De-centring the Securitisation of Asylum and Migration in the European Union: Securitisation, Vulnerability and the Role of Turkey

ABSTRACT This article contributes to the debates on de-centring the analysis of migration governance in Europe by focusing on the potential role of external actors in the securitisation of asylum and migration in the European Union (EU). Although there has been a growing amount of literature on the securitisation of asylum and migration in the EU, the role possibly played by external actors in this securitisation process has not been considered to date. This article addresses this gap using the case of Turkey. Theoretically, it contributes to the development of the securitisation framework by de-centring the study of securitisation processes. It argues that, from the vantage point of an external actor, a securitisation process highlights the existence of a vulnerability to a specific phenomenon that is perceived to be threatening. An external actor can then decide to exploit this vulnerability for its own gain, notably by making threats that play on the fears of the other political actor. Empirically, the article demonstrates how the Turkish government has been able to exploit the vulnerability of European countries to migration flows, which had been highlighted by the social construction of asylum and migration as security issues. By repeatedly threatening to send more asylum-seekers and migrants Europe’s way, the Turkish authorities have managed to secure some significant financial and political benefits for themselves in the last few years.


Introduction
Migration has become one of the most salient political issues in Europe in recent years. This has notably resulted from the linkages that have been made between migration and security, both discursively and through the deployment of specific practices. Regarding the former, a wide range of actors have made statements linking migration to security in the last few years. For example, Hungarian Prime Minister Viktor Orbán described migration as the 'Trojan wooden horse of terrorism' at the 2017 Malta Congress of the European People's Party (Brunsden 2017), whilst French Presidential hopeful and migration is Turkey, as epitomised by the signing of the controversial socalled ' EU-Turkey Statement' of 18 March 2016(European Council 2016.
As a consequence, this article addresses the topic of the role of external actors in the securitisation of asylum and migration in the EU, which has been neglected to date. It does so using the case of Turkey, as it appears to have played a significant role in the governance of asylum and migration in the EU. Methodologically, a range of documents are analysed, including EU official documents and media reports on EU-Turkey relations over asylum and migration. The remainder of this article is structured into five sections. First, it considers the burgeoning literature on the securitisation of asylum and migration in the EU and highlights that it has neglected considering the role of external actors in this process. Second, it introduces the securitisation framework and examines whether the role of external actors in securitisation processes in general has been conceptualised. As this question is answered negatively, the article puts forward the argument that a crucial aspect of a securitisation process from the vantage point of an external actor is that it reveals a vulnerability that can be exploited by this external actor for its own advantage. Moving on to the specific case of Turkey, the third section sets the context by discussing EU-Turkey relations in general and with regard to asylum, migration and border matters in particular. The following section discusses how Turkey has attempted to exploit the vulnerability of European countries to flows of asylum-seekers and migrants -given that asylum-seekers and migrants have been securitised in the EU -in the context of the Syrian refugee crisis. It examines the various migration-related threats made by the Turkish government and highlights what it has gained as a result. The final section offers some conclusions.

The Securitisation of Asylum and Migration in the EU
The last two decades have seen a significant growth in the literature on the securitisation of asylum and migration in the EU. This has notably reflected the increasing role that the EU has acquired in that policy area (Geddes, Hadj Abdou, and Brumat 2020;Léonard and Kaunert 2019). Although some have sought to put forward a more nuanced argument (see Boswell 2007;Neal 2009), a commonly held view in the existing literature has been that asylum and migration have been successfully securitised in the EU, that is, socially constructed as security issues. Within that debate, scholars have pursued two main lines of enquiry, which have not been mutually exclusive. First, some contributions have focused on the actors involved in and the processes underpinning the securitisation of asylum and migration (e.g., Bourbeau 2011;Ceyhan and Tsoukala 2002;Huysmans 2000Huysmans , 2006Lazaridis and Wadia 2015). An important contribution to the analysis of the processes through which migration has been securitised in Europe has been made by Bigo (1998Bigo ( , 2002, who has highlighted the importance of practices, in contrast to the emphasis placed on 'speech acts' (i.e. discourse) in the original conceptualisation of securitisation by the Copenhagen School. Huysmans (2006), Balzacq (2008) and Léonard (2010) have further explored the role of practices in the securitisation of asylum-seekers and migrants in the EU. Second, other contributions have critically examined the consequences of the securitisation of asylum and migration in Europe, including negative, unintended consequences (e.g., Bello 2017;Chebel d'Appollonia 2012;Karyotis 2011). Some of those have called for strategies to 'desecuritize' migration (Huysmans 1995) or to develop a more 'inclusionary' approach to asylum in Europe (Squire 2009). Thus, the idea that asylum and migration have been securitised in Europe has been widely shared in the academic literature.
Whilst these scholarly works have greatly contributed to our understanding of the securitisation of asylum and migration in the EU, they have omitted to consider an important factor, namely the potential role of external actors, that is, non-EU actors, in the securitisation of asylum and migration in the EU. This concern is prompted by the fact that, in recent years, the EU has steadily developed relations on asylum and migration matters with non-EU countries (i.e. 'third countries' in EU jargon), which has been documented by a growing amount of scholarly work. Whilst early academic literature emphasised the existence of a largely asymmetric relationship between the EU and third countries in favour of the EU (e.g. Lavenex and Schimmelfennig 2009), several scholars have more recently argued that one should not underestimate the agency of third countries in their relationship with the EU, especially when it comes to the governance of specific policy issues, such as migration (e.g., El Qadim 2015; Wolff and Trauner 2014;Wunderlich 2012). Others have gone even further by arguing that some governments have managed to instrumentalise flows of asylum-seekers and migrants towards Europe in order to successfully increase their leverage over European governments (Okyay and Zaragoza-Cristiani 2016;Paoletti 2011;Tsourapas 2017Tsourapas , 2019Völkel 2020). Such ideas echo the concept of the 'weaponisation' of asylum and migration flows originally developed by Greenhill (2010Greenhill ( , 2016, which entails 'the intentional creation, manipulation, and exploitation of migration and refugee crises' for foreign policy purposes (Greenhill 2010, 2). One of the most prominent examples of the instrumentalisation of migration flows was Libya under Muammar al-Gaddafi (Paoletti 2011;Tsourapas 2017). More recently, some scholars have turned their attention towards Turkey and its use of the issue of migration in its relations with European countries. For example, Okyay and Zaragoza-Cristiani (2016) have argued that, thanks to its position as a migration 'gatekeeper', Turkey has been able to gain leverage over the EU and to increase its bargaining power in the negotiations that led to the adoption of the EU-Turkey agreement on migration. For his part, Tsourapas (2019) has claimed that the Turkish government has engaged in a specific type of what he terms 'refugee rent-seeking behaviour' in its relations with the EU, namely 'blackmailing', which he defines as 'threatening to flood a target state(s) with refugee populations within its borders, unless compensated'. Thus, various scholars have documented the growth in the relations between the EU and third countries on asylum and migration matters. Recently, there has been an increasing acknowledgement that third countries are not just passive recipients of the EU's policies, but actually possess agency in migration governance. This therefore points to the potentially important role of third countries in the securitisation of asylum and migration in the EU, which has not been analysed to date. The remainder of this article examines this hitherto neglected topic by focusing on the case of Turkey. This country has been selected given the important role that it appears to have played in the governance of asylum and migration in the EU, as epitomised by the signing of the EU-Turkey Statement of March 2016 (European Council 2016).

Securitisation Processes and External Actors
Before one can analyse the role of Turkey in the securitisation of asylum and migration in the EU, it is necessary to first present the securitisation framework, before considering how -if at all -the role of outsiders has been conceptualised in the literature on securitisation. At the heart of the securitisation framework, which was originally developed by the so-called 'Copenhagen School', lies the idea that security threats do not objectively exist, 'out there' (Buzan, Waever, and De Wilde 1998). Rather, security threats come into existence as the result of processes of 'securitisation', which can be understood as 'processes of constructing a shared understanding of what is to be considered and collectively responded to as a threat' (Buzan, Waever, and De Wilde 1998, 26). Moreover, still according to the Copenhagen School, this social construction of security threats takes place through the utterance of 'speech acts' (Buzan, Waever, and De Wilde 1998: 26; see also Waever 1995, 54-55). In a 'securitising move', an actor presents a phenomenon as a threat to the survival of a so-called 'referent object'. After the audience of this securitising move has accepted it, securitisation has successfully taken place. Furthermore, the Copenhagen School argues that, by 'uttering security', the securitising actor 'moves a particular development into a specific area, and thereby claims a special right to use whatever means are necessary to block it' (Waever 1995, 55).
Over the years, the securitisation framework has generated a considerable amount of debates, leading to the development of a large body of literature (see Balzacq, Léonard, and Ruzicka 2016). In particular, some scholars have highlighted some under-developed aspects of the framework -such as the issue of the relationship between the securitising actor and the audience (e.g., Léonard and Kaunert 2010) -and have attempted to develop them further (Balzacq 2005;McDonald 2008;Stritzel 2007).
However, what is remarkable in the literature on securitisation is the overwhelming focus of most scholars on the internal dynamics of securitisation and the concomitant neglect of the potential role of external actors (or outsiders) in securitisation processes. In other words, there has not been any research explicitly considering the possible role of external actors in securitisation, which was not included in the original securitisation framework developed by the Copenhagen School either. However, it is conceivable that some scholars could have implicitly or indirectly addressed the role of outsiders whilst using other concepts. Given that outsiders evidently do not belong to the category of 'securitising actors', there are arguably two main concepts to consider in that respect, namely 'functional actors' and 'counter-securitising actors'. 'Functional actors' is a category that was identified by Waever and Buzan themselves. However, they left its definition extremely vague. They merely noted that 'functional actors' can be defined as 'actors who affect the dynamics of a sector' and 'who significantly [influence] decisions in the field of security' without being a securitising actor or the referent object (Buzan, Waever, and De Wilde 1998, 36). Moreover, for a long time, this aspect of the securitisation framework was not explicitly considered or further developed by other scholars. However, a recent article by Floyd (2020) has argued that functional actors have usually been referred to as audiences in the burgeoning literature on securitisation. Thus, apart from highlighting the importance of distinguishing outsiders from functional actors, the literature on functional actors does not shed any significant light on the potential role of outsiders in securitisation processes. As for the concept of 'countersecuritising actors', which can be coined on the basis of the limited literature on counter-securitisation that has been written (e.g., Stritzel and Chang 2015;Vuori 2008), it appears to refer to actors attempting to contest and resist against a securitising move. However, whilst the category of countersecuritising actors may comprise outsiders (Stritzel and Chang 2015, 553), it is a much broader category, which also includes those who are the referent objects of the securitising move. One therefore needs to distinguish outsiders from counter-securitising actors. Apart from that insight, an examination of the literature on counter-securitisation does not shed light on the specific issue of the role of outsiders in securitisation.
Thus, the scholarship on securitisation, including the research that focuses on functional actors and counter-securitising actors, has not conceptualised the role of external actors neither explicitly nor implicitly. In the absence of any available conceptualisation of the role of external actors in securitisation in general, it is suggested here to return to the securitisation framework in its fundamental aspects and, in contrast to the usual focus on internal dynamics, to consider securitisation processes from the vantage point of an external actor. This is fully in line with a de-centring research agenda. When doing so, it appears that one of the most salient aspects of a securitisation process is the fact that securitising an issue -i.e. socially constructing it as a threat to the survival of a referent object -simultaneously highlights a vulnerability in a referent object to a specific threat. For example, when political leaders are depicting migration flows as a fundamental threat to the survival of the customs and values of a given nation, they are simultaneously -albeit indirectly and presumably not deliberately -highlighting how a nation is vulnerable in that respect. They lay bare their fears and their concerns, which may be more or less shared across society, as they point themselves at a source of vulnerability and relative weakness. In turn, an external actor witnessing this securitisation process from the outside may decide to exploit this vulnerability for its own benefit.
Thus, for an external actor, one of the most important dimensions of a securitisation process is that it reveals a vulnerability of which it can take advantage. One of the most common ways in which an actor can exploit the vulnerability of another on the international stage is extortion -in a metaphorical way. Extortion refers to a situation 'in which one person takes advantage of another person against his or her will by means of threat of violence or threat of harm of any kind to the person' (Pehlivan 2014). In other words, extortion is about obtaining benefits through coercion. In this article, 'extortion' is used metaphorically, as this act has not been criminalised at the international level. Thus, it is argued here that, from the point of view of an external actor, the securitisation of an issue highlights a vulnerability, which this external actor can then choose to exploit for its own ends, notably through extortion. In such a scenario, an external actor threatens to take action targeting the vulnerability of another actor -which has been revealed through a securitisation process -in order to gain benefits. On the international stage, one can conceive of a wide range of possible benefits, which could be military, political or financial, amongst others.
Having presented the securitisation framework and conceptualised how an external actor can exploit a vulnerability highlighted by securitisation, it is now possible to examine the role that Turkey has played in the securitisation of asylum and migration in the EU in recent years.

Setting the Context: EU-Turkey Relations
In order to be able to assess the role of Turkey in the securitisation of asylum and migration in the EU, it is necessary to first locate it in its broader context. This section begins by examining Turkey's relationship with the EU in general, before focusing on their relations on asylum, migration and border matters in particular. With regard to the wider context of Turkey's relations with the EU, those can be seen as having two main, inter-related elements. The first is the association that was established between the then European Economic Community (EEC) and Turkey through the Ankara Association Agreement in 1963 (Ott 2017, 8). This agreement notably aimed to create a customs union in three stages and highlighted in its preamble that 'the support given by the European Economic Community to the efforts of the Turkish people to improve their standard of living will facilitate the accession of Turkey to the Community at a later date' (italics added). The second component in the EU-Turkey relations is precisely the status of candidate country that Turkey only received in 1999, that is, twelve years after it applied for EEC/EU membership. This delay can be explained to a large extent by the controversial character of the Turkish membership of the EEC, which was divisive for several reasons (Nas and Özer 2017). One of them was the sheer size of the Turkish population and the importance that Turkey would therefore acquire within EEC/EU institutions and policy-making after its accession, in particular because of its future voting rights in the Council and in the European Parliament. Another reason was the significant cultural and socio-economic differences between Turkey and EEC/EU Member States (Saatçioğlu 2020). Yet another reason was the unsolved dispute over the island of Cyprus, the northern part of which had been invaded by Turkish forces in 1974 (Suvarierol 2003). Since then, accession negotiations have progressed very slowly and membership prospects have considerably decreased. Nevertheless, for various reasons (see Aydin and Kirişci 2013;Tolay 2014, 2), the EU has managed to have some impact on Turkish policies, notably those on asylum, migration and borders, to which this section now turns.
Concerning asylum, migration and borders, the EU has exercised some degree of influence over Turkish policies as part of the accession process. However, this influence has been characterised by contradictions. On the one hand, some reforms have been made to align Turkish legislation with the EU acquis communautaire. On the other hand, foreign policy considerations have led Turkey to adopt or to maintain some measures that have diverged from the EU's own policies, notably in the field of visas (Yıldız 2016, 117-121). For example, Turkey introduced a mutual visa exemption regime with Syria in 2009. Another persistent bone of contention between the EU and Turkey has been the latter's refusal to lift the geographical limitation clause of the 1951 Geneva Convention and the 1967 Protocol (Aydin and Kirişci 2013, 383), which means that it does not grant full refugee status, but only conditional refugee status, to persons fearing persecution as a result of events occurring outside Europe (Güler 2020).
Nevertheless, in the last few years, Turkey has taken some steps that have brought it closer to the EU again in some respects. In particular, in December 2013, Turkey signed a readmission agreement with the EU, whilst a visa liberalisation process was also launched (Açıkgöz 2015). Nevertheless, the negotiation of that readmission agreement was long and cumbersome, as it had already started in 2002 (Ott 2017, 13). The Turkish authorities were rather reluctant to sign such an agreement, largely because of their fear that, as a result, Turkey could become a buffer area between the EU and various unstable countries in the region, which have sent significant numbers of asylum-seekers and migrants in the last few years. However, they were very determined to achieve visa-free travel for their citizens. After realising that this would not be achieved through rulings of the Court of Justice of the European Union interpreting the EU-Turkey association agreement (Ott 2017, 14), the Turkish government changed its negotiation strategy, thereby paving the way to the signing of the readmission agreement with the EU in December 2013. This agreement entails the readmission of Turkish nationals, as well as of third-country nationals and stateless persons as long as those have entered the EU through Turkey. In parallel, an EU-Turkey Visa liberalisation dialogue was launched, which mainly consists of a screening process of Turkey's legislation and administrative practices by the European Commission on the basis of the 'Roadmap towards the visa-free regime with Turkey' (European Commission 2013a, 2013b). The dialogue mainly aims to make progress towards abolishing the visa obligation currently imposed on the Turkish citizens travelling to the Schengen area for a short-term visit. The roadmap identifies five areas in relation to which Turkey is required to make legislative and administrative reforms, namely document security, migration and border management, public order and security, fundamental rights and the readmission of illegal migrants (European Commission 2013b). The launch of this visa liberalisation dialogue emphasised once again the importance for the Turkish government of achieving visa-free travel for its citizens on short-term visits. However, it is important to note that the roadmap did not include any specific timetable. On the contrary, it highlighted that '[the] pace of movement towards visa liberalisation [would] depend on Turkey's progress in adopting and implementing the measures and fulfiling the requirements set out in this Roadmap, including full and effective implementation of the readmission agreement' (European Commission 2013b). Moreover, although the process is called a 'dialogue', it has been significantly asymmetrical in favour of the EU, as it has been based on a document drawn up by the European Commission that identifies a long list of conditions that Turkey must fulfil before reaching its objective of being granted visa-free travel for its citizens.

The Impact of the Syrian Refugee Crisis on Turkey and the EU
At the same time as Turkey has engaged in these negotiations with the EU, it has had to deal with the arrival of hundreds of thousands of people displaced by the Syrian conflict since 2011 (Tolay 2014). In line with the de-centring agenda, it is important to also consider this situation from the viewpoint of the Turkish authorities. Before doing so, it is important to recall that, as previously mentioned, Turkey is one of few countries that has maintained the geographical limitation in the 1951 Geneva Convention. This can be explained by several long-term features, which are also highly relevant to understanding the attitude of the Turkish authorities in the face of the Syrian refugee crisis. First of all, there has been a concern in Turkey since the signing of the 1951 Geneva Convention that it is a country prone to receiving large flows of asylum-seekers because of its geographical location and the political situation in the neighbouring countries. It has been feared that the country could be overwhelmed with large numbers of asylum-seekers and irregular migrants, which could affect national security. In that respect, maintaining the geographical limitation in the 1951 Geneva Convention has been seen as a way of maintaining some degree of control over these migration flows, whilst managing relations with the governments of neighbouring countries more smoothly (Kirişci 1996, 308-309). Moreover, it is important to recall that nation-building in the Turkish Republic has emphasised the 'Turkishness' of the country in contrast with the multi-religious and multi-ethnic character of the Ottoman Empire. In other words, traditionally, Turkey has tended to largely view itself as not being a country of immigration. Flows of asylum-seekers and migrants have therefore been regarded with a certain degree of suspicion given the impact that they could have on Turkish culture (Kirişci 1996, 309). Furthermore, the Turkish authorities have often perceived the international community, including European states and the EU, as not doing enough in terms of burden-sharing in times of humanitarian crises, whilst being overly critical of the Turkish efforts to deal with asylum-seekers and migrants (Kirişci 1996, 309-310).
From spring 2011 onwards, Turkey began to receive large numbers of asylum-seekers from Syria (Tolay 2014). Those were originally treated as 'guests', who did not receive any residence permit or official document confirming their status. Thus, although they enjoyed de facto protection from the Turkish state, they were not provided with the minimum guarantees laid down by the 1994 Turkish legislation on asylum. Following criticisms from the international community, the Turkish authorities decided to grant Syrian asylum-seekers temporary protection in October 2011 (Tolay 2014, 4), which was eventually enshrined in a new Law on Foreigners and International Protection adopted in April 2013 (Zaragoza-Cristiani 2015, 15). This went hand in hand with significant efforts to support Syrian asylum-seekers, although the services provided to those outside camps -the majority of asylum-seekers -were more limited (Kirişci and Salooja 2014). By December 2013, more than 600,000 Syrians had found refuge in Turkey (Dinçer et al. 2013, 2). As emphasised by Kirişci and Salooja (2014), 'the Turkish government [. . .] [had] never opened its borders to refugees from the Middle East in this way'.
Over the next few months, the challenge for Turkey continued to grow, as the rate of arrival of Syrian asylum-seekers increased. From November 2014 until August 2015, the number of registered refugees in Turkish camps went from 1,060,279 to 1,938,999 (Zaragoza-Cristiani 2015, 9). This notably resulted from the restrictive policies put in place by other countries in both the Middle East and Europe. Also, it is important to acknowledge that the difficulties faced by the Turkish authorities were partially self-inflicted. Whilst other countries introduced restrictions, the Turkish government prided itself on its so-called 'open door' policy (Oktav and Çelikaksoy 2015). This mainly aimed to raise its profile amongst Syrian refugees and across the Middle East and to develop its role as a regional leader (Kirişci and Salooja 2014).
By 2015, Turkey's reception capacities were significantly overstretched as asylum-seekers continued to flee from the violence in Syria. The Turkish government called for more support from the international community (Tolay 2014, 5), of which it was simultaneously very critical at times. As underlined by some observers, this stance might have undermined its efforts at garnering more international support (Oktav and Çelikaksoy 2015). It was around that time that the effects of the Syrian refugee crisis began to be significantly felt in Europe, particularly in Greece and Italy. For example, in less than five days in August 2015, more than 40,000 people arrived in Greece by sea from Turkey (Zaragoza-Cristiani 2015, 2). In the following months, growing numbers of asylum-seekers and migrants made their way from Turkey to European countries by sea or by land. This resulted from a combination of factors, including advances of the so-called 'Islamic State' in Syria and a decrease in the ability and the willingness of the Turkish authorities to control the flows of asylum-seekers and migrants (Zaragoza-Cristiani 2015). In the space of a few months, more than one million people crossed the Mediterranean and entered the EU, which was widely perceived as a 'crisis'. The fact that these flows of asylum-seekers and migrants were securitised in the EU -notably through a discursively constructed association between asylum-seekers and terrorists (Léonard and Kaunert 2019) -highlighted for the Turkish government a European vulnerability to migration that it could exploit for its own benefit.

Exploiting the EU's Vulnerability: Turkey and Its Migration-related Threats
Against this background of a growing 'crisis' both in Turkey and in Europe, an EU-Turkey Joint Action Plan was adopted, which had two main parts. The first aimed to increase support for the Syrian displaced population and the Turkish communities hosting them, whereas the second focused on reinforcing cooperation in order to prevent irregular migration, including intensified information exchange and further support to strengthen Turkish capacities for addressing migrant smuggling (European Commission 2015). Part II also required Turkey to accelerate return procedures in line with the established bilateral readmission provisions, referring to the EU -Turkey Readmission Agreement of 2013 (European Commission 2015).
At an international summit between EU leaders and the Turkish government on 29 November 2015, it was decided to activate the Joint Action Plan (European Council and Council of the European Union 2015). The EU also agreed to significantly increase its financial support through the 'Refugee Facility for Turkey', indicating that '[the] EU [was] committed to provide an initial 3 billion euro of additional resources' (European Council and Council of the European Union 2015, 1). In addition to re-opening the accession negotiations, both parties agreed that the EU-Turkey readmission agreement would start becoming fully applicable from June 2016, whilst it was envisaged that the lifting of visa requirements for Turkish citizens on short-term stays in the Schengen area would be completed by October 2016 should all requirements listed in the Roadmap be met (European Council and Council of the European Union 2015). According to Reuters (2016), at a meeting with EU officials earlier in November, Turkish President Erdoğan would have demanded to receive six billion euros over two years and would have made threats towards Europe. He was quoted as having said 'We can open the doors to Greece and Bulgaria anytime and we can put the refugees on buses [. . .] So how will you deal with refugees if you don't get a deal? Kill the refugees?' (Reuters 2016 Council 2016, 1). In addition, the EU committed to both accelerating the disbursement of the 3 billion euros that had initially been allocated to the Facility for Refugees in Turkey and mobilising an additional 3 billion euros (European Council 2016). Although this agreement has been very controversial and criticised on various grounds (Batalla Adam 2017), the European Commission has generally evaluated it positively on the basis that its implementation has led to significant decreases in the numbers of migrants attempting to cross the Aegean Sea to the Greek islands and of lives lost in the Aegean Sea (e.g., European Commission 2016, 2). However, this EU-Turkish deal had barely been agreed that the Turkish government took a threatening tone in April 2016. Continuing to exploit the fears of migration flows in European countries, Turkish Prime Minister Ahmet Davutoğlu emphasised that Turkey expected to receive funding and visa-free travel, as well as seeing progress in the EU membership negotiations, in return for controlling migration flows towards Europe. He stated that '[these] pledges are mutual and that '[if] the EU does not take the necessary steps, it would be unthinkable for Turkey to do so' (France 24 2016). In other words, 'obviously nobody [could] expect Turkey to hold its side of the deal' (France 24 2016). This followed similar statements made earlier on the same day by the Turkish Minister for Foreign Affairs Mevlüt Çavuşoğlu. He notably stated that '[if] the EU doesn't keep its word (. . .) we will cancel all agreements' (France 24 2016).
The Turkish government put renewed pressure on the EU in July 2016, seeking to exploit the vulnerability of the EU Member States in the field of asylum and migration again. In an interview, Foreign Minister Mevlüt Çavuşoğlu emphasised that the Turkish government viewed the promised lifting of short-term visa restrictions by the EU as a crucial matter and that its government could back out of the agreement with the EU by October, if the EU did not grant visa-free travel to Turkish nationals. He highlighted the 'very serious measures' taken by Ankara to stem the flows of migrants and asylumseekers towards the EU, but emphasised that 'all that is dependent on the cancellation of the visa requirement for our citizens, which is also an item in the agreement of March 18ʹ (Al Jazeera 2016). Çavuşoğlu specified his threat when he stated that '[if] visa liberalisation does not follow, we will be forced to back away from the deal on taking back [refugees] and the agreement of 18 March' (Nielsen 2016).
These direct threats in rhetoric were reinforced by indirect threats exercised through relaxed migration controls. Those enabled over 1,000 migrants and asylum-seekers to cross from Turkey to the Greek islands in the last two weeks of July (Nielsen 2016). This was a clear reminder by the Turkish government to the EU and its Member States that it had the power to increase the migration flows that have caused them so much concern. Importantly, these threats and the increase in the migration flows towards Europe coincided with the aftermath of the attempted military coup against Erdoğan, which saw sweeping and extremely controversial purges across Turkish society (Amnesty International 2017). Thus, the Turkish authorities made migrationrelated threats, which they backed up with a demonstration of their ability to increase migrant arrivals in Europe, just as they were engaged in implementing repressive measures against large sections of Turkish society. Although some voices expressed their alarm at the unfolding post-coup crackdown, the official reactions of European governments and of the EU institutions were rather restrained. For instance, the High Representative of the Union for Foreign Affairs and Security Policy, Federica Mogherini, merely called for the 'end of violence' and the 'protection of the rule of law' (Deutsche Welle 2016).
The Turkish government reiterated migration-related threats at the end of November 2016. Referring to Kapikule, a Bulgarian border checkpoint where large numbers of migrants and asylum-seekers had gathered, President Erdoğan told European governments that '[if] you go any further, these border gates will be opened' (Shaheen, Wintour, and Rankin 2016). Again, the Turkish government did not make this threat fortuitously. Erdoğan's threat of abandoning the agreement made with the EU came in response to the adoption of a resolution on EU-Turkey negotiations in the European Parliament, which called on the European Commission and the Member States 'to initiate a temporary freeze of the ongoing accession negotiations with Turkey' (European Parliament 2016). This resolution, which was largely symbolic given its non-binding character, was itself prompted by the purges that followed the failed coup attempt against Erdoğan in July 2016.
New migration-related threats were issued by the Turkish government in March 2017. Playing on the widespread fear of large flows of asylum-seekers and migrants amongst EU Member States, Foreign Minister Mevlüt Çavuşoğlu declared that Turkey 'may cancel' the agreements that it had signed with the EU in the field of migration (Saeed 2017). He specified that the Turkish authorities 'are not applying the readmission agreement at the moment, and (. . .) are evaluating the refugee deal' because of delays in the implementation of the visa liberalisation promised by the EU. Fellow Minister Suleyman Soylu, who is known as a hardliner close to President Erdoğan, went further by threatening to 'blow the mind' of European countries by letting go the '15,000 refugees that we don't send each month' (Euractiv 2017). Again, these migration-related threats against European governments and the EU were deliberately made by the Turkish government at a very specific time. They came amidst renewed diplomatic tension between Turkey and European countries, which saw Erdoğan compare the Dutch and German governments to the Nazi regime (Yilmaz 2019, 243). This followed the cancellation by several European governments of rallies organised by the Turkish government for local Turkish communities in the run-up to a referendum on constitutional changes aiming to concentrate more power in the hands of the President (BBC News 2017;Saeed 2017 Official reactions to the Turkish referendum in Europe were very restrained in public, as epitomised by a statement made by EU Commissioner for Enlargement, Johannes Hahn, in July 2017. He merely noted that the referendum 'raised concerns about the direction and the structure of some constitutional amendments' in Turkey (EUobserver 2017).
To date, the most advanced attempt by the Turkish authorities at exploiting the vulnerability of European countries and the EU when it comes to migration flows took place in winter 2020. On 27 February 2020, the Turkish government announced that it was opening its western borders and would no longer impede the movement of asylum-seekers and migrants trying to reach Europe (McKernan and Boffey 2020). In the space of a few hours, hundreds of people arrived at the Turkish border with Greece. According to various reports, the Turkish authorities did not merely tolerate the movement of migrants and asylum-seekers, but actually 'encouraged and facilitated movement to the Greek border' (Amnesty International 2020, 4). Amnesty International reported having 'observed buses travelling towards the [Greek] border in convoy escorted by ambulance and police vehicles, state officials paying for fuel for buses and gendarmerie officers disembarking people and instructing them to walk the short remaining distance to the river Evros border' (Amnesty International 2020, 10).
Erdoğan's move can be interpreted as an attempt at instrumentalising migration flows in order to put European countries, the EU and the North-Atlantic Treaty Organisation (NATO) under pressure to offer their support to Turkish military activities in Syria, in particular the 'Spring Shield' operation. This was launched after an airstrike -allegedly carried out by the Russian air force in the context of its wider support to the Syrian regime -had killed dozens of Turkish soldiers (McKernan 2020) in the midst of the battle for the control of the Syrian province of Idlib, which borders Turkey. President Erdoğan himself linked both issues when he declared to the Turkish Parliament that '[if] European countries want to resolve the issue [of migration], they must support Turkey's efforts for political and humanitarian solutions in Syria' (Pitel, Foy, and Cornish 2020). Importantly, this threat to Europe was also made just before Erdoğan was due to meet with Russian President Putin to discuss a possible ceasefire agreement (Pitel, Foy, and Cornish 2020). In addition, it was suggested that, as Turkish public opinion had become increasingly critical of the military casualties in the conflict in Syria, President Erdoğan must have also hoped that polemic on migration would shift attention away from domestic politics to criticising European countries (Aydintasbas 2020).
Over the next few days, thousands of people continued to move towards the Turkish border with Greece (Pitel, Foy, and Cornish 2020). On 4 March 2020, it was estimated that about 12,500 people had gathered on the Turkish side of the Turkish-Greek border (Oltermann 2020). Given the announcement of the Turkish authorities, they had expected to be able to cross the border and enter Europe (Amnesty International 2020, 10).
However, far from opening their border, the Greek authorities had rapidly stepped up border security measures and met these migration flows with a strong response. Teargas and stun grenades were fired to dissuade migrants and asylum-seekers from approaching and crossing the border (Oltermann 2020;Statewatch 2020). In a concerning development, asylum-seekers reported the use of violence by Greek border guards to repel them (Amnesty International 2020; Statewatch 2020). The Greek authorities also requested the deployment of rapid border intervention teams (RABITs), which are coordinated by Frontex with a view to giving immediate assistance to a Member State in a situation of urgent and exceptional pressure at its external borders Moreover, on 2 March 2020, a government decree was passed, which suspended the registration of asylum claims for a month for people entering the country irregularly from 1 March, enabling the authorities to return them to their country of transit or origin without their asylum claim having been registered (European Council on Refugees and Exiles 2020). The Office of the United Nations High Commissioner for Refugees (UNHCR) criticised this decision (UNHCR 2020). However, the European Commission declined to comment on the legality of the decision of the Greek authorities. A Commission spokesperson argued that '[the] commission is not a court and does not have the authority to have a definitive legal opinion' (Nielsen 2020). EU Home Affairs Commissioner Johansson merely recalled that ' [individuals] in the European Union have the right to apply for asylum' (Rankin 2020b). European Commission President von der Leyen went controversially further by praising Greece for 'being our European "aspida" in these times' (i.e. 'shield' in Greek) (Jamieson 2020), whilst announcing that it would receive €700 m in EU funds, notably for strengthening its border infrastructure (Rankin 2020a). The strong response of the Greek authorities also received the backing of the German government. Although protesters called for the European borders to be opened to asylum-seekers and migrants coming from Turkey, German Interior Minister Seehofer stated that the government did not want 'a repeat of the year 2015ʹ (Oltermann 2020). The Interior ministry also sent tweets in several languages in strong support of the Greek authorities: 'We need orderly conditions at the EU's external border. We will use our best efforts to support Greece in achieving this. Europe's borders are not open for #refugees from #Turkey, and neither are our German borders' (Oltermann 2020).
Confronted with this firm response, which indicated that European states would not allow for a new 'refugee crisis' to unfold, the Turkish government backed down and began to organise the transportation of migrants stuck near the Turkish-Greek border back to Turkish cities (Stevis-Gridneff and Kingsley 2020b). On 9 March 2020, Turkish President Erdoğan met with the President of the European Commission, Ursula von der Leyen, and the President of the European Council, Charles Michel, in Brussels. At the subsequent press conference, von der Leyen acknowledged that 'Turkey needs support', which 'involves finding a path forward with Turkey' (European Commission 2020a).
According to media reports, following this episode, the Turkish government continued to '[press] European leaders to make fresh cash pledges to prevent tens of thousands of refugees from leaving the country and trying to reach Europe' (Wintour and Smith 2020). Despite the slow pace at which returns to Turkey have taken place and the fact that the Turkish government decided to suspend returns altogether in March 2020 citing the pandemic, the EU agreed in July 2020 to provide an additional 585 million euros in humanitarian bridge funding after the EU Facility for Refugees in Turkey, in order to ensure the continued financing of the Emergency Social Safety Net (ESSN) and Conditional Cash Transfers for Education (CCTE) programmes (European Commission and High Representative of the Union for Foreign Affairs and Security Policy 2021, 8-9). Moreover, in December 2020, the European Commission confirmed that eight more contracts had been signed under the EU Facility for Refugees in Turkey, meaning that the operational budget of this six billion euros facility had been fully committed and contracted (European Commission 2020b). Thus, by threatening to send more migrants and asylum-seekers towards Europe and by demonstrating that it was able and ready to do so in practice, the Turkish government put pressure on the EU to ensure that the entire budget of EU Facility for Refugees in Turkey was committed and contracted. Moreover, despite that fact that it has not entirely fulfilled its end of the bargain when it comes to returns, the Turkish government has even managed to secure some additional funding to support those who have found refuge in Turkey.

Conclusion
This article set out to contribute to the debates on de-centring the analysis of migration governance in Europe by focusing on the potential role of external actors in the securitisation of asylum and migration in the EU. It addressed this hitherto neglected topic using the case of Turkey, which has considerably developed its relations on asylum, migration and borders with the EU in recent years.
Theoretically, this article has contributed to developing the securitisation framework by de-centring the study of securitisation processes. It has put forward the novel argument that, from the vantage point of an external actor, a securitisation process highlights the existence of a vulnerability to a specific phenomenon that is perceived to be threatening. An external actor can then decide to exploit this vulnerability for its own gain, notably by making threats that play on the fears of the other political actor.
Empirically, this article has demonstrated how the Turkish government has been able to exploit the vulnerability of European countries to migration flows, which had been highlighted by the social construction of asylum and migration as security issues. By repeatedly threatening to send more asylum-seekers and migrants Europe's way, the Turkish authorities have managed to secure some significant advantages in the last few years. Some of them have been financial, although the Turkish authorities have regularly complained about delays in receiving financial assistance from the EU. Others have been political, even though they have not all materialised yet, as shown by the slow progress made towards visa liberalisation. Unless the EU and European governments fundamentally change the way they socially construct migration flows, which appears unlikely at the time of writing, one can expect such dynamics to continue to characterise EU-Turkey relations in the years to come.