Problems
In the light of the Iraq, Bosnia and Rwanda case studies one can infer conclusions firstly about the effects of a safe haven policy on a refugee policy level, secondly on the policy of safe havens 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 notion of the right to remain we must 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 protection to refugees fleeing violence. This is clearly the case in our three examples of Iraq, Bosnia and Rwanda. In the case of Iraq, Turkey closed its border to 400,000 persecuted Kurds when the ‘safe zone’ was implemented. The international community decided to alleviate the pressure on Turkey and instead put it on Saddam Hussein. In Rwanda, would-be refugees were prevented from fleeing to Zaire when Operation Turquoise implemented Zones Humanitaires Sure in the border 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 were clearly preventing Bosnians from fleeing by imposing strict visa restrictions. Another element of the safe area regime that greatly affected refugee status is 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, in cases 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 has impeded refugees’ right to seek asylum in a third country. It is clear from our three examples that the establishment of safe havens operation has hindered their right to flee. The most striking example is that of Rwanda, asylum demands from Rwandans were being rejected by France as the killings at Kibeho were taking place. In all three cases third countries, along with UNHCR, rejected demands for asylum based on the fact 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 the fact that safe havens have been established on highly politicized criteria. The decision to establish a haven is based on political imperatives rather than on humanitarian needs. A safe zone was implemented in Iraq because the plight of the Kurds took on strategic importance after the Gulf war. In the same respect nobody reacted to the Genocide in Rwanda because it was not of strategic importance to any of the ‘big powers’; when the international community finally intervened to create safe zones it was because France had an interest in the region. Quality of Protection 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 case of Iraq was probably the most successful attempt, where the safe zone saved thousands of Kurds thanks largely to the operation’s credible military threat and strong political will. Even though the two other cases are less clear cut the safe areas did protect many people: safe areas have brought humanitarian relief and temporary safety to thousands of IDPs 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 this case 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. This lack of safety is largely due to the nature of the conflicts that were causing people to flee. As B. Frelick rightly points out, protection during an ethnic conflict is another matter than providing assistance for 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 where 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 do not solve the original predicament that produces violence. It is an interim measure and should be treated as such. The difficulties faced by the current ‘safe area’ policies are due to three main failures: the failure to demilitarize the areas; the failure to have consent of all the parties involved; and the failure to have a credible military backing. Demilitarization 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’ which represents a territorial or a military threat is a contradiction in terms. 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 it could be argued that the non-military nature of the safe haven is one of the reasons it was more successful. In Iraq the operation was clearly identified as protecting Kurdish civilians, which explains why it was generally respected. At the same time one could argue that the presence of PKK members within the civilian population represented a threat for the Turkish state and that could explain the Turkish incursion in the ‘safe zone’. Consent The other key criteria for a truly safe ‘safe haven’ stated in the Geneva Convention is that they need to be based on the consent of the parties. This criteria can be implicitly linked to the demilitarization criteria 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 for their ethnic cleansing policy and their territorial conquest. The same can be said in the cases of Rwanda and Iraq, where consent for the operation was also absent. Furthermore, 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 due to the simple fact that civilians are the target. Credible Political and Military Backing In cases where consent for the operation is not clear, the key to effectiveness of the safe haven operation seems to be its military credibility and the commitment of the international community to protecting them. For example, in Iraq the safety zone was imposed externally rather than being based on consent. Despite this, the operation was generally successful in protecting the Kurds because it was backed by a credible military threat and clear political will to use the threat. In contrast, in Bosnia the number of ground troops was pitiful, and the threat of air bombings was not credible since two safe areas fell to the Serbs without significant NATO reaction. It was only after Gorazde was threatened a second time that there was a real attempt at creating a genuine ‘safe haven’. Again in the case of Rwanda, even though Operation Turquoise was authorized to use ‘all necessary means’ to protect the ZHS, it failed to do so credibly by not intervening when the government attempted to close the camps. | 
