Double two-level games and international negotiations: making sense of migration governance in EU-Africa relations

ABSTRACT

This article conceptualises international negotiations between two groups of states as the result of double two-level games, whereby the two negotiating teams negotiate with their counterpart at the international level while simultaneously engaging with their constituents at their respective domestic level, trying to reconcile pressures from both arenas. Furthermore, it points to the importance of contextual and normative explanations and problem-solving approaches to explain negotiation outcomes. Empirically, it sheds light on one of the most contentious aspects of the Post-Cotonou Agreement between the European Union (EU) and the Organisation of African, Caribbean and Pacific States (OACPS), thanks to exceptional access to the negotiation process and subsequent interviews with most members of the negotiating teams. While the EU-OACPS Agreement may be dismissed by its critics as a relic of the past, the significance of its provisions on migration and mobility may mark a turning point in EU-Africa relations. A detailed mechanism, in fact, was agreed upon to ensure that African (as well as Caribbean and Pacific) states readmit unwanted migrants in the EU without conditionality and formalities other than a swift identity verification, with serious consequences if they fail to comply.

KEYWORDS:

• EU migration policy
• EU-Africa relations
• Post-Cotonou Agreement
• migration governance
• two-level game model

Introduction

When the European Union (EU) commenced negotiations on the successor to the Cotonou Agreement in September 2018, one of its main objectives – which collided with those of its counterpart, the African, Caribbean and Pacific (ACP) Group of states – was agreeing on a specific mechanism to swiftly repatriate irregular migrants, particularly from Africa (ACP Group 2018; Council of the EU 2018). The task at hand for EU negotiators was formidable, for at least three reasons. First, the ACP Group – renamed Organisation of African, Caribbean and Pacific States (OACPS) since April 2020 (Carbone 2021b) – had successfully opposed the inclusion of detailed provisions on return and readmission before the Cotonou Agreement was signed in June 2000 and then in its second revision in June 2010, at a time when the EU could have used the European Development Fund (EDF) as a side payment (Arts 2005; Crush 2015; Knoll 2017). Second, individual sub-Saharan African states had refused to conclude bilateral readmission agreements with the EU, despite the prospect of substantial financial incentives (İçduygu and Aksel 2014; Wolff 2014; Adam et al. 2020; Slominski and Trauner 2022). Third, the negotiating capacity of Africa had significantly improved, as witnessed by the contribution of the African Union (AU) to the 2018 Global Compact on Migration and, before that, the outcome of the 2015 Valletta Summit on migration and the 2017 AU-EU Summit in Abidjan (Maiyegun 2019; Schöfberger 2019).

The unexpected outcome of the negotiations between the EU and the OACPS on the issue of migration governance begs the following question: why does the Post-Cotonou Agreement contain a specific mechanism on return and readmission, when on other occasions the EU had failed to achieve the same objective? To address this conundrum, this article firstly introduces a conceptual framework, drawing on and contributing to the literatures on international negotiations and migration governance. In particular, building on existing theorisation and application of the two- or three-level bargaining model (Putnam 1988; Young 2003; da Conceição-Heldt and Mello 2017; Gstöhl 2021), it conceptualises EU-OACPS negotiations as the result of a double two-level game with the EU and the OACPS negotiating teams engaging simultaneously at both the international level and their respective domestic levels. Furthermore, it emphasises the importance of contextual and normative explanations, rather than prevailing rationalist explanations, to understand the outcomes of migration negotiations (Betts 2011; Reslow and Vink 2015; Hampshire 2016; Adamson and Tsouparas 2019). The second half of the article unpicks the dynamics underpinning the conclusion of the chapter on migration and mobility in the EU-OACPS Agreement, which potentially marks a turning point in one of the most controversial aspects of EU-Africa relations. In fact, the return rate of numerous African states, much lower than the world average (see Table 1), has been identified as a key indicator of an ineffective EU migration policy. It has also been used to allocate significant amounts of resources with the aim of tackling the root causes of migration, or to make the approval of visa applications more difficult for citizens from countries with low return rates (Mouthaan 2021; Stutz and Trauner 2022; Zaun and Nantermoz 2022). The central argument of this article is that the surprising conclusion of the EU-OACPS Agreement with regards to return and readmission is due to an enabling negotiation context which allowed the two negotiating teams to craft an optimal compromise, constructively engaging with their international counterparts while respecting the key concerns of their respective domestic constituents.

Table 1. Third country nationals ordered to leave the EU and return rate of ACP states.

	2008		2012		2016		2020
	Ordered to leave	Return rate (%)	Ordered to leave	Return rate (%)	Ordered to leave	Return rate (%)	Ordered to leave	Return rate (%)
Africa	81,320	14.09	71,370	10.62	64,495	7.86	72,445	5.06
Caribbean	4,620	13.96	3805	22.99	2,630	25.85	1,420	21.83
Pacific	15	0	10	100	35	42.85	5	0
World	533,520	30.71	434,280	30.38	426,255	45.15	396,435	17.71

Source: Eurostat

To reach this conclusion, it was necessary to open the ‘black box’ of the negotiations, which were conducted by a restricted number of officials and ambassadors in Brussels between September 2018 and April 2021. The EU negotiating team consisted of a small group of officials operating within the Post-Cotonou Task Force – which included officials of the Directorate General (DG) for International Partnerships (hereafter, DG IntPa) and the European External Action Service (EEAS) – assisted by other officials from other DGs and the EEAS, who periodically reported to the ACP Working Party of the EU Council.¹ The OACPS negotiating team was composed of a select group of ambassadors serving as members of the Central Negotiating Team, supported by the OACPS Secretariat, who regularly informed the OACPS Committee of Ambassadors. Negotiations were conducted behind closed doors and attracted limited public attention. In fact, this article benefits from exceptional access to primary sources (including unpublished documents) and direct engagement with the two negotiating teams, as the author actively participated in the preparation and negotiation rounds of the Post-Cotonou Agreement from December 2015 to April 2021. It also draws on 34 semi-structured interviews conducted by the author between May 2021 and January 2022 with people directly and indirectly involved in the negotiation process. These interviewees include seven members of the EU’s Post-Cotonou Task Force; six additional policy makers from DG IntPa, DG Home, the Legal Service of the European Commission (EC) and the EEAS; five representatives of the EU Member States in the ACP Working Party of the EU Council; nine ACP ambassadors, seven of whom were members of the OACPS Central Negotiating Team; four officials of the OACPS Secretariat; and two AU representatives.² Access to the negotiation rounds and subsequent interviews – further triangulated with press releases, policy makers’ social media communications, newspapers, and analyses done by think tanks and civil society organisations – allowed the author to ascertain the policy positions of the various actors involved in the negotiations with a high degree of accuracy.

Conceptual framework: double two-level games and migration governance

Migration has traditionally been regarded as a stronghold of domestic policy, and national sovereignty has been the dominant principle in the philosophy of migration governance, with economic interests and security considerations linked to the flow of people acting as an impediment to greater international cooperation. Nevertheless, a multitude of formal and informal initiatives have arisen to address mobility of people at bilateral, regional, or inter-regional levels, often having a North-South dimension underpinned by a managerial migration governance philosophy (Betts 2011; Kunz, Lavenex, and Panizzon 2011). This emerging philosophy of migration governance sees its full expression in (one of) the 2015 Sustainable Development Goals: migration can and should be governed, and when this happens it may be beneficial to receiving and sending countries and to the migrants themselves, thus resulting in a potential triple win (Pécoud 2021). EU-Africa cooperation on migration governance encapsulates these trends, as it includes bilateral instruments in the form of mobility partnerships or migration compacts and EU support to (and dialogue with) regional communities such as ECOWAS in West Africa or IGAD in Eastern Africa, and continental or trans-continental organisations like the AU and the OACPS (Schöfberger 2019; Zanker 2019). All these initiatives seek to balance the costs and benefits of migration and mobility but share the same key objective: tackling irregular migration. In fact, despite official figures unequivocally showing that most African migration occurs within Africa, more than being a South-North phenomenon (Flahaux and De Haas 2006), migration from Africa to Europe is generally portrayed as uncontrollable, and thus European leaders are under pressure to effectively address the issue (Knoll 2020).

To better understand EU-Africa cooperation on migration, three interlinked literature strands are helpful. The first strand looks at negotiation processes, and specifically uses the two-level game metaphor, in the case of the EU extended to three levels to include the European level (EU institutions) along with the national (EU Member States) and the international (EU and third actors) levels, to show how chief negotiators play different games simultaneously. Their goal is to find a compromise which satisfies the other negotiating party while also being acceptable to their domestic constituents and legislators so that the negotiation outcome can be ratified; thus, to reach agreement, chief negotiators must define win-sets that can overlap with the win-sets of their counterparts (e.g. Putnam 1988; Elgström and Strömvik 2005; da Conceição-Heldt and Mello 2017; Gstöhl 2021). A second strand refers to migration diplomacy, which is interested in the strategic use of migration flows to reach other objectives or the use of diplomatic methods to achieve goals related to migration. Like traditional diplomacy, migration diplomacy is affected by the interests of states and the power relations between them, with issue-linkages used to compel weaker actors to accept undesirable provisions (e.g. İçduygu and Aksel 2014; Tsourapas 2017; Adamson and Tsourapas 2019; Geddes and Maru 2020). A third strand is concerned with the externalisation of EU migration governance, which investigates the EU’s attempts to shift the challenges of migration (essentially unwanted migrants) away from its Member States. In return for their cooperation, the EU promises third countries support in other areas, including development, trade, or conflict prevention (e.g. Boswell 2003; Lavenex 2006; Wunderlich 2012; Hampshire 2016; Adam et al. 2020; Geddes et al. 2021; Zaun and Nantermoz 2022). The response of African states has varied significantly, ranging from collaboration to confrontation or disdain. Thus, power asymmetry does not necessarily result in cooperative behaviour by weaker actors; in fact, African countries have increasingly shaped the EU’s negotiating positions and shifted power relations in their favour (Mouthaan 2019; Adam et al. 2020).

This article builds on these insights and adds to them in various ways. Rather than the result of a two- or three-level game, it conceptualises EU-OACPS negotiations as the result of a double two-level game, or three games played at two levels that are closely linked: the first level (‘international’) involves negotiations between the EU and the OACPS, while the second level (‘domestic’) refers to interactions of both the EU and the OACPS with their respective constituents (see Figure 1). One of the main appeals of the two-level game model is that it links domestic politics and international relations; yet, the domestic level, particularly in EU studies, is often inadequately integrated (Young 2003). It should be considered, in fact, that supranational negotiators are not necessarily accountable to their domestic constituents but to national policy makers operating at the EU level (İçduygu and Aksel 2014; Weinhardt and Moreland 2018). Therefore, this article treats ‘Brussels’ as the EU’s domestic level, but it unpacks it as much as possible. In this regard, it gives more weight to the positions of Member States expressed in the Council (including its Working Parties and Coreper), as this is the venue where problems related to the future approval of an agreement are pre-emptively raised, but it also takes into account pressures coming from various DGs, which tend to push their sectoral interests (Frennhoff Larsén 2007; Mouthaan 2021). The negotiating team of the EU’s counterpart, when it is made up of a cluster of states or a regional organisation, faces similar constraints, as pressures can come from the national or the supranational level. It may transpire, however, particularly in the case of developing countries, that domestic constituents remain detached from the negotiating process, due to states’ weak coordination structures or limited capacities, and thus supranational negotiators may not fully appreciate the preferences of each state they represent (İçduygu and Aksel 2014; Weinhardt and Moerland 2018).

Figure 1. Double two-level game in EU-OACPS negotiations.

The ‘international level’ refers to the long-standing relationship between the EU and a group of ACP states: 48 states in sub-Saharan Africa, 15 in the Caribbean, and 15 in the Pacific.³ Despite the EU’s emphasis on the principle of partnership, EU-ACP negotiations are generally regarded as asymmetrical, with EU internal talks, often lengthy and adversarial but usually followed by an external display of strength and unity, being more influential on negotiation outcomes than the actual negotiations with ACP states (Elgström 2000, Forwood 2001). Perhaps the sole exception is the conclusion of the first Lomé Convention (1975–80), due to a favourable international context combined with the strong negotiating skills of some visionary ACP leaders (Carbone 2021b). Since then, negotiations have largely reflected the EU’s positions, with the ACP Group generally seeking to contain damage, lamenting coercion, and shaming the EU for lack of generosity (Ravenhill 1985; Carbone 2013). This power asymmetry is also a consequence of the fact that many ACP states have not invested in the ACP Group, which has remained a rather weak inter-governmental organisation, highly dependent on EU funding and increasingly divided due to the growing diversity of its membership – and unsurprisingly, all this has compromised the ACP Group’s international agency, which in part explains the process of transformation into the OACPS. Its performance has also been affected by the fact that negotiations with the EU are times conducted by inexperienced ambassadors (with several notable exceptions), largely disconnected from their capitals, and supported by a rather understaffed OACPS Secretariat (Keijzer 2016; Carbone 2018; Carbone 2020; Carbone 2021b).

The ‘domestic’ level for the EU in the area of migration governance is particularly contentious. Existing literature has documented how tensions between and within the European Commission and the Council constrain the ability of the EU to be an effective actor in international settings (Reslow and Vink 2015; Hampshire 2016; Badell 2020). On the one hand, EU Member States, which experience different levels of migration and have different needs and histories in terms of immigration and colonialism, advance diverging views on migration governance: some are more prone to inter-state cooperation and formulate common approaches, while others are more protective of their national sovereignty and tend to shift blame to third states (Hampshire 2016; Schöfberger 2019; Knoll 2020). On the other hand, diverse views co-exist within the EU’s supranational institutions: DG Home has traditionally promoted the securitarian narrative and DG IntPa the migration-development nexus, whereas the EEAS, which is perceived as lacking a clear vision, has sought to balance different aspects while preserving good diplomatic relations with third states (Hampshire 2016; Knoll 2020; Mouthaan 2021). However, there seems to be an emerging consensus that, after the 2015 migration crisis, the EU has made a deliberate effort to overcome competing interests and adopt a strong bargaining position vis-à-vis third actors (Knoll 2020; Mouthaan 2021; Zaun and Nantermoz 2022).

The ‘domestic’ level of the OACPS when it comes to migration governance is even more complex. An element to consider is that this is not an area of central concern for the Caribbean and Pacific states (Crush 2015; Knoll 2017). As for Africa, generally governments have not treated migration as a priority issue, in part because they face other significant challenges, but also because they see mobility as a natural phenomenon, driven by processes of development and social transformation (Flahaux and De Haas 2006; Geddes and Maru 2020). Nevertheless, several African states have gradually adopted policies and action plans with particular emphasis on fighting irregular migration, with the EU being instrumental in their promotion. This is the case for West Africa, where the EU’s externalisation of migration governance has clashed with and in part weakened the regional project for the free movement of people (Bisong 2019). Migration has traditionally not been very politicised in ACP states, except again for West Africa where it has attracted more interest from the media, civil society organisations, and the diaspora, as well as elected politicians: this reflects the economic importance that different countries attach to migration, but also depends on the role that former colonial powers play in the region (Bisong 2019; Zanker 2019; Adam et al. 2020). Lastly, the AU has emerged as a credible actor in coordinating positions in international settings, and has become forceful in defending the relevance and predominance of its own migration policy frameworks for the whole of Africa (Maiyegun 2019; Knoll 2020; Maru 2021).

As shown in Figure 1, in a double two-level game, the two negotiating teams negotiate with their counterpart at the international level while simultaneously engaging with their constituents at their respective domestic level, trying to reconcile pressures from both arenas. In doing so, they must consider the ‘cost of no agreement’, remaining as close as possible to their negotiating mandates or taking account of their domestic win-sets if the mandate is not explicit. Importantly, they are not mere agents of their domestic constituents, and their actions and strategies are not necessarily driven solely by interests and calculations, with issue-linkage always requiring the use of side payments as suggested by the migration diplomacy literature (Adamson and Tsourapas 2019; Putnam 1988; Gstöhl 2021). Norms, socialisation, and persuasion, in fact, may also play a role in shaping how actors understand different issues and the relationship between them, in reducing clashes between the negotiating parties, and in reaching mutually beneficial agreements; negotiation behaviour and outcomes may also be affected by the proposal of innovative solutions outside the parties’ initial positions, reframing of issues, communication of new information more openly and fully, or changes in a party’s perceived alternatives during the negotiations (Odell 2010). Furthermore, problem solving, not just (hard) bargaining, is also an important logic underpinning international negotiations (Elgström and Jönsson 2000) or, to put it differently, there may be cases of coercive diplomacy, where one party manages to impose its preferences, but also cases of cooperative diplomacy, which may benefit all sides in the negotiations (Tsouparas 2017).

Empirical analysis: migration and the Post-Cotonou Agreement

The framework elaborated in the previous section is used in this section to understand the negotiation dynamics that characterised one of the most contentious aspects of the post-Cotonou negotiations (Carbone 2021a). Interestingly, migration governance was also a very divisive issue in the negotiations of the Cotonou Agreement. This was due to a last-minute proposal on return and readmission, strongly pushed by the EU following indications of the Tampere European Council of October 1999. The ACP Group reluctantly accepted the inclusion of a general obligation into Article 13 stipulating that each ACP state ‘shall accept the return of and readmission of any of its nationals who are illegally present on the territory of a Member State of the European Union, at that Member State’s request and without further formalities’ (EU 2000) – and mutatis mutandis the same obligation applied to EU Member States. It was also agreed that the operationalisation of this readmission clause would be left to subsequent bilateral readmission agreements (Elgström 2000; Arts 2005). Similarly, before the second revision of the Cotonou Agreement was signed in June 2010, the EU proposed making repatriations of unwanted migrants self-executive (meaning that there would be no need for complementary bilateral readmission agreements). This approach was rejected by the ACP Group, with the two sides deciding not to make any changes to Article 13, not even for those elements where there was consensus; eventually, a bland joint declaration committing the two sides to engage in dialogue was attached to the revised Cotonou Agreement (Crush 2015; Knoll 2017).

The issue of return and readmission was addressed in other contexts, not least because the number of irregular migrants increased after 2010. For instance, it was high on the agenda of the 2015 EU-Africa Summit on migration held in Valletta, when the EU sought, in vain, to ‘sneak in at the last moment’ language on the use of EU travel documents to facilitate swift repatriation of irregular migrants (Interview 9). Similarly, the Abidjan AU-EU Summit of November 2017, which was celebrated by both sides as the start of a real ‘partnership of equals’, was marred by tensions on migration governance, as the AU resisted prescriptive language proposed by the EU on forced return and instead insisted that returns must be voluntary (Agence Europe, 1 December 2017; Africa Confidential, 6 April 2018; Interviews 9 and 10). Most importantly, since the early 2000s the EU has sought to negotiate various readmission agreements with individual African states, but its efforts have proved futile as the offering of various incentives (for instance, border management and operational support, visa facilitation, financial assistance) could not compensate for loss of remittances and the societal costs to reintegrate returnees. Furthermore, civil society and the media often raise concerns related to (perceived) violations of human rights, and thus African governments do not want to be seen as facilitators of the deportation of their own citizens. For this reason, the EU has increasingly resorted to informal arrangements, which are less in the public eye, to allow a more pragmatic and flexible approach to return and readmission (İçduygu and Aksel 2014; Wolff 2014; Mouthaan 2019; Adam et al. 2020; Slominski and Trauner 2022).⁴

Unpacking the negotiation process

The expiration of the Cotonou Agreement, with negotiations for its successor to start no later than September 2018, provided another opportunity to raise the question of return and readmission, made even more urgent by the 2015 migration crisis. The agenda was set by the EC, jointly with the EEAS, which proposed ‘migration and mobility’ as one of the six strategic objectives for the Post-Cotonou Agreement and called for the achievement of ‘a comprehensive, coherent and balanced approach’ between the positive and negative aspects of migration and mobility (EC 2017). Yet, it was clear that its main objective – and here the fingerprints of DG Home were clearly visible (Interviews 17 and 18) – was to ‘set-up a mechanism to ensure that this obligation [that is readmission] is effectively fulfilled’ by ACP states, most notably ‘through identification of their nationals, issuance of travel documentation for return or acceptation of the EU travel document for return, [and] timely response on formalities of return operation’ (EC 2017). Indeed, several EU Member States had expressed dissatisfaction at the poor return rate of African countries and, as one EU official admitted, they had threatened that ‘there would be no Post-Cotonou Agreement without a mechanism on return and readmission, and this was the main reason to keep relations with the ACP Group alive’ (Interview 13). The EC proposal was largely endorsed by all Member States, and the EU negotiating mandate, which was rather detailed, was adopted in June 2018, after Hungary had held the EU hostage for several weeks with protracted negotiations at different levels in the Council (involving the ACP Working Party, Coreper, and the Foreign Affairs Council), lamenting inadequate appreciation of the negative consequences of irregular migration (Agence Europe 25 May 2018, 29 May 2018, 22 June 2018).⁵ While Hungary’s concerns were raised publicly, it should be noted that other EU Member States were extremely cautious on expanding pathways for legal migration, and were eager to have a stringent mechanism on return and readmission (Interviews 20, 21, 22, and 23).

The negotiating mandate of the ACP Group adopted in May 2018 in Lomé was brief and vague on migration, included under the pillar of ‘political dialogue and advocacy’, which was one of the three proposed pillars for the Post-Cotonou Agreement together with development cooperation and economic cooperation. The short section on migration contained four generic paragraphs, with nominal references to legal migration, intra-ACP migration, remittances, respect of human rights, and political dialogue, whereas on return and readmission it stated that returns ‘should be on a voluntary basis’ (ACP 2018). It was evident that ‘ACP states did not want to make migration one of their priorities for the new agreement’ (Interview 1) and that, considering the widely publicised EU interest in return and readmission, the main goal of the OACPS, as an ACP ambassador put it, ‘was essentially to arrive at an agreement on a process, which would allow dialogue and ensure that the EU Member States would not dump people on our doorsteps at the airport’ (Interview 5).

Negotiations on the Post-Cotonou Agreement commenced in September 2018, but the part on ‘migration and mobility’ was not dealt with until May 2019 (Carbone 2021a). The two negotiating teams – for the EU, the EU Post-Cotonou Task Force, and for the OACPS, the ambassador of Zimbabwe together with the ambassadors of Samoa and Guyana, assisted by two or three OACPS officials – easily found an accord on the positive aspects of migration, notably on joining forces to enhance the migration-development nexus, improve international protection, combat trafficking and smuggling of migrants, and develop aspects linked to legal migration yet in full respect of national competences. Various observers and civil society organisations, however, have complained that these provisions employ ‘extremely vague language’ and show ‘little commitment to concrete actions’, far from the ‘highly detailed and concrete’ language, including ‘specific timeframes for action’, used for return and readmission (Concord 2021, 17 and 18). This outcome is the consequence of the fact that, as admitted by an ACP ambassador, ‘the OACPS did not push enough on the positive aspects of migration, as our mindset was to limit the EU capacity to act unilaterally on return and readmission’ (Interview 5). Another element to consider is the fact that the migration constituency within DG IntPa sought to better operationalise the development-migration nexus, but their proposals were ‘internally blocked’ (Interview 14), because ‘it was not clear what the basis was for their suggestions, considering that the political line from the Member States and College was different’ (Interview 12).

The negotiation of the provisions on return and readmission, conversely, was controversial. The first move was made by the EU negotiating team, who expanded on the EU negotiating mandate to propose a very detailed mechanism, largely in line with readmission agreements, to improve return rates. The rationale of the initial EU proposal, as bluntly put by an EU official, ‘was to eliminate the other side’s administrations to a maximum possible, because African administrations are often inefficient and obstructive’ (Interview 12). Return and readmission requires close cooperation from the requested states, yet many developing countries do not have capacity to deal with readmission requests, or refuse to issue travel documents to people without sufficient documentation, particularly if migrants dispute their origin, destroy their identification documents, or claim no nationality to avoid forced returns (Stutz and Trauner 2022). The EU specifically proposed that returns should be ‘without conditionality’, fearing that ACP states could limit the number of returns, exclude certain categories of migrants, or make returns dependent on the adoption of reintegration packages.⁶ It also proposed very strict time limits for the procedure of readmission (ten days for the requested state to respond, which could be extended by five extra days in exceptional circumstances) and to issue EU travel documents if the requested state did not respond in a timely manner and the requesting state presented reasonable evidence for the identification of the irregular migrant. Remarkably, an EU official admitted that ‘to have all this agreed with 78 states altogether where the EU had failed with individual states was quite an endeavour, an optimistic scenario’ (Interview 11).

The OACPS reacted negatively to the EU proposal of a prescriptive return and readmission mechanism and for several weeks refused to even engage as, in the words of an OACPS official, ‘certain ambassadors thought that the EU proposed something that was disrespectful and even insulting, and certainly was deemed to be against the sovereignty of states to decide on their own who should be admitted in their territory’ (Interview 2). Following numerous negotiation sessions, the OACPS eventually committed to responding to readmission requests within 30 days (which is indeed the internationally agreed deadline set out by the Convention on International Civil Aviation), and to issue travel documents within ‘the shortest possible time’ in cases where the EU provides adequate evidence identifying irregular migrants. It accepted that returns should not be restricted to the use of scheduled flights, but rejected the proposal to use EU travel documents, unless these could be verified by the requested state (EC and OACPS 2021) – and, interestingly, an EU official acknowledged that this was ‘a serious red line for ACP states, but it was not surprising, because our arguments to be frank were not so strong, as there is no international reference for this: it’s a matter to be solved between states’ (Interview 11). Finally, the OACPS accepted, albeit reluctantly, the inclusion of a provision stipulating that failure to swiftly respond to readmission requests could trigger a formal procedure that could result in some forms of sanctions (‘proportionate measures’) – but, to use the words of an ACP ambassador, ‘this was a fair price to pay for having toned down all EU initial proposals’ (Interview 7).

Explaining the negotiation outcome

The compromise was achieved at the end of November 2020, when the win-sets of both sides eventually overlapped, with the operational provisions on return and readmission included in an Annex, which is legally binding. Within the EU, reactions were overwhelmingly positive: as claimed by an EU official, ‘the Post-Cotonou Agreement may have solved our problems and put an end to a discussion which has lasted for more than twenty years; if implemented well, technically we would not need to negotiate bilateral readmission agreements’ (Interview 8). Various policy makers from the EU Member States were happy with the work of the EU negotiating team, not only because the outcome on migration was in line with the negotiating mandate (Interviews 20, 21, 22 and 23) but also because it was ‘by far more ambitious than we could expect considering the initial ACP position and also what we couldn’t get in the Cotonou Agreement’ (DevEx, 14 January 2021). Hungary, however, opposed the compromise, claiming that some of the commitments on legal migration could result in more migrants into the EU (Reuters, 20 May 2021; DevEx, 14 March 2022) – interestingly, one of its representatives acknowledged that ‘it is much more than we could have expected on return and readmission, but it is still too positive on legal migration, so even if the language was softened, we would not support it, for political reasons’ (Interview 22). An alternative, in fact complementary, reading was advanced by different OACPS representatives. An official maintained that ‘from a visual point of view, it may appear that the EU has won, as they got a mechanism on return and readmission, but in reality there is nothing new beyond what we have already accepted in other settings’ (Interview 3). An ambassador even welcomed the inclusion of ‘a clear mechanism: if you have something clear, it can be respected, and if you do not respect it, there are consequences, and nobody wants to have sanctions applied against them’ (Interview 8).

Achieving a compromise was difficult because the two negotiating teams had to reconcile diverging pressures coming from their counterpart (at the ‘international’ level) and their respective constituents (at the ‘domestic’ level). Within the OACPS, the negotiating team was confronted with a (latently) divided Committee of Ambassadors, as priorities on migration vary by region. On the one hand, there were several ambassadors, and the number grew as time passed, who were of the view that ‘if these are our citizens, why we should resist? Instead, we should come up with some clear rules and mutually beneficial arrangements for their return’ (Interview 4). This was the case of the Caribbean and the Pacific, but also Southern Africa and some states in Central and Eastern Africa, whose return rates from the EU are above the world average, or have no problems with irregular migration to Europe; as an OACPS official noted, ambassadors from these states ‘started becoming a bit impatient to conclude the negotiations, as there was some uncertainty if the current EU-ACP arrangement would lapse’ (Interview 6). On the other hand, there was significant resistance from West Africa (and some East African states), as confirmed by several ACP ambassadors: ‘Countries like Senegal, Côte d’Ivoire, and Nigeria send many people to Europe, so you can expect that they can’t have the same views as others: they were very vocal, and at times it was difficult to act as a group’ (Interview 7). These different sensitivities created some frictions, as admitted by an OACPS official: ‘Some ambassadors from West Africa were afraid that the lead negotiators could make concessions to the EU on things for which they had no mandate; in some cases, lead negotiators were even asked to backtrack from an agreed compromise’ (Interview 1). These divergences were eventually resolved when the OACPS negotiating team decided to ‘socialise’ the most outspoken West African ambassadors by inviting them to the negotiating table, which meant that ‘they participated actively in the discussions, though they did not add much to the substance of migration cooperation; but they understood the constructive spirit under which the negotiations were conducted, and eventually a compromise was achieved more easily’ (Interview 4).

Within the EU, the negotiating team faced pressure from most Member States, as well as from DG Home. An EU official confirmed that ‘we had to do a lot of convincing with the Member States in various sessions of the Council’ (Interview 15), as their widely shared line was to obtain the maximum result from their negotiating mandate, as acknowledged by a policy maker from an EU Member State: ‘We knew that the EU wouldn’t get all it wanted on return and readmission, but we were asked by our capitals to raise this issue over and over again, and for many of us this was a political line’ (Interview 22). DG Home also put strong pressure on the EU negotiating team for more stringent, yet in some instances unrealistic proposals as lamented by a member of the EU negotiating team: ‘It was strange that DG Home pushed hard for us to do things they did not obtain when they were chief negotiators; they wanted us to solve their problem, but we, not them, had to strike the right balance between different issues’ (Interview 12; also Interviews 17, 18 and 19).

A sort of parallel game was played beyond the official negotiating table by the senior management of the EC and some EU Member States aiming to co-opt the most recalcitrant members of the OACPS, specifically in West Africa (Interviews 11, 12, and 13). A first attempt was made by the EU’s Chief Negotiator, Commissioner Jutta Urpilainen, both informally with her OACPS counterpart Minister Robert Dussey, and formally in the political negotiating rounds with different ACP ministers. In July 2020, Urpilainen even issued an ultimatum of a few days for the OACPS to accept the EU’s proposal on the return of undocumented migrants, but that initiative, which surprised the EU negotiating team, risked being counterproductive and was simply ignored (Interviews 2, 3, 7, 8; DevEx, 10 July 2020). A second try was for ‘the best placed Member States to reach out directly to a few West African states’ (Interview 20), and that meant, for example France talking directly to Senegal (Interviews 10, 11, 12). Interestingly, an ACP ambassador stated: ‘I have a long experience of negotiating with the EU, so we knew that sooner or later they would contact our capitals, but our capitals resisted, and that strengthened our negotiating position in Brussels’ (Interview 8). This was also the impression shared within the EU negotiating team: ‘Whoever is familiar with the psychology of diplomacy knows that these types of activities often are counterproductive, as they expose that country to the others, and it would be hard to justify why they changed position’ (Interview 11).

Clearly, the EU and the OACPS held stances on migration governance which initially seemed irreconcilable, but the logic that characterised their interactions, besides some episodic tactical moves, was problem solving, that is finding a solution to the problem of irregular migration, as confirmed by various members of the two negotiating teams. A first EU official acknowledged that the two sides ‘managed to build a very constructive atmosphere on some very critical points, particularly on migration, so each issue was presented by the two sides with reasonable arguments’ (Interview 12). A second EU official noted that ‘there was a significant amount of trust among the negotiators; many issues were discussed informally, at times away from the main negotiating room; it was not a planned strategy, it was the natural way of doing things’ (Interview 10). As for the OACPS, an ambassador who was part of the negotiating team, declared to the EU: ‘we have a long history behind us, we are partners, so let’s try to find the best balance between our different positions and let’s avoid crossing red lines’ (Interview 8). A second ambassador claimed that ‘EU-OACPS negotiations are normally regarded as asymmetrical, but this time was different: we understood that the EU negotiating team was under pressure from its Member States, so we worked hard to find a compromise that could suit both sides’ (Interview 4).

Conclusion

Cooperation on migration and mobility between the EU and Africa is a complex proposition, as it includes bilateral and inter-regional arrangements, both formal and informal. The main objective of all these initiatives has been to achieve a balanced approach between legal and irregular migration, yet the 2015 migration crisis has added pressure on European policy makers to effectively address the surge of irregular migrants. The EU, therefore, has sought on various occasions and in different settings to compel third states to accept the return of their irregular migrants, by including (more or less detailed) readmission clauses in framework agreements and by signing specific readmission agreements with individual states. This article has shed light on how migration governance is dealt with in the Post-Cotonou Agreement, initialled in April 2021 and expected to be signed in 2023 – with delays caused, in large part, by Hungary’s opposition to the commitments agreed upon on legal migration. While the EU-OACPS cooperation framework may be dismissed by its critics as a relic of the past and unfit to address contemporary challenges, the significance of the provisions on migration and mobility may mark a turning point in EU-Africa relations. In fact, the long sought-after bilateral readmission agreements may no longer be needed as the Post-Cotonou Agreement contains a detailed mechanism to ensure that African (as well as Caribbean and Pacific) states readmit their citizens illegally present in the territory of an EU Member State without conditionality or formalities other than a swift identity verification, with serious consequences if they fail to comply.

This article has provided some explanations for this outcome, which was unexpected because African states had managed to counter protracted pressure at bilateral and inter-regional levels, but also because the EU, at least officially, was not willing to link it to compensatory financial resources. Contrary to expectations that the EU would co-opt its negotiating counterparts, it has been demonstrated that the EU and the OACPS struck a deal that met their respective demands without causing significant harm to either side: in fact, the two negotiating teams engaged simultaneously in two different tables (in a sort of a double two-level game) to craft language that could be accepted by EU Member States and African states. This outcome was also possible because of the specific forum in which negotiations took place: African ambassadors were somewhat insulated from domestic pressure, exercised when readmission agreements are negotiated bilaterally; the Pacific and the Caribbean (but also different regions in Africa) successfully highlighted the importance of finding a balance with other aspects of the Post-Cotonou Agreement; and the EU negotiating team persuaded some Member States in the Council and different units within EU institutions to tone down their most unrealistic demands.

These conclusions, which are reached thanks to exceptional access to all EU-OACPS negotiation rounds and subsequent confidential interviews (34 in total), have broader implications which merit further research. On the one hand, they point to the relevance of alternative (contextual and normative) explanations for the emerging migration diplomacy literature, which so far has rested on rationalist explanations. They also highlight the fact that developing countries do not always benefit from negotiating as part of a larger group: this is the case for some African states, which have managed to extract concessions or resist pressure when negotiating with the EU bilaterally, but may well be the case also for other states and regional groups in other circumstances. On the other hand, this article probes prevailing views that the EU’s internal negotiations are the best predictor for the outcome of the EU’s international negotiations, particularly in the case of developing countries, where the EU has often pursued a ‘take-it-or-leave-it’ approach. Indeed, more attention must be paid to the internal dynamics within the EU’s negotiating counterparts, and thus the double two-level game metaphor presented here may be appropriate for other inter-regional negotiations as well.

Acknowledgments

An earlier version of this article was presented at Maastricht University (25 November 2021), Vrije Universiteit Brussel (3 December 2021), Ghent University (7 December 2021), and the European University Institute (9 February 2022): many thanks to Sophie Vanhoonacker, Florian Trauner, Jan Orbie, and Andrew Geddes for inviting me and the participants in the four seminars for their suggestions. I am also grateful to Niels Keijzer, Mehari Maru, Sebastian Steingass, Gerasimos Tsourapas, and the peer reviewers for their comments. I am indebted to the members of the EU and OACPS negotiating teams for the numerous exchanges we had between 2015 and 2021 and other officials who accepted to be interviewed.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Note that DG for International Cooperation and Development (DG DevCo) was renamed DG IntPa in January 2021. Jutta Urpilainen, as member of the EC led by Ursula von der Leyen (2019–2024), replaced Neven Mimica as EU Chief Negotiator in December 2019.

2. All interviews, following ethics approval by the College of Social Sciences Research Ethics Committee of the University of Glasgow (application number: 400200183), were conducted either on the phone or via Zoom in English, French and Italian. No formal interview was conducted with any representative of the European Parliament or civil society, but the author had the opportunity to informally engage with a handful of them throughout the negotiation process.

3. The OACPS consists of 79 states, though Cuba is not a signatory of the Cotonou Agreement.

4. Cape Verde is the only African state that has signed a formal readmission agreement with the EU, whereas Côte d’Ivoire, Ethiopia, Guinea, and The Gambia have agreed on informal arrangements (Stutz and Trauner 2022).

5. The compromise that was eventually found included some minimal yet significant changes: the need to separate national from supranational competences was clearly spelt out; language on irregular migration was somehow toughened; some references on the migration-development nexus were deleted (Carbone 2021a).

6. The choice of the term ‘conditionality’ in this context may be misleading, as this generally refers to the conditions that donors attach to aid disbursements.