Blog post by Dr Özlem Gürakar Skribeland, Postdoctoral Researcher at the University of Oslo*


Recent years have seen vibrant debates on the future of international refugee law and the continued relevance of the Refugee Convention (i.e., 1951 Convention and/or 1967 Protocol). As part of this, the enduring focus on the laws and practices of a select few states/regions of the world and the perception of non-signatories as exceptions or deviants have increasingly come under criticism. Some 44 UN-member states have not ratified the 1951 Convention or the 1967 Protocol, among them key refugee hosts like Lebanon, Jordan, India, Pakistan, and Bangladesh. The European Research Council-funded BEYOND Project aims to bring to the fore some of these key non-signatory states, exploring the interactions and impacts between them and the international refugee regime.

Studying Turkey as a non-signatory

Turkey – the country hosting the biggest refugee population in the world since 2014 – does not technically fall in the non-signatory category. It has ratified both the 1951 Convention and the 1967 Protocol but, in both cases, with the optional geographical limitation. This means that as a matter of treaty law, Turkey is required to extend the protection of these instruments only to those persons who become refugees “as a result of events occurring in Europe.” The close to 4 million people seeking international protection in Turkey do not fulfil this criterion. In fact, a parliamentary report from 2018 (see p. 11) states that there are only about 70 persons in Turkey with refugee status as per the Refugee Convention. Arguably then, Turkey can be framed and analyzed as a non-signatory state.

As part of my research on Turkey’s protection regime and its place in the global refugee regime, I have, among other things, been tracing the Turkish courts’ engagement with key instruments of international refugee law. As would be well-known to the readers of this blog, Article 1 of the 1951 Convention defines the term “refugee”, which then appears in each of Articles 2-34. Given that the geographical limitation is part of the refugee definition, one would not expect to see the Turkish courts engage with those articles in the many judgments involving refugees from Syria, Iran, Iraq, Afghanistan, Pakistan, and other non-European countries – except maybe Article 33, considering the customary nature of the principle of non-refoulement. This, however, is not the case: Turkish court judgments do refer to multiple provisions of the 1951 Convention in cases involving non-European refugees under a range of different circumstances. I am unpacking this engagement, including its development over time, in detail in an article that I am currently working on. In this blog post, I would like to briefly draw attention to some Article 2 and Article 31 jurisprudence from Turkish courts.

Case law on Article 2 of the 1951 Convention

Article 2 of the 1951 Convention states that every refugee has duties to their host country, which requires in particular that they conform to the host country’s laws and regulations. Lambert (para. 37) observes that the jurisprudence relating to Article 2 of the 1951 Convention is hard to find but that it has been used by courts to justify acts going beyond the scope of that article and the overall humanitarian purpose of the 1951 Convention. In Turkey, too, this provision seems to be used by courts in a rather self-serving manner in recent years, in cases involving non-European refugees.

In 2013, Turkey passed the Law on Foreigners and International Protection (LFIP), the country’s first law (i.e., primary legislation) dealing with all aspects of migration and asylum. Under the LFIP, international protection applicants are subject to certain reporting and residency requirements, and not abiding by these rules leads to their applications for international protection to be deemed withdrawn (see LFIP Art. 77). When these deemed-withdrawal cases come before the Turkish Council of State (i.e., the highest administrative court), that court often agrees with the administration’s strict application of the relevant Turkish law provision, and interestingly, references Article 2 as additional justification (see, for example, (i) Council of State 10th Chamber E. 2016/3216 K. 2017/911 T. 21.2.2017, and (ii) Council of State 10th Chamber E. 2016/128 K. 2019/54 T. 7.1.2019, decided in 2017 and 2019 respectively).

Case law on Article 31 of the 1951 Convention

There seems to be a relatively bigger number of cases on Article 31 of the 1951 Convention, which deals with the situation of “refugees unlawfully in the country of refuge” (see Costello et al). However, the case law on this provision is still limited and the relevant Turkish jurisprudence is not internationally known. Interestingly, the Turkish Council of State does reference different parts/elements of Article 31 in cases involving non-European refugees under a number of different circumstances both in the pre- and post-LFIP period. However, a line of pre-LFIP judgments is particularly interesting and points to what I view as “real engagement” (whereby the provision’s application to the facts decides the outcome of the case), especially with the last sentence of Article 31 which reads: “The Contracting States shall allow such refugees a reasonable period and all the necessary facilities to obtain admission into another country.”

At the time of these judgments, asylum matters were governed by a rather short piece of secondary legislation (i.e., Regulation on the procedures and principles related to possible population movements and foreigners arriving in Turkey, either as individuals or in groups, wishing to seek asylum either from Turkey or requesting a residence permit in order to seek asylum from another country, commonly referred to as the 1994 Regulation). This regulation required asylum seekers to register with the Turkish authorities within a certain timeframe/without delay (see 1994 Regulation Art. 4); however, some would approach only the UNHCR and not register with the Turkish authorities within the required timeframe or at all. The aforementioned line of cases involves administrative decisions to deport asylum seekers for failure to register with the authorities and the resulting irregular presence. At that point, the persons concerned have often already been recognized as refugees by the UNHCR with resettlement efforts under way. When these refugees then challenge the deportation decisions at court, the Council of State rules in their favor and the last sentence of Article 31 forms the primary basis of the final judgment (see, for example, (i) Council of State 10th Chamber E. 1999/154 K. 2000/2756 T. 25.5.2000, and (ii) Council of State 10th Chamber E. 2004/8769 K. 2007/2735 T. 18.5.2007, decided in 2000 and 2007 respectively).

Looking forward

Turkish courts’ engagement with the 1951 Convention in these and other cases involving “non-European” refugees is overall very interesting. One reason for this is that there is much to be explored here, including certain shifts in the courts’ approach to similar fact patterns over time. And a bigger and related question that we as members of the BEYOND team are concerned with is how treaties influence states that have not ratified them.

There is much to be learned from a closer study of non-signatories to the Refugee Convention.

* I would like to thank Professor Cathryn Costello for drawing my attention to the limited case law on Article 31 and for advising me to write a blogpost to share some of my initial Turkish case law findings on this.


The views expressed in this article belong to the author/s and do not necessarily reflect those of the Refugee Law Initiative. We welcome comments and contributions to this blog – please comment below and see here for contribution guidelines.