is that it involves the buying and selling of people, and so is analogous to the trading of chattel or slaves. This argument rests on an overstatement. What is being traded are not people but duties to protect, that is, visas. Refugees and asylum seekers within the scheme will be protected no matter which state ends up responsible for issuing the visa. They are, furthermore, given a chance to rank preferred destinations. In what sense is their position, then, analogous to slaves or chattel?

Perhaps the claim is that buying and selling quotas are like the buying and selling of chattel in the sense that it treats refugees as things that can be bargained over and sold. Again this strikes me as an overstatement. If buying and selling visas is equivalent to treating people objectionably as things to be bargained over, then why isn’t any labor market? Take, for example, the transfer market in football (soccer), or any other professional sport. Players on the transfer list are available to be bought by other teams, which can bargain to acquire the best players. Of course, in such cases, it is not literally people that are bought and sold, but their availability to play for clubs. From this point of view, the TRQ scheme seems to score more highly than the transfer market, since the visas are bought and sold before the identity of refugees allocated to states is known. And, crucially, the buying and selling of quotas do not affect the overall number of visas granted, which is fixed in advance. So if you are a refugee within the system, you know that you will be protected somewhere, even if there is a bargaining process determining exactly where (also recall that family reunification and language are taken independently into account, and that there is a preference matching algorithm that improves your chances of getting your preferred location).

A better version of the objection has the following form. Isn’t a market in refugee quotas likely to encourage some states to reduce standards of protection while accepting as many offers as possible to take in and process refugees from other states? Such states would then be objectionably using refugees for monetary gain while doing nothing to protect them. The states in question might both make acceptance of claims more difficult (when processing applications) and hope that the low standards of protection would encourage those who do receive protected status to move on to another state (illegally).

This is an important objection. But there are mechanisms available that might address it (some of which have already been implemented within the EU, albeit with mixed success ¹⁸ ). The mechanisms involve setting minimum standards for reception, screening, processing and return. Establishing a floor beneath which no state can sink in standards can safeguard against abuse of the TRQ scheme. These mechanisms can then be buttressed by stronger human rights protection. ¹⁹ Of course, the implementation of such a scheme faces huge political, practical and administrative hurdles. Procedures for harmonization in standards, for example, are likely to be used primarily to streamline procedures rather than protect refugees, and to make rejection and return of failed asylum applicants more efficient rather than fairer. ²⁰ But similar obstacles will face any policy in this area given the current political and economic climate (including fixed quotas). The important point is that there are policies that are available and that, with the right degree of support, would be successful in mitigating the dangers just canvassed.

I will now argue for a more convincing objection to the market aspect of TRQ schemes. Let us return to what is in many ways the most difficult aspect, both morally and politically, of a TRQ scheme, namely the initial allocation of quotas (before trading). ²¹ One might think that an obvious way of making such an allocation is to assign equal shares to each member state. But this would be absurd: Why should Luxembourg have as large an initial allocation of quotas as Germany, a much larger country? And why should Bulgaria, given that it is both smaller and less rich? The initial allocation should surely take into account population and GDP. We may also think that how many refugees and asylum seekers the state in question has previously hosted is relevant: where a state has already hosted many, it would be justified in hosting relatively fewer in the future. Finally, we might also believe that the level of unemployment in a country is relevant (reflecting the relative difficulty that a member state would have in integrating refugees and asylum seekers during periods of high unemployment). In line with this reasoning, Rapaport and Moraga suggest using the EU Commission's distribution key (established during the Syrian crisis): national GDP (weighted 40%), size of the population (40%), unemployment level (10%) and the number of asylum-seekers already hosted (10%). After those quotas have been assigned, let the trading begin.

What other factors might we want to include to determine the initial allocation? Could member states argue, for example, that the expected welfare and administrative costs of integrating refugees and asylum seekers should be included? According to this rationale, a state that pays more per refugee or asylum seeker in welfare and administrative benefits should be assigned a lower initial quota. And what about other, more intangible, political, cultural and social costs – for example, the costs to political stability and legitimacy that admitting more refugees and asylum seekers would incur, or the costs to cultural (or ethnic or racial) homogeneity, or the costs in increased criminality and social dysfunction? And what should matter: the social, political and cultural costs that a member state's citizens perceive they must pay to integrate refugees and asylum seekers or some objective estimate of those costs? Furthermore: Whether or not one has objective or subjective estimates of such costs in mind, how does one commensurate and measure them? Consider, for example, that states have strategic incentives to misrepresent the costs they will incur in order to attain a lower quota.

Indeed, this is why Rapaport and Moraga defend a market solution: the market functions to reveal state preferences across all these further costs and equalizes opportunity costs across states. This is easy to see. Suppose we stay with the initial allocation implicit in the Commission's distribution key. Now ask yourself: why and under what circumstances would states pay other states to take over a portion of their quota? They would if the marginal (cultural, social, political, administrative…) costs of integrating an additional refugee within their quota are greater than the marginal costs of paying another state to take over that visa. And why and under what circumstances would a state be willing to take an additional visa? They would if the marginal (cultural, social, political, administrative…) costs of integrating an additional refugee or asylum seeker are lower than the marginal benefit secured by the price offered. Trades are likely, that is, when it is more efficient for one state to integrate refugees and asylum seekers than for other states to do so. The price of a visa on the market reflects the opportunity costs to other states of taking on an additional refugee; if there are no distortions or externalities, the market will equalize the opportunity costs of taking on more refugees and asylum seekers across states (without, nota bene, needing to collect any information independently about those costs). All states end up better off (by their own lights) with the market than without, and no state can complain that they would rather exchange their particular bundle of visas ²² with another state.

Return now to the (perceived) costs that are likely to lead a state to pay other states to take over a portion of its quota: welfare and administrative costs, the contribution of refugees and asylum seekers to political instability and loss of legitimacy, pressure on public services, perceived increases in crime and social dysfunction, preferences for cultural, ethnic and racial homogeneity. We saw that it would have been difficult to include such costs in determining the initial allocation. This was because such costs are difficult to determine (states have incentives to misrepresent) and because it would be difficult to devise non-arbitrary metrics to commensurate them. Creating a market solves these problems since it is, in practice, a mechanism for revealing preferences. But now suppose that such costs would have been possible to determine ex ante. Would it have been morally appropriate to include them? On the more plausible side of the continuum are welfare and administrative costs. It seems clear, however, that preferences for, say, racial or ethnic homogeneity, or (false or exaggerated) beliefs about the impact of refugees and asylum seekers on crime or public services, or political unrest or dissatisfaction based on any of these, would not be permissible as a basis for the initial allocation. This is because these further ‘costs’ reflect discriminatory (and often false) beliefs that target the self-respect of refugees and asylum seekers in the same way as discriminatory laws that target the self-respect, say, of minorities or the disabled. ²³

But, if that is right, and if the operation of a market is bound to reflect those very same costs then – the objection concludes – we ought to believe that a market is morally problematic for the very same reasons. After all, the rationale for a market solution was, as we saw above, precisely to take these costs into account in determining a final allocation. The market serves, that is, to legitimate the expression of those preferences. By licensing states to act in them, the market ensures that the final allocation will not just only reflect the criteria used in the initial allocation but also the preferences expressed in trading.

In making this objection, an analogy is useful. In a well-known article, Rae Langton argues that pornography silences women by making it permissible to subordinate women (e.g., by treating a woman's ‘no’ as a ‘yes’). ²⁴ Leave aside whether this is a convincing analysis of pornography. ²⁵ Here, I want to focus on the logic of that argument. In what sense does pornography ‘make it permissible’ to subordinate women? It is not as if pornography can make it morally permissible to do so. The way pornography operates, Langton argues, is by legislating a social convention that determines women's inferior civil status. It is important that the impact on the effectiveness of women's voice is not causal but normative. Of course, it may be true that pornography may cause and depict women's subordination. But her point is that, whether or not pornography makes subordination more likely, it enacts and reinforces a societal norm that such subordination is permissible. Put another way, through its portrayal of women, pornography doesn’t simply send a signal that it is (conventionally) ok now and around here to treat women as inferior but makes it the case that it is ok to treat women as inferior. Langton uses Austin's speech act theory to make her case. Verdictive speech acts are speech acts that value, rank or place. An example is an umpire declaring the ball ‘out’. Exercitative speech acts are speech acts that permit, prohibit or enact. An example is an apartheid law that declares ‘blacks shall not vote’. Notice that both speech acts, for their success, require that the issuer be an authority. Langton argues that, because of the pervasive presence and use of pornography (especially among the young), pornography carries authority in legislating (especially young people) what is (conventionally) ok and not ok in relating to the opposite sex (Langton doesn’t discuss gay pornography). It is hence in the positions, analogously, of the apartheid legislator and the tennis umpire.

The objection to a market scheme works in a similar way: by licensing and legitimating what are often hostile and discriminatory preferences, it not only signals that these preferences are ok now and around here but also makes it the case that they are ok. It enacts a new societal convention: it is now legitimate for member states to shift duties to protect refugees and asylum seekers when they are perceived as criminals, racially or religiously alien, unwilling to work (or willing to take work from citizens) or parasites. Just as in the pornography case, this is not a causal argument. The argument does not rest on the idea, that is, that instituting a TRQ scheme – and hence allowing (some) states to reduce quotas through trades – will make discriminatory or hostile preferences more prevalent. The argument is, rather, that instituting such a scheme will make it the case that, from now on, it is socially permissible to express discriminatory and hostile preferences towards refugees.

This objection is much more powerful than the objections from commodification and objectification. The objection, however, doesn’t work for several reasons. First, it is unclear how the TRQ scheme legitimates the expression of discriminatory attitudes towards refugees. After all, the TRQ scheme makes it the case that states must pay more to get rid of their quotas (given the initial allocation); states that demand reductions, that is, are penalized by having to pay for the privilege of reducing their quota.

Second, it is unclear how instituting a market in refugee quotas legislates a societal convention. In the pornography example, it is crucial that pornographers be taken as authorities in the sexual realm. The assumption is that (especially young) people look to pornography for instruction regarding the moral and social norms governing sexual conduct. Although of course not all pornography depicts the subordination of women (take, e.g., gay porn or porn that is made to counter such norms), most of the porn to which (especially young) people are exposed does. But who is the relevant verdictive authority in the TRQ case (and who is the relevant audience)? And which social norms, exactly, are legislated as a result of the institution of a market in refugee quotas?

Consider first that, while it would be possible to observe and/or infer the presence of pressure on governments to reduce the numbers of refugees processed and protected through a TRQ scheme, it will be unclear in any given instance how much of that pressure is due to xenophobic or otherwise discriminatory preferences. Let us leave this complication aside. Let us take a simple case in which we know that a particular government's willingness to pay other states to take over part of its quota is due mostly, if not entirely, to such preferences. For the objection to work, it must be the case that some social norm governing the (perceived) permissibility of expressing such preferences shifts or is reinforced as a result of a judgement from an authority. But which authority? The authority must be (in our regional example) the EU as a whole. For the argument to work, the EU's institution of a market scheme must make it the case that publics (a) take the EU to be an authority on what is morally or socially permissible to express by way of preferences, and (b) believe that the EU has confirmed that it is morally or socially permissible to express negative, hostile or discriminatory preferences towards refugees. Neither seems likely, especially in countries, such as Hungary, which have strongly aversive preferences regarding refugees. ²⁶ Consider that these preferences would exist whatever the EU says. It is also likely that publics in such countries do not believe that expressing such preferences is socially or morally impermissible to begin with (or to be in any doubt about its permissibility; compare young people in the pornography case). In any such country, there are likely to be parties that express and reinforce the preferences whether or not there exists a market in refugee quotas, and whatever the EU says.

A better analogy is the common policy of giving clean needles to heroin addicts. Does a city that enacts such a policy make it the case that it is now socially or morally permissible to take heroin? Much depends on the point of the policy, the way it is presented, and its implementation. If the policy is enacted in order to get heroin users off the streets, give them information that might help them get off drugs and reduce the likelihood of disease and death, then it seems clear that the policy does not legislate a societal convention that it is ok to use drugs. (Note that it is also highly unlikely for any users and potential users to regard a clinic of this kind as an authority on what it is permissible or impermissible to do.) In the clinic, drug use is tolerated not encouraged. If, on the other hand, the policy were not enacted for any of these reasons – if, say, there were posters on the walls of such clinics celebrating drug use, if the personnel at such clinics were highly respected people within the community of users and potential users, and if the personnel encouraged drug use – then things would be different. In the latter case, it would be right to say that the clinic legitimated drug use.

The institution of a TRQ scheme (such as the one proposed for the EU) would be much more like the former than the latter. Consider that one of the most compelling reasons to adopt a TRQ scheme (rather than a fixed allocation scheme) is that we can expect that a greater number of refugees and asylum seekers would be accepted into the scheme, that states would be more willing to accept the scheme, and that the integration of refugees and asylum seekers would be more successful. These predictions are warranted because a market scheme would give states flexibility and a chance to opt-out (at a price). Compared with a fixed quota scheme, publics would be less likely to feel that they were being compelled to accept refugees they do not want. Furthermore, states that are more likely to be relatively hospitable and successful in integrating refugees and asylum seekers would be more likely to accept a larger share of quotas than states that are less likely. ²⁷ And refugees would have a greater measure of freedom to choose which states to end up in than is currently the case. There are further advantages. In a fixed scheme, there would be deadweight losses: there would be (barred) trades available that would make everyone better off, and no one worse off. States for whom it is relatively less costly to integrate refugees would end up taking up fewer visas compared to states for whom it is more costly to integrate. Comparative advantages in integration would be wasted. In short, one of the main reasons for adopting a TRQ scheme is that it is likely to increase the number of refugees processed and protected overall and decrease the overall costs of doing so, while at the same time promoting their successful integration in the host state. For all of these reasons, it seems evident that a TRQ scheme would be much more like the standard heroin clinic than the non-standard one.

To be sure, this conclusion might seem conditional. In particular, it might seem conditional on the EU's (or other regional institutions) not becoming more like the nonstandard heroin clinic. Couldn’t it be argued, for example, that the EU, as a whole, promotes discriminatory and hostile preferences towards refugees? Given the performance of the EU in dealing with the Syrian refugee crisis (with better results in Ukrainian), the EU's recent reallocation deals with unsafe or inadequate third countries (e.g., Libya and Turkey), ²⁸ the EU's safe country of origin policy, ²⁹ and the EU's human-rights free border policing ³⁰ and criminalization of rescue NGOs, ³¹ it would not be far-fetched to argue the EU reflects and reinforces hostile and discriminatory practices vis-à-vis refugees. Against that background, wouldn’t a TRQ scheme be equivalent to the heroin clinic that encourages rather than mitigates drug use?

No. There would still be an important difference. Recall that the whole point of a TRQ scheme is to increase the number of refugees protected and processed. This is equivalent to the heroin clinic's aim of reducing death, addiction and disease among heroin addicts. If a TRQ scheme could be implemented in the EU without changing any of the EU's other policies, it would most likely serve to counterbalance the negative legitimating functions of the EU's other policies. The analogy would be a heroin clinic that opens despite a city's overall promotion of drug use.

One might object: But what do we say about those countries (like Hungary in our example) that seek to lower their refugee intake due (let us assume) to hostile and discriminatory preferences? Suppose they supported the TRQ scheme (or were forced into it) merely to reduce, as much as feasible, their refugee burden; they then go on to pay sizeable sums to other states to take over their share. And suppose all of this were commonly known. Wouldn’t the state's policies serve to legitimate those preferences? In response, there is no doubt that the state is criticizable for its policies in part because of the attitudes it expresses towards refugees, and the way it legitimates those preferences among its population (compare a state many of whose citizens have hostile and discriminatory preferences but that actively tries to suppress or discourage those preferences – perhaps the German state's 2015 promotion of a ‘welcome culture’ is an example ³² ). The problem with this argument is that it does nothing to impugn the TRQ scheme. To be sure, those who institute the hypothesized EU TRQ scheme would know that there will be countries like Hungary that will take advantage of the system to reduce their overall burden. But they tolerate rather than encourage such use of the system, recall, in order to increase the overall number of refugees protected overall. So, once again, the EU and its hypothesized TRQ scheme are more like the good heroin clinic, with Hungary playing the part of the drug user that instrumentalizes the clinic's services to supply and sustain their own drug use.

The aim of this article has been to argue that markets in tradable refugee quotas have a lot more going for them than most generally believe. The major ethical objections to such a scheme – from commodification, objectification and legitimation – all either fail or must be heavily qualified. To be sure, there are many practical, administrative and political hurdles that lie in store for any serious attempt to implement such a scheme. But ethical ones do not lie among them.