Last week’s pictures of the tragic death of a refugee toddler brought into focus the ongoing crisis facing Syria and other conflict zones, which have given rise to increased numbers fleeing persecution worldwide, and in the European Union. The events of recent weeks (increasing numbers landing in Greece, the journey of many via the Western Balkans, Hungary and Austria to Germany) demonstrate that the Schengen and Dublin systems are coming under strain as a result.
How should the EU address this issue next? Should it abolish or reform the Schengen and/or Dublin rules? Are Member States complying with EU and international law in their response?
To answer these questions, I will examine in turn (a) the international law framework; (b) the EU law framework; (c) whether Schengen is at ‘fault’; (d) whether Dublin is at ‘fault’; and (e) what the EU should do next. My main purposes are to explain the legal background, to point out some legal errors, and to suggest the best way forward in light of the international refugee law framework.
International law framework
While it is often strongly asserted that 'international law requires refugees to apply for asylum in the first safe country they enter', in fact the position is rather vaguer than that. The United Nations (Geneva) Convention on the status of refugees does not contain any express rule to that effect in the rules on the definition of refugee, or on the cessation (loss) or exclusion from being a refugee, as set out in Articles 1.A to 1.F of that Convention.
However, there are some indirect suggestions in the Convention that the number of countries which a refugee has crossed through might be relevant. Article 31 of the Convention bans States from imposing penalties upon refugees for entering a country illegally - a rule which contradicts some of the rhetoric about refugees being 'illegal immigrants'. The drafters of the Convention clearly recognised that it might be necessary and legitimate in practice for a refugee to breach another country's immigration laws to escape threats to his or her life or freedom. So it is not necessary to be on the territory legally in order to qualify as a refugee.
This rule is, however, subject to several conditions - including the requirement that the refugees were 'coming directly' from the country which they had to flee. If that rule is interpreted narrowly, then refugees can only benefit from the exemption from penalties for breaching immigration law in neighbouring states, not states further afield. But refugees’ failure to satisfy this condition only permits States to prosecute them for breach of immigration law; it does not allow those States to exclude the refugees from protection. As I pointed out already, the rules on definition and exclusion of refugees in the Convention are quite separate from the rule on non-prosecution for breach of immigration law. And it is also possible to interpret this condition more generously - in the sense that the 'coming directly' requirement does not exclude all refugees who have merely transited through other countries, but only those who have stopped and obtained protection in another State already.
Another relevant provision is the 'non-refoulement' rule in Article 33 of the Convention, which prevents States removing refugees to an unsafe State. But it does not prevent refugees from being removed to a safe State. Furthermore, the restrictions in the Convention on expelling refugees to any State, and many of the substantive benefits which the Convention gives to refugees (such as welfare and access to employment) are reserved for those who are lawfully resident or present in the territory; and the Convention does not require States to give refugees a lawful status under national immigration law.
So overall, the Geneva Convention gives States a degree of flexibility to insist upon a 'safe third country' requirement, but there is no absolute rule that refugees must always apply in a safe third country. If the Convention had intended to impose a firm rule in that regard, it would surely have said so expressly and provided for obligations for the first safe country to readmit the refugee. Moreover, the preamble to the Convention refers to the heavy burden which the grant of asylum may place upon some countries, and the need for international cooperation to avoid refugees becoming a source of tension between States. Taken as a whole, then, the drafters of the Convention recognized that a strict safe third country rule could impose undue burdens on countries neighbouring a conflict in some cases, but left it to States to work out the details of how to address such burdens when they occur.
EU law framework
The EU law framework consists first and foremost of the Schengen rules - which set out conditions for crossing external borders, in conjunction with no controls on internal borders. While some appear to assume that the Schengen rules require EU States to refuse entry to refugees at the external borders, that is clearly not the case. The Schengen Borders Code contains general exceptions relating to refugees and human rights, as well as specific asylum exceptions from the normal rules on the grounds for admission, and from the requirement to penalizing those who cross the external border without authorization. The latter exception obviously reflects Article 31 of the Geneva Convention, discussed above.
While Member States are generally obliged to ensure control of the external borders (just like non-Schengen States aim to control their borders), there is no specific requirement to build fences, as several Member States have done. Building fences is not ruled out by the Schengen rules, but it is Member States which decide to build the fences as a means of controlling the border, not the EU. And anyone who makes it to those fences and applies for asylum is entitled to be admitted to have their asylum application considered.
This is confirmed by the EU’s asylum legislation, which says that it applies to all those who apply at the border or on the territory. There are some optional special rules for asylum applications made at the border, but there is no rule saying that an application must be refused because it was made at the border, or because the applicant entered the territory without authorization. Reflecting the interpretation of the Geneva Convention discussed above, the EU’s asylum procedures Directive states that an application might be inadmissible if the asylum-seeker gained protection in a ‘first country of asylum’, or has links with a ‘safe third country’. The application of these rules doesn’t mean that the asylum-seeker is not a refugee; rather it means that another State is deemed responsible for resuming protection, or for assessing the asylum application.
The problem for refugees is reaching the territory in the first place. EU law imposes carrier sanctions on transport companies if they transport persons without visas, which explains why refugees do not buy relatively cheap tickets to travel instead of paying smugglers for unsafe journeys. Arguably the EU’s visa rules require Member States to issue special visas for those in need of humanitarian protection (see discussion here), but they rarely do so. Taken as a whole, then, the Schengen borders rules have a rather uneasy co-existence with asylum law: but they clearly allow for the admission of asylum-seekers who do reach the external borders of the EU.
There’s another uneasy relationship between the Schengen rules and the Dublin rules, ie the rules which require an asylum-seeker to apply usually in the first EU country which he or she reaches. That’s because it’s obviously harder in practice to enforce those rules without border controls between countries.
Is Schengen at fault?
As noted already, the Schengen system does not ban people from seeking asylum at EU borders, although it makes it harder for them to reach the territory and more likely to risk their lives trying. But there is nothing in EU law to prevent Member States from resettling large numbers of refugees directly from conflict zones if they wish to. The decision of most Member States to resettle few Syrians or others in need of protection is up to them alone.
While the Schengen rules logically make the Dublin system harder to enforce, this difficulty is relative: after all, if an asylum-seeker moves on from Greece, he or she will have to cross further borders anyway before reaching the rest of the Schengen area (Romania, Bulgaria, and Croatia are not in Schengen yet; and most asylum-seekers prefer instead to travel via the non-EU countries in the Western Balkans anyway).
And it is naïve in the extreme to assume that reimposing border controls would stop all movement of asylum-seekers between Member States. As discussed in Agnes Hurwitz’ book, the Dublin rules have their origin in Council of Europe discussions in the 1980s, and were not initially connected to the Schengen project. There was already a ‘refugees in orbit’ problem of refusal to take responsibility at that point – when border controls were still fully in place. Otherwise States would not have opened discussions on the issue. So clearly borders are permeable even when border controls exist, particularly the long land borders on the continent. Indeed, Germany last received huge numbers of asylum-seekers during the Bosnian war of 1992, again when the border controls were fully in place, which was years before countries like Hungary joined the EU (Austria was not yet a member either).
Does Schengen encourage asylum flows, as some claim? Looking at the statistics, it clearly does not. The flows of asylum-seekers last reached 2014 levels back in 1992 and then dropped off. There have been increases and decreases over the years, which can clearly be linked back to events in countries of origin (Bosnia, Kosovo, Syria, et al). If Schengen caused asylum flows, one would expect the numbers to increase after the borders were abolished in 1995, and again when the system was extended to Central Europe in 2008. But they did not. Nor did the numbers of irregular migrants.
Think about it. The asylum-seekers who reach the EU have often crossed many Asian or African borders already, and those arriving in Greece plan to cross some European borders anyway after they arrive, unless (improbably) Greece is their intended destination. They have fled poverty or persecution, paid smugglers a small fortune, often been ill-treated on the way to the EU, and endured an appalling and dangerous sea crossing. The EU’s assumption that withdrawing rescue vessels in the Mediterranean would deter them from coming proved to be tragically wrong. Compared to all that, why would the reimposition of Schengen border controls deter anyone who would otherwise come?
Conversely, is Schengen at fault for the treatment of asylum-seekers by Hungary? In principle, if third-country nationals do not (yet) apply for asylum, they are irregular migrants, and so Hungary had an obligation to remove them under the EU’s Returns Directive. But that Directive does allow Member States to apply higher standards, and it would have been foolish not to do so (as Hungary eventually did) in circumstances where Germany had already signalled its willingness to consider their planned asylum applications.
Is Dublin at fault?
The Dublin system has undoubtedly shifted a significantly higher burden to certain Member States (Greece, Italy and Hungary at present), which is increasingly difficult to manage as migration flows have increased and the Greek economy in particular has suffered from austerity. It should be noted, however, that in principle the courts have ruled since 2011 that Greece is not responsible for all the asylum-seekers who come there. The normal assumption that each EU country is safe has had to be suspended, since the ECHR and the EU courts have ruled (in the cases of MSS and NS) that Greece is not safe, due to the collapse of the asylum system there.
So it’s clearly legally incorrect to claim that ‘Greece is safe’. Moreover, since the Dublin rules are effectively disapplied to Greece, asylum-seekers can’t be criticised for evading the Dublin system if they move on from that country. However, Greece still has the initial burden of dealing with the large number of asylum-seekers which reach its territory first; and it may be difficult in practice for those asylum-seekers to move on quickly to other Member States. It’s not clear if the first other Member State they reach after Greece then becomes responsible under the rules or not.
The large number of asylum-seekers moving in recent days has probably been partly affected by the German decision that it would not apply the Dublin rules to Syrians. Germany is legally entitled to do this: the Dublin Regulation has a sovereignty clause (allowing States to consider asylum claims that are not their responsibility) and the CJEU has ruled that there are no limits on how Member States may use that clause. This will undoubtedly exacerbate the burden on Germany in turn.
While it makes more sense from a burden-sharing perspective to allocate asylum-seekers between Member States more evenly, there are a number of political problems with that idea. Since Member States are not keen to have more asylum-seekers, the would-be net recipients always resist the suggestion. Various proposals for burden-sharing have repeatedly failed since the 1990s. The EU is on the brink of sharing burdens for the first time, by adopting a Decision on relocating some asylum-seekers from Greece and Italy to other Member States. But the number involved is modest (less than 40,000) and the offers to take asylum-seekers are voluntary. The Commission’s proposal was for 40,000 people distributed by a mandatory scheme, but Member States would not accept this.
This brings me to my next point: did ‘the EU stop the refugees coming’? Yes, to the extent that it did not give sufficient relief (in the form of relocation of refugees) to Greece and Italy; but the EU’s response on this was only insufficient because Member States, voting in the Council (made up of national ministers), blocked the EU from adopting even a modestly more ambitious relocation decision. As for refugees coming from outside the EU, as noted above the EU does make it hard for them to get here (because Member States want it to do so), but Member States are free to resettle people as much as they like. If the EU did not exist, would Member States really suddenly be willing to admit many more refugees?
The central issue then is the role of national politicians. While the European Parliament has a joint decision-making role on most asylum legislation (not a purely advisory role as some have claimed), it does not have that usual role when it comes to the relocation decision, since it is an emergency measure. And in turn, national politicians are not on an anti-refugee frolic of their own, but responding to the significant proportion of public opinion that still does not want to see significant numbers coming to stay in the EU (see this recent UK opinion poll, for instance). Those who want a more liberal approach to refugees and migration need to focus on convincing their fellow citizens, not blaming the EU.
What should the EU do next?
With all this in mind, what should the EU do next? As I upload this blog post, Commission President Juncker is about to make his ‘State of the European Union’ speech with new proposals. It seems likely that they will include some of the ideas already previewed in May, as part of the EU Migration Agenda (discussed here): a common list of ‘safe countries of origin’; the greater use of Frontex in returns procedure; and new rules on emergency exceptions from the Dublin rules. The latter will be much more ambitious than the first decision to this end (aiming for 120,000 more asylum-seekers relocated, according to reports), although whether Member States will accept this remains to be seen.
I will blog about the details of the new proposals when they emerge, but for now here’s an overview of what I think the priorities should be. (For other suggestions, which partly cross-over with mine, see the position of the UNHCR and Human Rights Watch. I agree with their suggestions to step up rescues and to ensure safe passage for more refugees, but I will elaborate here on some other ideas).
First of all, looking at the situation within the EU, it’s obvious that there needs to be some form of relief (far more than provided for in the relocation Decision about to be adopted) for the Member States bearing a very large share of the burden. As I pointed out at the outset, the concept of burden-sharing is intrinsic to international refugee law (see also the comments on this by my colleague Geoff Gilbert, and by Michael Ignatieff). It’s also recognized in EU law by Article 80 of the TFEU, which refers to the principle of solidarity between Member States.
On that point, it’s reported that a group of newer Member States is resisting not only the idea of relocating asylum-seekers, but even the compromise suggestion of making an extra financial contribution in lieu of this. This is a flagrant breach of the burden-sharing principles of international and EU asylum law. Moreover, since these countries have benefited enormously from their citizens’ refuge-seeking in and economic migration to other countries (both into and outside the EU) as well as a substantial inflow of EU funding, their position is morally untenable.
Secondly, it’s apparent that there needs to be more coordination of national policies within an overall EU framework. The overall impression given is shambolic, although this has not stopped some commentators from ascribing responsibility for every decision of the Hungarian government or even a non-EU state’s police (in the Former Yugoslav Republic of Macedonia) to the EU. Some new forum – perhaps a framework for ministers, civil servants and the heads of EU agencies to meet to discuss rapid reactions – needs to be established. Clearly the overall amount of humanitarian assistance for those reaching Member States’ shores also needs to be increased.
Thirdly, on a similar theme, the consistent interpretation of EU rules needs to be enhanced. There are too many examples of divergent approaches to EU law which should in principle be ‘uniform’ (although Member States have scope to apply higher standards). These sort of issues can partly be addressed through the coordination framework discussed above, but further measures are clearly necessary.
Let’s look at several examples. The Hungarian government has just passed a law to impose custodial criminal penalties upon those who cross its external borders fence. As we have seen already, refugees must be exempt from penalties for irregular border crossing, at least in some cases. Moreover, the CJEU has ruled that irregular migrants should not be subject to custodial penalties for unauthorised entry, since that delays their removal (for a summary of the case law to date, see the opinion in the pending case of Celaj).
Several Member States also have stated that they do not wish to take Islamic refugees. This is again a clear breach of international and EU law: the Geneva Convention specifically states that it applies without discrimination on ground of religion, while the EU Charter of Rights bans discrimination on grounds of religion when applying EU law (and the asylum process in all its aspects amounts to applying EU law). As for the bizarre argument that ‘our country doesn’t have a mosque’, it is in fact possible for Muslims to pray elsewhere. Refusing refuge to Muslims is not necessary to keep Europe Christian; rather It’s a Christian duty to offer refuge, as Pope Francis and the Archbishop of Canterbury have pointed out.
Also, there have been examples of individual mistreatment of groups of refugees and irregular migrants, in the form of police action or border guard pushbacks, or inadequate living conditions. While Member States’ border guards and military forces have surely rescued many more people than they have mistreated, there needs to be a commitment to ensure redress for the latter cases. Although the Schengen Borders Code generally requires border guards to behave courteously and fairly, it would be useful to agree common standards on prohibited behaviour.
On the living standards point, note that it is not accurate to say (on one journalist’s blog) that someone who applies for asylum in one Member State even though another Member State is responsible for their claim under the Dublin is a ‘migrant’. The CJEU expressly ruled in its judgment in CIMADE and Gisti that such a person is an asylum-seeker and is entitled to the relevant benefits until they are transferred to the responsible Member State under the Dublin rules, The same rule is now expressly set out in the preamble to the Dublin III Regulation.
More generally, the EU should reflect on whether more far-reaching moves to achieve more harmonisation in practice should be attempted: for instance, transforming the European Asylum Support Office into a body able to make decisions on asylum applications in ‘overflow’ cases (if need be, by seconding national officials for the duration), and/or creating a common European asylum appeal court. There are still wide differences in recognition rates of refugees between Member States, despite a common legal acquis: the low recognition rate of Eritreans in France as compared to other Member States may have contributed to the ‘Calais crisis’, for instance. In the meantime, the Commission needs to step up infringement actions for breaches of EU immigration and asylum law.
Fourthly, more generally, should Schengen survive? It follows from the analysis above that ending the Schengen system and reimposing internal border controls would likely have little impact on the overall flows of migrants and refugees coming to the EU. It might have a modest impact on restricting their movement between Member States, but the desirability of that outcome is tied up with the merits of the Dublin rules, to which I turn below.
But if Member States are serious about ensuring that irregular migrants and asylum-seekers do not cross borders to flout the Dublin rules, they would have to bring back internal border controls with a vengeance, building fences and installing border guards across many thousands of kilometres of common borders. Except for Malta, the Schengen states do not have the built-in border control that the UK, Ireland and Cyprus (also not in Schengen) have, for obvious geographic reasons. The likely outcome would be dozens of ‘Calais’ situations at the various internal borders of the Union.
Furthermore, the basic obligation to drop internal border controls is set out in the Treaties, and so a permanent suspension of Schengen would require a Treaty amendment. A reasonable compromise might be to amend the Schengen rules to allow for more temporary internal border checks whenever intelligence shared between Member States suggests that a large number of irregular migrants or asylum-seekers is likely to cross an internal border.
Fifthly, should Dublin survive? I will forego answering this question until we see whether Member States are willing to back a much more ambitious relocation system. If they are, then a significant chunk of the excess burden being borne by some Member States would be redistributed. Asylum-seekers would also benefit from better conditions and a fairer chance of getting protection as a result – although it would be best to take account of their preference in any relocation system as far as possible, so as to reduce ‘secondary movements’.
Sixth, the individual enthusiasm of refugee advocates should be harnessed as far as possible. Why not try to find money from the EU budget to pay those who are willing and able to give humanitarian or other support to refugees and asylum-seekers in an over-burdened Member State, or even a third State? (This would supplement the EU’s existing programmes for volunteers). Why not address a recommendation to Member States, encouraging private sponsorship of refugees to reduce the burden on taxpayers? As well as family and friends, corporations and NGOs should also be allowed to sponsor, and (for instance) universities could ask if students, staff and alumni wanted to sponsor refugees who would meet the criteria to enrol in courses.
Critics of refugee advocates often argue that those advocates should put refugees up in their homes – but many thousands of those advocates have in fact offered to do so. (For my part, let me reassure those panicking about the admission of refugees that I was not beheaded by my Middle Eastern refugee roommate at university). But let’s turn this argument around: why don’t critics of refugees spend a year in Syria, or in the refugee camps which many Syrians have fled to, before they criticise the refugees for leaving or others for wanting to welcome them?
This brings me neatly to the international framework. To repeat, burden-sharing in refugee matters is not just an EU principle, it’s a requirement of international law too. As regards Syrians, the neighbouring States (Turkey, Egypt, Jordan, and Lebanon) have borne a far heavier burden than the EU or any other countries have done. So it is entirely right for the EU to assist them more with this burden than it is doing already (the EU and Member States already give significant financial assistance).
While refugees in those countries may be safe from immediate persecution, they face difficult living conditions: see the further analyses here of the position in Turkey and the nearby Arab States. Moreover, food support from UN agencies has just been halved. The EU’s further support for these States does not have to be solely in the form of admission of refugees: it could make a further financial contribution to increase the refugees’ living standards. Nor does the EU alone have to offer this assistance: the international law rules apply across the globe, and the nearby Gulf states and other wealthy or middle-income countries could do more by way of taking refugees and offering financial assistance.
In the medium term, some more imaginative solutions are possible. One problem facing Syrian refugees in Turkey is that they are technically not refugees there, since Turkey applies the optional geographical limitation to the Geneva Convention, applying it to European refugees only. This means that many Syrians are not allowed to work. The EU could offer to waive visa requirements for Turkey in return for it lifting the geographical limitation (along with other conditions).
Another idea is the creation of a UN agency with the remit of the Syrian refugee crisis. This would mean that a special regime in the Geneva Convention and EU law would apply: if the agency could not take adequate care of Syrian refugees for any reason, they would be fully entitled to refugee status. Otherwise, their applications could be refused (see further the CJEU judgment in El-Kott, on the application of these rules to Palestinians). At the very least, the EU needs to take the initiative to hold a major international conference on the Syrian refugee situation, in order to encourage and coordinate many other countries’ efforts to take efforts to take a greater share of the burden of relieving this unfolding human tragedy.
Barnard & Peers: chapter 26