[David A. Martin is the Warner-Booker Distinguished Professor of International Law, University of Virginia.]
I start with a high-altitude view of the history and contours of refugee protection, to provide perspective on the current use of interdiction – and also on the contrasting stances taken by the U.S. Supreme Court in Sale v. Haitian Centers Council (509 U.S. 155 (1993)) and the European Court of Human Rights in Hirsi Jamaa v. Italy (Application no. 27765/09, Eur. Ct. H.R. 10 (2012)).
Refugee protection is not, at its core, dependent on fixed or expansive legal obligations of states or other political actors. Since Biblical times, refugees have been protected, even in the absence of treaty or legal edict. These were policy decisions by political leaders, influenced by compassion, but also by pragmatic considerations that ranged from assessments of absorptive capacities, food supplies, and the tolerance level of the leader’s subjects or fellow-citizens; through perceived advantages to be gained if the refugees seemed a particularly skilled or enterprising lot; to judgments about whether the exodus would strengthen or weaken the state against its enemies.
Protection that rests on policy is uncertain and unpredictable – by definition not wholly or even principally guided by humanitarian considerations. But it has nonetheless at many times afforded true shelter to multitudes. And after 60 years of operation under the major international treaties relevant here, we can hardly say that treaties assure reliable and consistent humanitarian responses either.
Comes now a new generation of tribunals and treaty bodies that seem to believe they can end the modern era’s inconsistency and usher in a virtually pure humanitarian practice of refugee protection. Their methods include reading selected provisions in refugee and human rights treaties quite expansively, deploying broad notions of a state’s jurisdiction – territorial and otherwise – and projecting onto the relevant treaties a mono-thematic conception of object and purpose, seeing only an objective to provide protection and issuing their interpretations accordingly.
This newly ambitious legal effort, most clearly exemplified in the Hirsi Jamaa decision, certainly is capable of improving protective outcomes for certain asylum-seekers in specific flight situations. But there are reasons for deep skepticism that this idealistic effort to carpet the entire field with judicially enforced legal prescripts will somehow overcome or crowd out all those benighted considerations of tawdry policy that state leaders tend to take into account. It may even interfere with efforts to optimize protection within the real-world constraints that government officials must account for. I offer here four reasons for skepticism.
Original intent and government buy-in.
First, this mono-thematic focus was not what states agreed to – not in 1951 or 1967, not in the European Convention on Human Rights. The treaties have a more complex set of objectives. Humanitarian shelter is one of them, to be sure, enough to motivate setting into law a floor reflecting modest but important commitments, focused on persons already present on the state’s territory – a floor, it must be emphasized, not a whole edifice. Other important treaty objectives, given only muted expression but still decidedly on the minds of the diplomats creating the treaties, were to preserve the core of national control over migration by foreigners and also to protect against criminal and security threats. There was no clear template for the mechanisms to be employed toward these ends, but states clearly regarded it as important to keep their protection commitments in balance with the sovereign right to make deliberate decisions about inbound migration. At each stage governments provided clear signals that they were not writing a “blank cheque” (a frequent assertion in the travaux of the 1951 Convention). They acknowledged that it would be good to do more, when possible, but they left the “more” in recommendations meant to inform ongoing policy decisions.
The dateline limitation in Article 1 of the 1951 Convention relating to the Status of Refugees may be the clearest indication of this caution (treating as “refugees” only those who fled “as a result of events occurring before 1 January 1951” – hence a largely known and finite population already present in the West European countries leading the drafting effort). But strong markers of the same attitude also appear in the exclusion from coverage of persons who had committed serious crimes, and especially the broadly worded exception to the nonrefoulement protection of Article 33 when “there are reasonable grounds for regarding” a person “as a danger to the security of the country” of refuge. Commentary at the time, even pieces written by scholars of strong humanitarian instincts, generally accepted that Article 33 did not include non-rejection at the frontier.
Some have suggested that the 1967 Protocol, which eliminated the dateline, should be seen as global acceptance of a more purely protective stance. This claim has to downplay the continuing exclusions from treaty coverage of serious criminals and national security threats. But there is an even stronger indication that the UN General Assembly, in adopting the text of the Protocol, was not propounding an absolute bar on interdiction or other barriers to arrival. That same year the General Assembly adopted a formal Declaration on Territorial Asylum (GA Res. 2312 (XXII), 22 U.N.GAOR Supp. (No. 16), at 81). Article 3 is widely quoted for its general provision that seems at first to bar “measures such as rejection at the frontier.” But the very next clause, often omitted by the commentators, states: “Exception may be made to the foregoing principle only for overriding reasons of national security or in order to safeguard the population, as in the case of a mass influx of persons.” The General Assembly clearly did not consider non-rejection at the frontier an absolute or nonderogable obligation for states. Governments still insisted on keeping in their hands certain tools to meter their obligations or keep them politically manageable, albeit at a higher protective level than in pre-treaty days. The Sale decision essentially recognized this tradeoff at the foundation of the legal obligations in the Protocol. The Court, like the Convention’s drafters, acknowledged that going further toward protection would be desirable in many circumstances – but that is a task for policy, not treaty obligation (509 U.S. at 188).
The insufficiency of alternatives.
Second, supporters of expansive readings of nonrefoulement sometimes counter that non-entrée policies, which include maritime interdiction but perhaps also strict visa regimes, are not necessary to serve the state’s interests in migration control and in stopping crime, terrorism, or spying. Instead those aims can be addressed in a far more precise and scientific way through careful application of the governing definition, which requires an asylum seeker to show a “well-founded fear of persecution” on account of five specified grounds, and which also excludes serious criminals and security threats.
This is an attractive claim, and it seems to accept that the treaties embody a more complex set of objects and purposes. But government officials generally are not reassured that such reliance will sustain the needed balance, because of hard experience with adjudication systems. In fact we still are not very good at accomplishing timely and accurate asylum decisions, with sufficient checks against fraud, even though intensive effort and millions of dollars have been devoted to improving adjudication since the surge in asylum applications in the 1980s. Moreover, as government officials see it, the problem is not just insufficient progress in procedures. The substantive standard also keeps expanding, now reaching far beyond fear of persecution or the five familiar grounds – especially under the ECHR jurisprudence.