© T.M.C. Asser Press and the authors 2015Mónika Ambrus and Ramses A. Wessel (eds.)Netherlands Yearbook of International Law 2014Netherlands Yearbook of International Law4510.1007/978-94-6265-060-2_9
9. Temporary Protection: Hovering at the Edges of Refugee Law
International Institute of Humanitarian Law, San Remo, Italy
9.2.1 Time as ‘Attachment’
9.2.2 Time as Deadline
9.2.3 A Clash of Times
9.3.1 A Time of Crisis
9.3.2 After the Crisis
9.4 Hovering on …
As a subject of international law, the refugee is inherently temporary: refugee status exists in order to fill the gap caused by a breakdown of the normal bond between citizen and state with ‘international protection’, until this bond can be restored, either with the original state of nationality, or with another state. While this ambition is clear, the practice of refugee protection under the current, post-1951 regime has exposed the serious problem international refugee law (IRL) faces with regard to the tail-end of protection, also known as ‘durable solutions’. This chapter studies the most prominent attempt made by UNHCR and states within the refugee regime at re-invigorating the temporary character of international protection—namely the mechanism known as ‘temporary protection’. While the concept can be traced to formulations of ‘temporary refuge’ in the 1980s, temporary protection (TP) truly emerged as a term of art in the 1990s, as Western Europe was faced with a large-scale influx of forced migrants from the former Yugoslavia. In 2001, TP was the subject of an EU directive, which partly clarified the relationship of TP to mainstream IRL—notably whether TP should be seen as a substitute for, or a prelude to, the operation of the 1951 Refugee Convention. Ambiguity has remained the hallmark of the TP concept, however, not least in UNHCR’s attempts at formalising a TP regime outside the European ambit. After examining the main features of these attempts, the chapter concludes that, while the EU directive should be taken seriously, a continuing doctrine of temporary protection outside established IRL is both legally unsound and politically unconvincing.
KeywordsRefugeeInternational protectionTemporary protection1951 ConventionCessationRepatriationDerogationEmergency
Programme Director, International Institute of Humanitarian Law, San Remo, Italy; Research Associate, Refugee Studies Centre, University of Oxford, UK.
In the words of two leading scholars in the discipline, international refugee law is ‘an incomplete legal regime of protection, imperfectly covering what ought to be a situation of exception’.1 This finding is doubly relevant to the overarching theme of this volume of the Netherlands Yearbook of International Law. Indeed, as this chapter will show, the time dimension of refugee status at international law is the source of both its acceptability as a regime of exception, and of some of the most glaring gaps in that regime. As a subject of international law, the refugee is inherently temporary: refugee status exists in order to fill the gap caused by a breakdown of the normal bond between citizen and state with ‘international protection’, until this bond can be restored, either with the original state of nationality, or with another state. ‘The claim for inclusion in a host community is … a consequence of exclusion by the host community’.2 The international protection of refugees has therefore been described as surrogate protection, a ‘back-up to the protection one expects from the state of which an individual is a national … meant to come into play only in situations when that protection is unavailable’.3
It is fairly obvious that, within the Westphalian social order that underpins modern public international law, surrogate protection is meant to be exceptional. This exception tends to be recurrent, however, and its regulation is as old as that of the ‘normal’ condition—the protection exercised by the state of nationality—from which it derogates. Asylum has a long and distinguished history indeed, both in practice and in doctrine.4 Contemporary international refugee law (IRL), however, is a rather under-developed segment of the vast body of post-WWII, UN Charter-based human rights and humanitarian law. Its customary basis is thin, consisting mainly of the rule known as non–refoulement, which ‘prescribes, broadly, that no refugee should be returned to any country where he or she is likely to face persecution, other ill-treatment, or torture’.5 Whether non–refoulement, in either this or another formulation, has attained the status of a principle of customary international law is actually debated among scholars.6 Treaty law of universal vocation, which at this time binds over 140 states, is limited to one Convention and one Protocol, supplemented by one refugee-specific instrument at the regional (namely, African) level.7 The 1951 Convention has definitely proved its resilience, having survived countless political crises, and offered relief to millions of refugees, since its adoption. It must be noted, however, that the bulk of its provisions regulate the status of refugees once admitted, and need to be transposed into the Parties’ domestic legal order in order to be effective. International cooperation, while crucial to the functioning of the international refugee regime, is referred in non-binding terms only in the Convention’s Preamble.8 An international organisation—the Office of the United Nations High Commissioner for Refugees (UNHCR)—was established in 1950 and vested with supervisory powers over the implementation of the Convention. It was also given a broad mandate to ‘provid[e] international protection, under the auspices of the United Nations, to refugees’ and to seek ‘permanent solutions for the problem of refugees’.9 The personnel of UNHCR, as well as the agency’s Executive Committee, have been instrumental in developing, over the years, a considerable amount of soft law in the form of conclusions, declarations and guidelines on a wide range of issues, including those on which treaty law is silent.10
It is interesting to note that UNHCR was set up for an initial period of three years.11 This is symptomatic of a cautious mood, which can be traced back to the League of Nations’ origins of the modern refugee regime. In the inter-war period, refugee problems tended to be defined and addressed one at a time.12 The Convention itself, with its time limitation and optional geographic reservation, does not represent a major departure from past approaches: it looks mainly towards the past.13 On the other hand, it contains an objective, non-situational definition of ‘refugee’, which clearly allowed the instrument to deal with a new refugee problem arising in the early 1950s—namely, fugitives from the Soviet bloc. Still, it is only through the adoption of the 1967 Protocol14 that the Convention became truly forward-looking, so that the legal regime instituted in 1951 could apply to (potentially) every new refugee situation. The ambition of the post-Protocol regime is huge: for every refugee in the world, at any time, there must be a state providing asylum. What is surprising is that this open-ended commitment was made without the backing of any fresh legal tools: no formula was proposed to determine how states should share responsibilities and otherwise co-operate, not only with regard to asylum, but also with a view to finding ‘permanent solutions for the problem of refugees’.15 This lacuna proved to be particularly problematic as, during the same years, many states in the developing world—starting with newly independent African states—acceded to the Convention and Protocol. With the multiplication of refugee crises in the global South, the modalities of responsibility—and burden-sharing in the international regime—changed radically: ‘physical’ burden-sharing, i.e. the distribution of refugee caseloads through relocation and resettlement, became the exception, while fiscal or financial burden-sharing became the rule, largely governed by donor states in the exercise of their discretion. In the absence of parameters to define what a ‘fair share’ may be, complaints of being short-changed are likely to be heard each time a new refugee situation appears on the map—especially where large numbers are involved.16
How to organise international cooperation and to assign protection responsibilities is even more problematic where frontline states are not party to the Convention. Conclusion No. 22 of UNHCR’s Executive Committee, adopted in 1981, purports to ‘reaffirm the basic minimum standards for [the] treatment’ of ‘asylum seekers forming part of large-scale influxes’. These standards are ostensibly less generous than the provisions of the 1951 Convention.17 However, this Conclusion must be read in context, namely the continuing outflow of refugees from Vietnam, whose rickety boats reached—if they were lucky—the shores of non-party states.18 The exodus of ‘boat people’ also provided the background for the emergence of a new concept in refugee doctrine, namely: ‘temporary refuge’. In the mid-1980s, Perluss and Fitzpatrick Hartman, subsequently supported by Goodwin-Gill, argued that a customary norm of temporary refuge existed, which prohibited the forcible repatriation of ‘refugees’ in a broader category than as defined by the 1951 Convention. Their argument was based in part on a general duty to rescue people in distress, and in part on an enlarged concept of non–refoulement. 19 This doctrine, which some commentators regard as a precursor to ‘temporary protection’,20 is better seen as an attempt at establishing, even for non–party states, a duty to admit refugees on their territories until a solution can be found elsewhere. Recourse to customary norms was a response tailored to a crisis affecting a region that was (and still is) by and large outside the universal legal regime. The refuge expected from Southeast Asian states for the Vietnamese refugees was temporary ‘not because the causes of flight were thought to be of short duration’, but because it included, on the part of Western states, ‘the promise of resettlement’.21
This chapter will primarily deal with the responsibilities of States that are party to the 1951 Convention (as amended by the 1967 Protocol) towards refugees in, or seeking entry into, their territories. Though international cooperation—or rather the lack thereof—looms large over this discussion, the focus of analysis will remain on what the Convention dictates, rather than on what it omits. The question we intend to address can be formulated as follows: how does international refugee law manage ‘refuge time’? In other words, borrowing from the dual mandate of UNHCR, how does it regulate the transition from ‘protection’ to ‘solutions’? After setting the general parameters of this discussion in the next section, we shall focus on the most prominent attempt made by UNHCR and states within the refugee regime at re-invigorating the temporary character of international protection—namely the mechanism known as ‘temporary protection’ (TP) as it emerged in Europe in the 1990s. The title of this chapter is a tribute to the late Joan Fitzpatrick, a leading thinker on both temporary refuge and TP, whose work will be amply cited herein.22 Her suggestion that TP continues to ‘hover … at the edges of refugee law’23 will lead us to clarify the relationship of TP to the 1951 Convention, then, in the final sections of this chapter, to put into question the continuing relevance of a doctrine of temporary protection outside the Convention-based regime.
9.2 Time in the International Legal Regime
The 1951 Convention manages ‘refuge time’ in two ways. On the one hand, it establishes a gradation of rights, the accrual of which is contingent upon, inter alia, the length of stay on territory. On the other hand, the Convention makes it clear that refugee status is not meant to last forever, and it provides for the circumstances in which it shall terminate. In the following we deal briefly with each of these aspects in turn, before exposing areas of tension between the two.
9.2.1 Time as ‘Attachment’
The content of refugee status under the Convention is determined by the superposition of ‘two cumulative conditions governing, first the criteria of entitlement (the applicability of the norm) and, second, the standard of treatment’.24 Entitlement criteria distinguish three levels of applicability: a core set of basic guarantees apply to ‘refugees’ without any further qualification25; those who are present within the territory benefit from additional rights, especially if they are lawfully present26; finally, residence entitles refugees to such rights and benefits as administrative assistance, access to wage-earning employment or liberal professions (where the residence is lawful) or the protection of intellectual property (where the residence is both lawful and habitual).27 Once these criteria are fulfilled, the precise content of the norm is determined by reference to the corresponding entitlements of ordinary aliens, most-favoured foreigners, or nationals of the country of asylum, as the case may be. One can agree with Chetail that ‘the legal regime deriving from the superposition of various entitlement criteria with different standards of treatment remains extremely complex’, and that the task of ‘finding a cogent rationale in such byzantine gradation’ is not an easy one.28 A rationale can be found, however, once this complex architecture is recognised as an ‘incremental protection regime’ following the ‘intensity of the territorial bond’ between the refugee and her State of asylum. ‘In sum, the longer the refugee remains in the territory of the State Party, the broader the range of entitlements.’29 In the same vein, Hathaway refers to ‘levels of attachment’ and equates refugee status under the Convention with an ‘assimilative path’.30 In sum, the logic of the Convention is to provide refugees with rights and obligations that over time get gradually closer to those of citizens of the host state, with naturalisation as the recommended outcome.31
9.2.2 Time as Deadline
Until the refugee is naturalised (which is neither mandatory nor extremely common in state practice worldwide),32 she remains protected by her special status. However, to suggest that this status might last longer than necessary would run counter to the other objective of refugee law, which is to resolve the ‘anomaly’ represented by refugeehood and surrogate protection in the international social order. The Convention provides, therefore, for a number of circumstances under which its special protection ‘shall cease to apply’.33 To start with, an individual refugee may at any time, through her own actions, indicate that international protection is no longer required because she no longer experiences a fear of persecution. Cessation of refugee status thus flows from circumstances such as voluntary repatriation and re-establishment in the country of origin, or the acquisition of a new nationality and the protection which derives therefrom.34 These cessation clauses based on voluntariness will be discussed in this chapter only in relation to the doctrine of voluntary repatriation. It is rather the cessation clause in Article 1C(5) of the Convention that will be focused on because, under this provision, the will of the refugee yields to objective circumstances, and the protection of refugee law comes to an end if the refugee ‘can no longer, because the circumstances in connexion with which [s]he has been recognized as a refugee have ceased to exist, continue to refuse to avail [her]self of the protection of the country of [her] nationality’.35
Even though this cessation clause responds to a legitimate state concern about migration control, including the prerogative of deporting aliens whose stay is no longer justified,36 states Parties rarely invoke it. There are, according to Fitzpatrick, three main reasons for this. The first has to do with administrative convenience, namely ‘[t]he factual predicates for cessation may be too resource-intensive to satisfy’, requiring lengthy procedures for the (re-)examination of individual cases. So ‘[w]hy divert scarce resources to remove persons who do not pose a threat and who may be productive members of the community?’37 This question already points to the second reason, which is that ‘many asylees obtain a more enduring immigration status prior to the occurrence of major political changes in their state of origin.’38 As time goes by, the development of ties with or within the country of asylum tends to become more relevant to the refugee’s right to remain than the changing conditions in the country of origin.39 Finally, it is evident that ‘the circumstances in connexion with which’ a refugee has been recognised are seldom purely ‘individual’. The same set of circumstances (say, persecutory practices against a particular ethnic or religious group) force large numbers of potential victims out of their countries in search of asylum, and where many refugees of the same nationality find themselves in a same asylum state, cessation under Article 1C(5) of the Convention may entail the repatriation of a substantial number of aliens—something which states are ‘often reluctant for political reasons to force’,40 let alone the administrative and logistical difficulties involved.41
Whether the initiative stems from the state of origin or from asylum states stricken with compassion fatigue, the fact is, with regard to the application of Article 1C(5) of the Convention, that states prefer to act collectively in declaring an entire refugee situation terminated. To this end, they tend to either seek, or follow, the advice of UNHCR.42 In practice, such instances of ‘group’ cessation are infrequent, and they invariably occur many years, at times decades, after the emergence of a refugee situation.43 UNHCR explains that a ‘premature or insufficiently grounded application of the ceased circumstances clause can have serious consequences’, including the severing of ties to family, social networks and employment in the host country, as well as ‘persons being compelled to return to a volatile situation’.44 It recommends, therefore, that (i) cessation be applied only once it has been ascertained that ‘conditions within the country of origin have changed in a profound and enduring manner’45; and (ii) procedures for determining general cessation be ‘fair, clear and transparent’.46 Furthermore, UNHCR’s Executive Committee has declared it to be a ‘general principle that all refugees affected by general cessation must have a possibility, upon their request, to have such application in their cases reconsidered on international protection grounds relevant to their individual case’.47 Durieux and Hurwitz have argued that these strict requirements ‘are a source of frustration for States’, especially for states in the global South that are faced with large refugee populations whom they are unable and/or unwilling to integrate. As a result, states ‘tend to take the legal and policy battle to a seemingly more open field, namely the doctrine of voluntary repatriation’, which is discussed in the next section.48
9.2.3 A Clash of Times
The management of ‘refuge time’ is the source of an obvious tension within the 1951 Convention regime. Premature return to a continuing situation of persecution or violence is clearly prohibited by the non–refoulement principle. Changes in the country of origin are usually slow to come about, and take much longer to stabilise. Meanwhile, attachment leads to integration, which becomes the default solution. In the words of Coles, refugee law has ‘tended to be biased in favour of local integration rather than repatriation’.49 Of course, it can be argued that the practice of Western states in implementing the Convention during the Cold War willingly aggravated such a bias, as it was then politically convenient to welcome and integrate the ‘champions of freedom’ who had run away from Communist rule—and the Iron Curtain made sure that the numbers would remain manageable. Yet, it is true that the letter of the Convention, coupled with UNHCR’s cautious approach to ‘changed circumstances’, creates an imbalance in respect of how cessation intervenes. Individual cessation of refugee status through voluntary repatriation is foreseen, but the Convention is silent about the transition from refugee status to reinstatement of the protection of the state of origin—i.e., about how to prepare (refugees, as well as the state of origin) for voluntary repatriation. Historically, this gap has been filled on a situation-by-situation basis, through the adoption of so-called ‘tripartite agreements’ expressing the commitments of one or several states of asylum, the state of origin/return, and UNHCR. Besides guaranteeing the voluntary character of repatriation, these agreements usually provide for conditions of safety upon return, non-penalisation for having sought and enjoyed asylum abroad, amnesties and the like.50 They also guarantee UNHCR’s right of access to returnees in order to monitor the consequences of repatriation. While such tripartite agreements have been instrumental—especially in Africa—in the affirmation of binding principles in the repatriation context, their limitations are equally evident: they are restricted to a particular movement of return and usually time-bounded; they tend to intervene late in the ‘life’ of a refugee situation—albeit usually well before cessation under 1C(5) of the Convention;51 and they rely heavily on the will (or at least the willingness) of refugees to return.
When refugees do not feel inclined to return, how far can the host state (or UNHCR, for that matter) go in order to change their minds—i.e., in UNHCR jargon, to ‘promote voluntary repatriation’? The 1951 Convention does not address this question.52 Its terms suggest that (i) fully voluntary, i.e. spontaneous and un-coerced repatriation is the best way of effecting return; and (ii) return to the country of origin regardless of the refugees’ will—or squarely against it—is only possible after invocation of the cessation clause in Article 1C(5) of the Convention. However, as noted by Durieux and Hurwitz, state practice paints ‘various shades of voluntariness (or, conversely, of coercion) … [b]etween these two clear-cut scenarios’.53 ‘Increasingly, “safe return” has become part of the policy thinking of governments’, occupying an ‘interim position’ between the refugee deciding to go back home on her own free will and the host state exercising its deportation prerogatives against overstaying migrants.54 The literature on the tension between ‘safe’ and ‘voluntary’ return—short of cessation—is too vast to be aptly summarised here.55 Suffice to note, for our purposes, that the parameters of the debate have shifted over time. As early as in the 1980s in Africa, the ‘safe return’ discourse aimed at dealing with ‘unreasonable’ objections to return, arising either within a scenario of promotion of voluntary repatriation, or among a ‘residual’ caseload hanging on after the majority of the refugees had returned. With the Cold War, however, the West discovered the appeal of repatriation as the ‘ideal’ solution to refugee problems. To take the state of origin to task and to engage its protection responsibilities vis–à–vis its exiled citizens became an objective in itself, alongside reducing the burden on asylum states. In a dramatic reversal of the former ‘exilic bias’ of the international regime, international protection could be thought of as a simple way-station on a course that had started in state X and would inevitably end up there, too. The irruption of TP into the policy discourse of European states and UNHCR in the early 1990s made it urgent to address the question as to whether this conception of international protection could be accommodated within the Convention-based regime, or could only be designed outside it. It is to this question that we now turn.
9.3 Temporary Protection in Europe, 1992–2001
Soon after the end of the Cold War, the violent break-up of Yugoslavia and the attendant egregious violations of human rights and humanitarian law constituted ‘a watershed in the formalization of temporary protection’.56 As massive flows of refugees, including many victims of ‘ethnic cleansing’ policies, started pouring into Central and Western Europe, the interest of scholars and policy-makers in temporary refuge and other ‘flexible’ and ‘pragmatic’ protection mechanisms, which had flagged in the 1980s, rose again with a vengeance. Precisely what kind of mechanism TP was, and what needs it was meant to meet, was far from clear at the outset—and a fair amount of confusion prevailed over both doctrine and practice throughout the decade.57 What matters the most, though, is that—unlike temporary refuge in Southeast Asia—TP was conceived by and for States that were all bound by the 1951 Convention, and it directly challenged either the Convention’s letter, or its spirit, or both. This challenge was directed very specifically to the time dimensions of the Convention, which we have outlined in the previous section. In essence, TP was conceived as a return–oriented protection mechanism, within which ‘the focus on return as the most appropriate solution’ provided the rationale for ‘standards of treatment which emphasize the provisional aspect of the refugees’ stay … and minimize, at least in the initial stages, efforts to promote integration’.58
9.3.1 A Time of Crisis
Writing in 2000, Fitzpatrick aptly noted that ‘[r]ecent European practice has emphasized circumstances of arrival in designating the beneficiaries of temporary protection, rather than the causes of flight’.59 The most relevant of such circumstances are definitely the size and suddenness of the influx. This is an important observation, echoed by Durieux and Hurwitz who ascertain that ‘two distinct issues, the size of the influx and the personal scope of protection, were merged into one’.60 As is often the case in mass influx situations finding their origin in generalised armed conflict, European states receiving asylum seekers from the former Yugoslavia could not be sure that every individual in the group was a bona fide refugee as defined in the 1951 Convention. Nonetheless, the predominantly refugee character of the flow could not be denied. When the High Commissioner for Refugees for the first time called for ‘a flexible system of temporary protection’ on behalf of ‘persons fleeing former Yugoslavia’, she was explicitly referring to people fleeing ‘[s]ystematic expulsions, forcible relocation, assassinations and other forms of persecution … for no other reason but their national, ethnic or religious origin’.61 Those were, simply put, Convention refugees. The challenge, faced by UNHCR and prospective asylum states, did not, therefore, reside in the causes of flight, but rather in the size of the (actual and potential) refugee caseload, and in the nature of the conflict. As regards the former, Kjaerum aptly observed that ‘[i]n Western Europe today … the major reason for the introduction of temporary protection is the inability of the European countries to co-ordinate their efforts sufficiently to allow for equal burden-sharing’.62 Lack of solidarity was indeed what made this refugee situation an emergency—i.e., a situation overwhelming the resources normally assigned to refugee status determination and refugee protection. Regardless of the numbers involved, a large-scale influx of refugees into several wealthy European states need not create an emergency, so long as responsibilities are shared among those states. Sadly, ‘burden-sharing was a dismal failure during the Yugoslav crisis’,63 which allowed those states that were most exposed, such as Germany, to seek the same kind of justification as states in the developing world for a less than welcoming attitude towards new and prospective asylum seekers.
As regards the latter issue—the specificity of the conflict in the former Yugoslavia—, High Commissioner Ogata’s words deserve close scrutiny: ‘I believe the nature of the conflict is such that a flexible system of temporary protection would respond adequately to the emergency situation and encourage return as the most desirable and feasible solution.’64 There seem to be two interwoven rationales in this sentence. In one sense, the nature of the conflict refers to the ‘new – horrifying – twist’ of the conflict in the former Yugoslavia, where ‘[d]isplacement seems to be the goal, not just the result of the war, with the motive clearly being ethnic relocation’.65 This characteristic made it morally compelling to reverse ethnic cleansing, which can only be achieved through ‘return home’. However, there is also a hint in the High Commissioner’s statement that this result can be achieved promptly—that the nature of the conflict is such that return following temporary refuge, that is, early return, looks like a ‘feasible solution’.66 Thus, Ogata’s move to encourage return as ‘the most desirable and feasible solution’ was based on a mix of moral argument and wishful thinking. On the side of moral argument,
[r]eturn … fundamentally … will require the reaffirmation of the basic human right to return home in safety and dignity and for an end to all practices and actions which force people out of their homes and block the prospects of return. I consider this point to be of crucial importance. Otherwise, measures to protect and assist refugees will amount to complicity in the policy of ‘ethnic cleansing’.67
The moral duty not only to end but indeed to ‘reverse’ ethnic cleansing came to justify and deepen European States’ reluctance to allow refugees to integrate. As for the wishful thinking part, it was captured well by Fitzpatrick: ‘Duration figures largely in the conceptualization of TP, though it is often impossible to predict the time frame of refugee crises’68—not least the crisis that was then unfolding in the midst of Europe. There was an unspoken assumption in the promotion of TP for former Yugoslavia refugees, that because they had fled precipitously they would be able to return promptly. As UNHCR noted in 1994,
[o]ne of the principal reasons for applying the term ‘temporary’ to protection given to persons fleeing conflicts or acute crises in their country of origin is the expectation – or at least the hope – that international efforts to resolve the crisis will, within a fairly short period, produce results that will enable the refugees to exercise their right to return home in safety.69
The UNHCR note candidly acknowledged that ‘[a]s the situation in Bosnia and Herzegovina shows, hopes for an early safe return are not always realized.’ It insisted, nonetheless, that the provision of TP had to be seen ‘as part of a comprehensive programme of concerted international action that includes prevention and solutions’—echoing the High Commissioner’s 1992 tabling of a ‘comprehensive humanitarian initiative in former Yugoslavia’ consisting of seven elements.70 The implicit (political) message to states reads: if you work harder on solutions (as well as prevention), you may be justified in offering a lesser quality of protection.
Indeed, quality and duration of protection are intimately inter-connected in the conceptualisation of TP. Precisely, TP involves putting brakes on the ‘integrating machinery’ of the Convention because of an expectation that refugees will return home within a near future: duration of sojourn is ‘linked to standards of treatment’ where ‘social and economic benefits facilitate local integration and deter repatriation’.71 With the withholding of such benefits, Kjaerum has argued, ‘a signal is sent to the refugee that his or her stay in the specific country is only temporary.’72 This echoes Fitzpatrick’s observation that TP regimes ‘may be consciously structured to cultivate [the refugees’] natural desire for repatriation and to encourage the recipients to conceive of exile as limited in time.’73 In sum, what European states expected from TP was a protection regime that, instead of facilitating local integration and deterring repatriation (which ‘normal’ application of the Convention was deemed to involve), would in fact deter local integration and facilitate repatriation.
184.108.40.206 The Quality of Protection
European states’ approaches to the contents of temporary protection—i.e., basically, what rights and benefits should be withheld—varied significantly, and there is no room in this chapter for a detailed comparison of national practices.74 It appears, however, that restrictions were mainly introduced in three areas—not surprisingly, those with the highest integration potential. First, the ability of new arrivals to lodge an asylum claim was severely curtailed in many states, and purely and simply excluded in a few. In terms of status, this had the effect of creating a category of protection-seekers below that of formal asylum seekers, i.e. those persons whose claims are being processed. From an international law point of view, the practice of denying access to an asylum procedure raised an even more serious problem, since states Parties to the 1951 Convention owe to each other, as well as to refugees, a duty to assess, whenever a person seeks the protection of the Convention, whether this request is well founded or not. The right to seek asylum is proclaimed in the Universal Declaration of Human Rights,75 and UNHCR warned—both during and after the crisis—that TP ‘should not become a substitute for the right of refugees to seek and enjoy asylum in accordance with internationally agreed standards.’76 On the whole, though, restrictions of this fundamental right were tolerated, to wit a document emanating from UNHCR’s Standing Committee in 1997, which, among several ‘procedural models’, envisaged an acceptable situation in which ‘States deny or suspend access to individual status determination during the stay in the country of refuge.’77 Whether or not they were allowed to lodge an asylum claim, ‘beneficiaries’ of TP were often denied the right to work—even in those countries that allowed asylum seekers to engage in wage-earning or self-employment pending a decision on their claims. The objective was clearly to avoid economic and social integration by cutting the refugee from the labour market, even though—in some states at least—the numbers were so high that they could not, in the short term, be absorbed by that market anyhow. Furthermore, the ‘return orientation’ of TP was manifested through an array of restrictions and adjustments to the right to education (notably by obliging refugee children to follow classes in their own native language); and to family reunification.78
220.127.116.11 The Duration of Protection
Cessation of TP, as Fitzpatrick astutely observed,
has a dual aspect. First, TP may be terminated, and its beneficiaries repatriated, when conditions in the state of origin permit return in safety and with dignity. Second, TP should be terminated, and a solution not involving involuntary repatriation provided, when danger persists in the state of origin and TP beneficiaries can no longer tolerably be maintained in uncertainty or under standards of treatment that fall substantially short of the 1951 Convention.79
In this sub-section we discuss these two distinct ‘cessation’ issues in turn.
For those States—most prominently Germany and Austria—that attached particular importance to TP as a prelude to repatriation, the defining factor triggering return had to be the end of armed conflict in the Western Balkans. In several jurisdictions, TP for Bosnians was extended from year to year, or for six months at a time, and was meant to terminate around six months after the end of hostilities. UNHCR chose to maintain some ambiguity over the issue, suggesting in 1994 that
if conditions in the country of origin change sufficiently for the better to make possible the refugees’ return in safety and dignity, arrangements can be made, in consultation with UNHCR, for [TP] to be phased out, ideally through voluntary repatriation.80
Then peace broke out in Bosnia—through the signing of the Dayton Peace Agreement (DPA) in November/December 1995—, forcing all stakeholders to show their cards and, if possible, to agree on criteria for refugee return. Germany, fixated on the end of hostilities and nothing else, refused to extend its deadline for TP expiry beyond August/September 1996, with plans to return 200,000 Bosnians by July 1997.81 However, UNHCR was able to convince all other refugee-hosting states that deadlines had better be accompanied, if not replaced, by safety benchmarks. Pursuant to the DPA, UNHCR was to develop a plan allowing for an ‘early, peaceful, orderly and phased’ return of refugees and displaced persons.82 By linking the phasing of return to implementation of the commitments assumed by the DPA Parties, UNHCR cleverly construed ‘peaceful and phased’ to actually mean ‘safe’, and to qualify the adjective ‘early’. Although the Repatriation Plan’s three phases were dated,83 the deadlines could only be indicative since repatriation could only occur once TP was lifted, and TP should be maintained until (i) full implementation of the military aspects of DPA; (ii) proclamation of the required amnesty; and (iii) the establishment and functioning of mechanisms for the protection of human rights—all of these being Dayton commitments. Furthermore, UNHCR recommended that TP should not be lifted until the OSCE had certified that conditions were satisfactory for elections.84 It also urged that persons unable or unwilling to return to their place of previous residence due to changes in territorial control and in the ethnic composition of the region concerned should be the final category considered for return, owing to their special vulnerability—clearly, another mechanism of protection through phasing. Last but not least, UNHCR warned that some refugees would be in need of continuing protection—without, however, explaining how this need could be allowed to surface, or be assessed.
The ‘elephant in the room’ within the Repatriation Plan was, of course, the existing legal framework based on the 1951 Convention, which the peace accords could not possibly displace. True to its promise of a ‘flexible’ protection system, UNHCR went out of its way to address all protection issues involved in the lifting of TP and ‘early’ return without referring to the Convention. This conscious effort notwithstanding, what shows through clearly in UNHCR’s Plan is the ‘spirit’ of the Convention, as well as UNHCR’s reluctance to abandon the relatively safe legal grounds on which it had, over the years, developed its doctrine and built its credibility. The rather ambiguous debate over ‘safe return’ that followed the signing of the DPA confirmed Fitzpatrick’s intuition that TP termination as a prelude to repatriation ‘involves issues analogous to those that arise under Article 1C(5) for Convention refugees’. She saw this as rather problematic, since ‘[w]ithout a promise that termination of TP will be simpler and more automatic than cessation of Convention refugee status … the incentives for states to adopt binding legal standards on temporary protection are diminished substantially’.85 UNHCR was clearly not prepared to concede any automatic cessation, because it knew that in the minds of states termination of TP was synonymous with (almost) immediate return, and the safety of repatriation could not be assessed on the basis of a single ‘event’—not even the signing of a peace agreement. Could the agency offer a simpler process of cessation—albeit without the name? Actually, the only important difference between the safety benchmarks in the Repatriation Plan and those generally read into Article 1C(5) of the Convention is that the latter can be assessed with relative calm, years after the end of the crisis, and often with the results of a ‘live test’ in hand—namely, the level of reintegration of spontaneous and other voluntary returnees. In contrast, where return is contemplated soon after the brokering of a fragile peace, basic conditions of safety must be not only assessed, but often actually engineered, so to speak, ‘on the run’. Beyond basic safety, the ‘profound’ and ‘enduring’ nature of the changes in the country of return cannot be objectively assessed—it becomes a matter of faith, hope, or mere speculation. Time spent waiting for durable, stable change is a luxury, which many states having espoused TP thought they could not afford. This may have made ‘termination of TP’ a simpler decision to make, but definitely not a simpler process overall. The risks involved in a premature or ill-conceived decision—that repatriation might not take hold, that minority returns might re-ignite conflict, etc.—were not eliminated by the decision to effect ‘early return’. Those risks were simply ‘repatriated’ along with the refugees, and they had to be managed post-return, inside Bosnia, by a host of political, human rights and humanitarian actors whose reports had to be analysed and acted upon on a regular basis.86 The jury is still out on the question whether, on the whole, termination of TP for Bosnians resulted in a lesser financial or political burden on European states, as compared to what would have happened if ‘normal’ cessation rules had been followed.