On March 31, 2003, thousands gathered outside the Bosnian town of Srebrenica for a mass funeral, commemorating more than 7,000 men and boys slaughtered when the Serbs attacked this U.N.-designated "safe area" in 1995. The massacre highlights the dubious value of safe-haven operations throughout the 1990s. However, in a world with a high civilian casualty rate (nine of 10 casualties in recent conflicts, compared with only one of 10 at the beginning of the 20th century), it is crucial to examine whether safe-haven operations can still be useful in protecting the lives of Internally Displaced Persons (IDPs) and would-be refugees.
Even though the application of safe-haven policies in the 1990s has been highly problematic, to say the least, there is still a clear role for a restructured system of safe havens, closer to the original definition in the 1949 Geneva Conventions. The reliance on asylum policy only is unsustainable. This is not only due to Western governments' growing reluctance to offer asylum, it is also linked to the sheer numbers of people in danger and the fact that this policy excludes many from its protection.
In this discussion, we will:
- set the discussion of safe havens within the context of international refugee law and recent policy;
- refer back to the original concept of what safe havens were intended to be;
- examine the application of the concept during the 1990s by looking at three case studies: Operation Provide Comfort in Iraq, the six Bosnian safe havens, and Operation Turquoise in Rwanda (these are presented in separate essays);
- summarize the crucial elements raised by the case studies, and attempt to understand the positive and negative elements of past safe-haven experiences; and
- assess the usefulness of safe-haven policy and consider the options for a restructured safe-haven policy.
'Safe Havens' in the Context of International Refugee Law
Nancy Ferrell discusses the importance of trust and safe spaces for communication within families.
The 1951 United Nations Refugee Convention spells out the obligation of states toward refugees and the rights of refugees. It stemmed from the realization, after the Second World War, that states could help citizens persecuted in other countries by offering them protection. Article 33 of the Convention develops the concept of non-refoulement, by stating that "no Contracting State shall expel or return a refugee in any manner whatsoever to the frontiers of a territory where his life or freedom would be threatened on account of his race, religion, nationality, membership of a particular social group or political opinion."
The United Nations High Commissioner for Refugees (UNHCR) has been the leading agency in regard to refugee issues for the past 50 years. It has promoted three ways to protect refugees:
- voluntary repatriation;
- local settlement; and
- third-country resettlement.
All these policies have taken exile as their starting point, and the main policy used to protect refugees has generally been asylum in third countries. The concept of non-refoulement is central to refugee protection and is at the basis of asylum policy. A state is obliged, under international law, to accept within its borders refugees in need of protection and thus to grant them asylum. However, in recent years, the reliance on asylum policy as the main protection tool has been questioned, since governments in third countries have tended to shy away from their legal duty.
Mary Anderson describes how one "scales up" interventions from the personal level to the institutional, structural, and societal level.
Northern countries, which have historically tended to grant refugees permanent resident status, have withdrawn from their obligation. Largely due to the political cost linked to high immigration rates, governments have started perceiving asylum policy as a 'back door' to immigration. They have thus developed a process of external deterrence (non-entr'e). Northern countries have violated the non-refoulement doctrine by establishing restrictions such as visa requirements, 'international zones' in French airports, or committing even more flagrant violations, such as Italy turning away a whole boat full of Albanian refugees. The recent proposal by the United Kingdom to 'outsource the asylum' process outside the European Unionand the statements made by Prime Minister Tony Blair promising that he will halve the number of asylum seekers in Britain by September 2003, should be seen as yet another attempt to strengthen 'Fortress Europe.' Countries that border refugee-producing conflicts have been less likely to block access to their borders. However, due to a severe lack of resources they have become increasingly resistant to refugee flows.
The current system of protection for refugees based on asylum is thus under threat. Asylum policy as the main method to protect refugees is weakened because states are less willing to accept refugees within their borders. One could also argue that the current understanding of who a refugee is, and thus who should receive protection, is very restrictive within this framework. States have interpreted the 1951 Geneva Convention in a very limiting light and have thus excluded many people from their protection. For example, people suffering from non-state persecution are not taken to be eligible for refugee status and thus for protection.
In the same context, one could highlight the plight of Internally Displaced Persons. IDPs are "persons who have been forced or obliged to flee or to leave their homes or places of habitual residence, in particular, as a result of, or in order to avoid the effects of, armed conflict, situations of generalized violence, violations of human rights, or natural or human-made disasters, and who have not crossed an internationally recognized state border."
IDPs are not eligible for refugee protection status because they have not crossed an international border, whether due to distance from frontiers, increasing restrictions, or simply by choosing to stay close to their roots and culture. Nonetheless, IDPs represent 20 to 25 million people, and thus a significant proportion of people who do not receive protection under the current asylum regime (there are actually more IDPs than refugees). The Center for Disease Control states that during emergency situations, the people who have been forcibly displaced within their own country have the highest death rate. These people are denied the right of citizens to be protected by their governments, just like refugees, and thus their protection should be of the highest priority.
Current asylum policy undeniably saves countless lives, but due to increasing reluctance from governments to open their frontiers it is clear that this policy cannot deal with the massive influx of refugees, nor does it help to protect individuals stuck in their countries. It is within this context that the idea of the right to remain has emerged. The UNHCR developed the idea that refugee movements could be averted if action was taken to reduce the threats that force people to leave their countries. The policy of safe havens that was implemented in different countries throughout the 1990s is an outcome of this new way of thinking. Critics have argued that this new policy has merely been a way to legitimize non-entr'e policies of the North and has not offered effective protection compared to that offered by the asylum policies. Nonetheless, in the light of the limitations of the asylum policy it is very relevant to evaluate the role safe havens could still play, as a parallel policy in dealing with the sheer number of people in situations of peril and the need of the IDPs.
Criteria For Truly Safe 'Safe Havens': The Geneva Convention Definition
The idea of implementing safe zones to protect civilians in times of war was not new to the 1990s; it was already a tenet of international humanitarian law. It is useful to refer back to its original meaning, to understand how recent safe-haven operations have tended to differ from this previous definition.
The general understanding of what a safe haven ought to be, under humanitarian law, boils down to a "location within the disputed country of territory, neutral, free of belligerent activity, to which humanitarian access is ensured." The 1949 Geneva Convention concerning the Protection of Civilian Persons in Times of War was the first attempt to define the guiding principles of a safe-haven policy. Article 15 states that any party to a conflict may propose to establish "neutralized zones intended to shelter from the effects of war" the "wounded and sick combatants or non-combatants" and "civilian persons who take no part in hostilities." The humanitarian nature of these zones is highlighted in Articles 23 and 59, which states that parties shall permit "the free passage of all consignments of essential foodstuffs," medical supplies, and clothing to these zones.
The understanding of how these 'safe zones' should be implemented is clearly based on the concept of consent. Article 14 states that the parties concerned may "conclude agreements on mutual recognition of the zones and localities they have created." The other important element that is made clear in the Convention is the nonmilitary aspect of the zones. They are limited to civilians, and it is stressed that the inhabitants must "perform no work of a military character."
The most successful example of this type of safe zone was one established in Shanghai in the 1930s, during the Second Sino-Japanese War. The safe zone, which resulted from negotiations between the parties, sheltered a quarter of a million Chinese. However, such nonmilitary? Geneva places, based on consent, have not often been implemented.
It is important to keep in mind the Geneva criteria for a safe zone (i.e. consent-based and purely civilian), for our analysis of the safe-area experiences of the 1990s, as it will become apparent that they have not followed these basic principles.
Applications of the Safe Area Concept During the 1990s
It is hard to generalize about recent safe-haven operations, since they have been conducted in an ad hoc manner without a uniform legal mandate. Some safe areas have been protected with extensive military power (e.g. Iraq) whereas others have been based on mere consensus (e.g. Sri Lanka). Nonetheless, all safe-haven attempts have been intended to protect potential refugees within their own borders. The idea was to "keep refugees close to home and to keep the pressure on human rights violators to clean up their act and to restore the rights and security of their own citizens."
Safe-haven policies have led the UNHCR to address the needs of people other than refugees. Indeed, in recent publications the organization describes its policy as "proactive, homeland-oriented and holistic." This new proactive policy, mainly channeled through safe havens, has been the primary vehicle for providing protection to IDPs in situations of conflict. This development is closely linked to the rhetoric of the right to remain. This mechanism, developed by the UNHCR, was designed to solve the problem of countries being increasingly reluctant to receive refugees on a large scale. It offered to eliminate the causes of refugee movement by protecting individuals in their homeland. Although the rhetoric of this policy sounds very appealing, the key is to evaluate first the quality of the protection being offered, and second to assess whether the right to remain seems to have undermined the traditional right to flee. Three prominent examples of safe-haven policies are examined in three linked essays: Operation Provide Comfort in Iraq, the six Bosnian 'safe zones', and Operation Turquoise in Rwanda.
Problems With Safe-Haven Policy
In the light of these three case studies, one can infer conclusions about the effects of a safe-haven policy on refugee policy, on the safe-haven policy itself and the quality of protection it has offered, and finally on the reasons why the policy has seemed to fail in recent attempts.
Since safe-haven policies have been implemented within this new concept of the right to remain, we ought to assess the effects of such a policy on the rights of refugees as well as on the duties of states towards them. Safe havens have generally been implemented as a second-best alternative when other nations have not been willing to provide asylum and thus protection to refugees fleeing violence. This is clearly the case in our three examples:
- In the case of Iraq, the 'safe zone' was implemented because Turkey closed its border to 400,000 persecuted Kurds, and the international community decided to alleviate the pressure on Turkey and instead put it on Saddam Hussein.
- In Rwanda, to prevent the would-be refugees from fleeing to Zaire, Operation Turquoise implemented Zones Humanitaires Sures in the bordering region of southwestern Rwanda.
- The same incentive was apparent in the case of the Bosnian safe areas, since they were set up because European states clearly prevented Bosnians from fleeing by imposing strict visa restrictions.
Another element of the safe-area regime that greatly affected refugee status was the fact that the implementation of these safe areas actually hindered the individual's original right to flee. To a certain extent, one could argue that this is actually the aim of safe areas, to stop would-be refugees from seeking asylum outside, but as we will see later, in the case where the protective regime is not of a comparable quality to asylum it is important to be aware of the extent to which implementing safe havens impedes refugees' rights to seek asylum in a third country. In our three examples, once again it is made clear that the establishment of safe-haven operations hindered the right to flee. The most striking example is that of Rwanda, in which France was rejecting asylum demands from Rwandans as the killings at Kibeho were taking place. In all three cases, third countries along with UNHCR rejected demands for asylum, based on the claim that the asylum-seekers were being protected in their own countries.
A final point that can have a crucial impact on the bearings of international refugee law is that safe havens have been based on highly politicized criteria. Here, we consider the argument that Posen has made about the risk of humanitarian intervention. He argued that deviation from the sovereignty norm has followed a double standard: "When the bad guys are weak, such as in Iraq in Kurdistan, intervention pops to the top of the agenda, when they are strong, such as Russia in Chechnya, little is said." In the case of safe-haven intervention, the decision to establish a haven is based on political imperatives rather than on humanitarian needs. Nobody has so far proposed to establish a safe-haven operation in Russia to protect the Chechens from governmental persecution. This point also seems clear in our three cases: for example, the safe zone was implemented in Iraq because the plight of the Kurds took on strategic importance after the Gulf War. Similarly, nobody initially reacted to the genocide in Rwanda because it was not of strategic importance to any of the big powers, and when the international community finally intervened to create safe zones it was because France seemed to have an interest in the region.
What do the case studies show, in terms of the quality of protection that was offered to IDPs and would-be refugees? Safe areas did offer a certain degree of protection. The safe zone in Iraq was probably the most successful: it saved thousands of Kurds, largely due to the operation's credible military threat and strong political will. Even though the two other cases are less clear-cut, the safe areas brought humanitarian relief and temporary safety to thousands of people. One can always ask the rhetorical question of whether it would have been worse without the implementation of safe havens, but a clear success of the safe-haven policy has been the protection of many Internally Displaced Persons who would not have been protected under the asylum regime.
Nonetheless, if we are to consider the safe-haven policy as a credible alternative policy to refuge in a third country, we have to evaluate the quality of the protection. In these cases, one can argue that the quality was poor because there was no guarantee of safety. This is clearly exemplified by the Srebrenica massacre in Bosnia, the Turkish incursion in Iraq, and the Kibeho incident in Rwanda. One can link this problem to the inherent nature of the conflicts that were causing people to flee. As Frelick rightly points out, protection during an ethnic conflict is another matter than providing assistance after a natural disaster, because there is no guarantee of safety. In the case of a genocide or politicide it is very hard to protect a specific group of people within the country in which the conflict is taking place.
Another failure of the safe-haven regime is that it is neither sustainable over time nor does it actually solve the problem. Safe havens can be a tool to reduce hardship, but they definitely do not resolve the original predicament that produces violence. They are an interim measure and should be treated as such. The sustainability of a safe-haven regime will depend on how quickly an agreement can be reached to resolve the conflict.
It is important to understand which elements of the safe-haven regime seem to have led to these different problems. The difficulties faced by the current safe area policies can be summarized by three common failures:
- failure to demilitarize the area;
- failure to gain the consent of all the parties involved; and
- failure to gain credible military backing.
The first and last are policy failures, and the second seems to be due to the nature of many current conflicts.
The demilitarization requirement that was made abundantly clear in the 1949 Geneva Convention has not been respected in recent safe-haven operations. It is clear that a safe zone that represented a territorial or a military threat would be a contradiction in terms. The risk faced by a safe area that is not clearly civilian is obvious; it will become an element in the conflict and may be attacked if it is seen as threatening. This failure is once again apparent in two of our cases, in Bosnia and Rwanda. In Bosnia, the areas were used by the Bosnian military as training and resting camps. This greatly undermined the Serbs' respect for the safe-area regime, since it was clearly seen as a military threat. In the case of Rwanda, the ineffective disarmament of Hutus made the camp a security threat and led the Rwandan government to feel endangered by it. In the case of Iraq, I would argue that its nonmilitary nature is one of the reasons why this safe-haven experience was more successful than the other two. In Iraq the operation was clearly identified as protecting Kurdish civilians, which explains why it was generally respected. However, one could argue that the presence of PKK members within the civilian population represented a threat for the Turkish state, which could explain the Turkish incursion in the safe zone.
The other key criterion for a truly safe 'safe haven' as stated in the Geneva Convention is that it needs to have the consent of the parties. It is the acceptance of a safe haven by the parties that "(underlies) subsequent respect for its condition." This criterion can largely be linked to the demilitarization criterion, because it is clear that a party will not consent to a safe haven if it represents a military threat. In Bosnia, the Serbs considered the safe havens as obstacles to their ethnic cleansing policy and their territorial conquest. The same can be said in Rwanda and Iraq, where consent for the operation was also absent. This underlines the dilemma of trying to establish consent for safe havens in situations of ethnic conflict or fratricide, where the shared interest in protecting civilians is often absent since civilians are the targets.
In the case where consent for the operation is not clear, the key to the effectiveness of the safe-haven operation seems to be its military credibility and the commitment of the international community to protect the area. For example, in Iraq the safety zone was imposed externally and was not based on consent. Nonetheless, the operation was generally successful in protecting the Kurds because it was backed by a credible military threat, and clear political will to use that threat. By contrast, in Bosnia the number of ground troops was pitiable, and the threat of air bombings was not credible since two safe areas fell to the Serbs without significant NATO reaction. Only after Gorzade was threatened a second time was there a real attempt at creating a genuine safe haven. In Rwanda, even though Operation Turquoise was authorized to use all necessary means to protect the ZHS, it failed to credibly protect the IDPs, by not intervening when the government attempted to close the camps.
Solution: Toward a Restructured Safe-Haven Policy
In light of these apparent problems, many analysts have argued that we should scrap the current safe-haven policy. Refugee-law specialists such as Hathaway and Neve have argued that the right to remain is based on a fallacy. They argue that safe havens have failed to offer surrogate protection and hence must no longer be used because they have been invoked to legitimize non-entr'e policies. Their proposal to reform the refugee system centers on asylum policy, and offers a twofold approach:
- the temporary nature of asylum has to be highlighted, to send a clear signal to Northern countries that asylum is not a back door to immigration;
- interest-convergence groups should be set up developing a common but differentiated responsibility toward refugees.
Even though the first aspect of this proposal seems conducive to ameliorating access to third countries for refugees, Anker, Fitzpatrick, and Shacknove criticize the second aspect as moving asylum-seekers from the realm of law to that of political bargaining. However the major flaw I perceive in Hathaway and Neve's argument is that even though they acknowledge that IDPs suffer more severely than refugee communities, they do not offer a solution to protect such an at-risk population; they center their reforms on the restructuring of asylum policies, a regime to which IDPs do not have access for protection.
The mere reliance on the reform of the asylum system is not sufficient to deal with the sheer numbers of people in need of protection, especially as we have seen third countries increasingly retreating from their legal obligations to provide asylum. Asylum does not protect all the individuals who fail to fall under the rigid category of refugees. Hence, I agree with Arulanantham when he argues that safe-haven policy should not be completely abandoned as a tool for protection. There seems to be a gap in the refugee-protection system that could be successfully filled by a restructured safe-haven policy, one that is closer to its original 1949 definition.
As was made clear in our analysis of the case studies, recent safe-haven operations have failed because they were neither demilitarized nor consent-based. Future safe-haven operations must be demilitarized from the outset, since they will be respected by the parties involved only if they are clearly neutral and civilian. This is obviously not easy to implement in cases of civil wars and ethnic conflict, when fighters do not wear uniforms or when they tend to hide among civilians. Nevertheless, there are some concrete steps that can be taken to ensure that the safe area stays a purely civilian area. The first step is to disarm anybody entering the delimited zone. This simple step has too often been ignored, as in Rwanda where the French made no attempt to demilitarize the Hutus, or in Bosnia where the resolutions concerning safe havens never contained a demilitarization clause. The second step is that safe areas could be restricted to women, children, and elderly people in cases where the situation is too muddled to decide whether individuals are fighters or civilians.
If a safe-haven operation takes place, it has to be based on the consent of the parties involved. The operation has to be clearly defined and understood by all parties. As Landgren points out, it would be preferable if safe zones were negotiated at ground level, by those familiar with local conditions. The case of Sri Lanka and its experience with Open Relief Center is very interesting, as an example of the importance of consent. Open Relief Centers (ORCs) are different from safe areas because safety is not supposed to be guaranteed; instead they are "temporary places where displaced persons on the move can freely enter or leave and obtain essential relief assistance in a relatively safe environment." In Sri Lanka, the Open Relief Centers have been described as relatively safe, since their small size and transitory nature make them politically insignificant. These Open Relief centers in Sri Lanka received governmental support and were negotiated within the country, rather than prompted by external states. This consent-based approach seems to have worked very well, since the ORCs have protected many displaced persons without having military backing.
In cases where consent is questionable, or when safe havens are established in violation of a state's sovereignty, then the safe haven has to be backed with credible military threat and strong political will to use it. The forces charged with the protection of the safe havens should be appropriately equipped with the adequate mandate. Future safe-haven operations must resemble the military credibility present in Iraq, rather than the inadequate resources and political will present in Bosnia. This is clearly problematic, because dealing with political will and resources in the field of humanitarian interventions is very difficult. The only solution in dealing with these discrepancies would be to have fixed guidelines on when to establish safe havens and what resources would be needed, based on a threshold. Arulanantham proposed to set a limit of 100,000 displaced people as the threshold at which a safe area should be automatically put in place. This could circumvent the politicized nature of recent safe-haven operations, since a predetermined threshold would make the intervention automatic instead of based on the current double standard. One could follow the same logic in determining how many troops should be made available, or the size of the haven itself.
A crucial element in the future use of the safe-haven policy is that it should not be used or considered to be an alternative for the current asylum policy. The aim here is to restructure safe havens so as to offer safety to those who wish to remain in their country, but this should under no circumstance impinge on an individual's right to flee. This attempt to restructure safe havens should be seen as a parallel attempt to an increased protection of the non-refoulement norm, which has been at the heart of the problem in recent safe-haven operations. Governments have refused to grant asylum to refugees on the basis that they were already being protected, therefore greatly impeding their right to flee, even though safe-haven operations have not offered adequate protection to allow for such logic. This vicious circle does seem hard to break, but a recent proposal by Arulanantham has shed a positive light on this dilemma.
He proposes that a new body should be established whose sole purpose would be to report on whether refugees' rights are being adequately protected during safe-haven operations. The Refugee Convention already contains a provision for periodic reporting and collective scrutiny, which was never implemented by the UNHCR. He argues that with increased information, governments would be forced to either accept refugees by the thousands, or to provide more suitable protection in safe havens. Governments would thus be faced with three choices:
- declining to honor their refugee obligations, which would be highly unlikely in the light of the historical tradition of refugee law;
- accepting refugees in large numbers, again highly unlikely in the current political climate; or
- increasing protection for safe havens.
As governments across the globe fail to honor their obligation to protect individuals at risk through asylum policies, and as conflicts unleash massive influxes of refugees, I argue that there is still a role for a restructured safe-haven policy. Safe havens can be a very useful tool in protecting would-be refugees and IDPs, and are a helpful mechanism for the delivery of humanitarian aid. Still, safe-haven policy should be implemented as a parallel policy to asylum and should under no circumstances hinder refugees' right to flee. As we have seen in the past, humanitarian disasters can occur when safe havens impede the rights of refugees to seek asylum without offering an adequate level of safety.
I propose to return to the original understanding of what a true safe area entails. Thus, future safe-haven operations should be demilitarized from the outset, they should be based on consent, and if they lack any degree of consent they should definitely be backed by credible military threat. The key to developing safe havens without obstructing refugees' right to flee seems to lie in the development of a new information mechanism that would credibly assess the quality of protection in safe-haven operations, either pressuring governments to offer higher-quality protection during safe-haven operations or forcing them to accept thousands of refugees.
 BBC World, March 31, 2003, www.bbc.co.uk
 UN Refugee Convention (1951), Article 33. Please see this page to search for the Convention http://www.unhcr.ch/cgi-bin/texis/vtx/publ
 J. Hathaway and A. Neve "Making International Refugee Law Relevant Again: A Proposal for Collectivized and Solution-Oriented Protection" in Harvard Human Rights Journal.
 D. Dombey "UK asylum proposal draws mixed response" Financial Times, March 29/30, 2003.
 F. Deng & R. Cohen, Masses in Flight: The Global Crisis of Internal Displacement, p. 18.
 Cited by R. Cohen in "Nowhere to run, no place to hide," Bulletin of the Atomic Scientists, (November/December 2002), p. 38.
 B. Frelick "Aliens in their own Land: Protection and Durable Solutions for Internally Displaced Persons" in World Refugee Survey 1998, p. 30.
 J. Hathaway and A. Neve (1997), p. ?
 K. Landgren (1995), p. 438.
 Geneva Convention Relative to the Protection of Civilian Persons in Time of War, 12 August 1949, Article 15.
 Ibid, Article 23 and Article 59.
 Ibid, Article 14.
 Ibid, Article 15.
 K. Landgren (1995), p. 439.
 A. Arulanantham "Restructured Safe Havens: A proposal for Reform of the Refugee Protection System", Human Rights Quarterly, Vol. 22 (2000), p. 21.
 B. Frelick, Safe Havens, Broken Promises, U.S. Committee for Refugees
 UNHCR, The State of the World's Refugees: Fifty years of Humanitarian Action (2000), p. 4. avaliable at http://www.unhcr.ch/cgi-bin/texis/vtx/template/+kwLFqQp1xceUh5cTPeUzknwBoqeRzlmiG+XXXeRDlmqnDteIybnM
 F. Deng & R. Cohen (1998), p. 281.
 B. Posen, p. 94.
 B. Frelick, U.S. Committee for Refugees.
 L. Franco, "Safety Zones for Internally Displaced Persons" in N. Al-Nauimi & R. Meese (eds) International Legal Issues Arising Under the United Nations Decade of International Law, p. 885.
 K. Landgren
 Larry Minear and Thomas G. Weiss, Mercy Under Fire: War and the Global Humanitarian Community, p. 168. Westview Press. 1995.
 K. Landgren, p. 440.
 F. Deng and R. Cohen, p. 281.
 B. Posen, p. 103.
 J. Hathaway and A. Neve, 1997.
 D. Anker, J. Fitzpatrick and A. Shacknove "Crisis and Cure: A Reply to Hathaway/Neve and Schuck" in Harvard Human Rights Journal, Spring 1998.
 A. Arulanantham "Restructured Safe Havens: A Proposal for reform of the Refugee Protection System," Human Rights Quarterly (2000).
 K. Landgren, p. 455.
 H. Senevirante and M, Stavropoulou "Sri Lanka's Vicious Circle of Displacement" in R. Cohen and F. Deng (eds) The Forsaken People: Case Studies of Internally Displaced, p. 383
 K. Landgren, p. 452.
 A. Arulanantham, p. 55.
 U.N. Refugee Convention (1951), Article 35.
 A. Arulanantham, p. 11.
Use the following to cite this article:
Haspeslagh, Sophie . "Safe Havens." Beyond Intractability. Eds. Guy Burgess and Heidi Burgess. Conflict Information Consortium, University of Colorado, Boulder. Posted: November 2003 <http://www.beyondintractability.org/essay/safe-havens>.