When politics trumps strategy: UK–EU security collaboration after Brexit

Both the United Kingdom (UK) and the European Union (EU) have significant incentives for close collaboration in foreign, security and defence policies, given their shared strategic interests, the clear potential for efficiency savings in working together, and the intensity of prior working relations. That the recently negotiated EU–UK Trade and Cooperation Agreement contains no provisions in this area is thus puzzling for followers of European security, who predicted prompt agreement, and for theories of international cooperation, which emphasise the importance of shared threats, absolute gains and prior interaction. We argue the failure to reach such an agreement stemmed from the politics of the withdrawal process itself, which resulted in acute problems of institutional selectivity, negotiating dynamics that polarised the relationship, institutional change that made an agreement less likely, and distributional scrabbling to supplant the UK. Our findings show that the dynamics of moving away from existing forms of cooperation are highly distinct from those motivating cooperation in normal times.


Introduction
Brexit -the withdrawal of the United Kingdom (UK) from the European Union (EU) -has been the subject of much interest, not least given the implications for European integration, for British politics, and for the prospects of international economic cooperation more broadly. While Brexit was highly disruptive economically, initial talk of its implications for security and defence cooperation were more positive, with the mutuality of strategic interests on both sides assumed to increase the prospects of an agreement. After all, the EU was losing a powerful politico-military actor, while the UK was losing a valuable forum for foreign and security policy coordination, all against the backdrop of a deteriorating regional security landscape. Moreover, some of the major obstacles to collaboration in other domains affected security cooperation less, since strategic interests were regarded as indivisible, levels of politicisation relatively low, and the EU's security and defence policy a largely intergovernmental affair.
Yet efforts to negotiate a comprehensive agreement on UK-EU security and defence collaboration did not play out as expected. The UK government's proposals for a bespoke partnership were rebuffed by Brussels, which objected to the UK's presumed right to participate informally in EU structures and to continue to collaborate on specific initiatives as a non-member. Proposals for an agreement were watered down in mid-2018 under Theresa May's premiership, before being further diluted by her successor, Boris Johnson, who finally ditched foreign and security policy issues from the negotiations in March 2020. The resulting EU-UK Trade and Cooperation Agreement (TCA), which provisionally entered into force on 1 January 2021, notably omits any mention of collaboration on foreign, security or defence policy (Cardwell, 2021).
Why should it prove so difficult to reach agreement on security and defence between two actors with shared strategic interests, a significant economic interest in efficiency savings, and a shared history of recent and deep interaction? After all, theories of international cooperation emphasise these very conditions -shared threats, mutual interests, prior interaction -as key facilitators of institutionalised collaboration. In this article we argue that the difficulties of negotiating an agreement on security and defence stem from the politics of the withdrawal process itself, which has undermined efforts to negotiate a deal in four respects: (a) First, by raising problems of institutional selectivity, including concerns for decision-making autonomy on both sides, worries about the consequences of selective engagement, and a risk of incentivising further departures; (b) Second, by establishing negotiating dynamics that introduced further difficulties, including the hardening of preferences, the politicisation of security and defence, and greater centralisation of decision-making on both sides; (c) Third, by precipitating and/or accelerating change on both sides, unleashing efforts to show meaningful divergence and facilitating reforms previously precluded; (d) And finally, by establishing new distributive dynamics, including demand for the spoils of the UK's departure and its former role(s), and competition over the institutional format(s) governing future cooperation.
A focus on the politics of withdrawal helps us to understand why a post-Brexit agreement in security and defence has thus far proven elusive for both sides. But it is also of broader theoretical relevance, helping us to theorise the distinct dynamics associated with the politics of exit in multilateral settings. Unsurprisingly, Brexit has unleashed a flurry of scholarship on the consequences of institutional withdrawal, but the distinct drivers underlying these processes have yet to be fully teased out.

The ease of the deal? Toward a post-Brexit security partnership
Managing the fallout from the Brexit vote in security and defence was expected to be the easy part, at least compared to the tricky issues involved in disentangling the UK's economic, regulatory and trading relationship with the EU . For one thing, the strategic incentives for security and defence cooperation are significant. Both the UK and the EU share the same security milieu within Europe, facing a number of serious security threats, including a resurgent Russia, civil conflict in the near-abroad, fears of transatlantic decoupling, and a heightened threat of terrorism (Sus and Hadeed, 2020: 433-434;Whitman, 2016: 519). Moreover, security and defence issues are, at least among allies, seldom subject to the kind of distributional bargaining or relative gains concerns which economic issues often are, partly because security is indivisible -if one actor loses, all actors lose -but also because of the high stakes involved and the tendency for security concerns to trump economic and identity-based ones at times of growing international insecurity. It was also clear that both sides would lose out from Brexit, since the EU was losing a major security and defence contributor and a significant global diplomatic and military presence, and since the UK would lose access to key forums for coordinating with its European partners (Martill and Sus, 2018: 853;Whitman, 2019: 528). The UK's relative strength in this area might have been expected to level the playing field, increasing the amenability of the EU to bespoke security and defence arrangements.
The strategic incentives for security and defence collaboration, moreover, have intensified since the Brexit vote. Fears of American disengagement have increased markedly in the months following the vote, partly as a result of President Donald Trump's crude efforts to convince North Atlantic Treaty Organization (NATO) allies to contribute more to the alliance, which raised genuine fears of declining American interest in Europe. Such seeming equivocation on behalf of the US contributed to renewed concern regarding the credibility of NATO -which Emmanuel Macron described as 'braindead' -leading to a corresponding interest in European mechanisms for insuring the continent against external threats (Dennison and Oliver, 2018). The economic consequences of Brexit also augur for greater collaboration in EU security and defence, not least since burden-sharing has been a major underlying driver of intra-EU collaboration in this policy domain (Martill and Sus, 2018: 860). The consensus among economists was that Brexit would result in a decline in UK GDP, although the scale of that decline varies depending on the nature of the post-Brexit relationship in the long term (Hix, 2018); and, politically speaking, the UK's desire to make up for lost institutional ties offers a further rationale post-Brexit for seeking to codify collaborative arrangements in security and defence, especially if these come with a lower sovereignty cost than membership of the single market and customs union (European Parliament, 2018: 5).
Several other factors reinforced the belief that an agreement on security and defence collaboration would be easy to achieve. To begin with, security and defence was not a highly politicised domain, at least compared with other aspects of the UK-EU relationship. The referendum campaign focused on questions of sovereignty, immigration, identity, trade and regulatory 'red tape', rather than on questions of security and defence (Hill, 2018). Although discussion of Britain's role in the world was commonplace, specific mentions of the EU's role in foreign, security or defence policy were difficult to find (Martill and Rogstad, 2019). Moreover, the intergovernmental nature of the EU's Common Security and Defence Policy (CSDP), a throwback to the pillar system of the Maastricht Treaty which kept the policy area ring-fenced from the EU's supranational legislative procedures, also held out the prospect of easier agreement: Not only would the sovereignty cost to the UK associated with continued participation be substantially lower in this domain, but the risk to the EU of keeping the UK involved would also be commensurately lower, since questions of judicial oversight or market-preserving values would not be at issue. Finally, the terms of collaboration were in no way a blank slate, since the UK and EU had been cooperating for many years, during which time they had developed a roster of shared practices and informal contacts at all levels (Michalski and Danielson, 2019;Svendsen and Adler-Nissen, 2019) and routinising intra-EU consultation and coordination (Whitman, 2016: 527).

The post-Brexit security puzzle
Given the incentives for continued security collaboration post-Brexit, it is not surprising that the UK government was initially optimistic about the prospects of reaching an agreement in this area. Prime Minister Theresa May's rhetoric on security and defence issues was consistently more conciliatory than the fierier tone used to refer to the negotiations more broadly, with talk of the need for a 'close partnership' a recurring theme. Indeed, May's depiction of the value of the CSDP was more positive than the UK's position had been prior to the referendum, a turnaround the EU ascribed to concerns about American disengagement and missing the boat vis-à-vis intra-EU security developments (European Parliament, 2018). The UK signalled early on its willingness to maintain a number of its CSDP commitments, including maintaining Northwood as the operational headquarters (OHQ) for Operation Atalanta -the EU's anti-piracy mission off the Horn of Africa -and continuing to participate in the Galileo satellite-navigation project.
As the months went by, collaborative rhetoric was matched with concrete proposals. The February 2017 White Paper, The United Kingdom's exit from and new partnership with the European Union, stressed the need for continued security and defence collaboration with the EU, stating that the UK 'will remain committed to European security and add value to EU foreign and security policy' after withdrawal (HM Government, 2017a: 63). Later that year, the UK released a position paper setting out its desire to develop 'a deep and special partnership with the EU that goes beyond existing third country arrangements', the beyond element of which envisioned regular high-level coordination of foreign policy, cooperation on sanctions, UK participation in existing (i.e. EU) foreign policy mechanisms, contribution to CSDP missions, and participation in European Defence Fund projects and Galileo (HM Government, 2017b: 18-20). In May 2018, the government published its Framework for the UK-EU Security Partnership which again proposed 'deep and special' arrangements based on coordination, consultation and collaboration taking place across different fields of foreign and security policy and at various different levels of seniority, coupled with a new treaty guaranteeing collaboration on matters of internal security (HM Government, 2018).
And yet the response to May's proposals from Brussels was not as cordial as had been expected. What the UK regarded as something of a concession to Brussels, namely, the articulation of proposals for a comprehensive partnership in security and defence (HM Government, 2017a, 2017b), the EU rejected on the basis of concerns regarding institutional autonomy and concern about the British 'cherry-picking' approach to integration. Brussels initially declined the UK's offer to maintain existing CSDP commitments, including the hosting of the OHQ for Atalanta and the command of Operation Althea by a British general (as Deputy Supreme Allied Commander in Europe (DSACEUR)); and precluded also the UK's terms for a partnership going far beyond existing precedents, ruling out 'associate' membership of decision-making forums, rejecting proposals for a UK contribution to CSDP mandates, and precluding British participation in Galileo (European Parliament, 2018: 15). Thus the agreement contained in the ensuing Political Declaration stayed close to the proposals of May 2018, yet without provisions that would go beyond existing third country arrangements (Whitman, 2019: 394). But the resulting trials and tribulations of May's deal, which failed to pass the House of Commons on three occasions in early 2019, led to May's downfall and the renegotiation of the deal, and of the Political Declaration, by her hardliner successor, Boris Johnson.
Johnson's revised document still envisaged case-by-case collaboration in CSDP missions, continued dialogue, industrial collaboration, and some aspects beyond existing models (Institute for Government, 2020). But the draft agreement on the future relationship, published in May 2020, showed that the UK had decided to wholly deprioritise the foreign, security and defence policy aspects of the agreement, which were absent from the text, owing partly to the government's prioritisation of economic considerations and partly to its belief that an agreement in security and defence was not a necessity, given the UK's membership of NATO and its bilateral security ties with EU's member states (Bond, 2020;Whitman, 2020). Unsurprisingly, the resulting EU-UK TCA, finalised on 24 December 2020 and provisionally entering into force on 1 January 2021, contained no provisions for foreign and security policy collaboration outside of the specific provisions for cooperation on cyber security and information sharing. And, as if the absence of an agreement were an insufficient indicator of the priority the UK attached to the EU as an international actor, an immediate spat erupted over the UK's decision not to bestow full diplomatic status on the EU's Ambassador to London (Cardwell, 2021).
In spite of prevailing expectations that negotiating a security partnership would be the least contentious aspect of Brexit, then, the experience to date has been anything but (Sweeney and Winn, 2020: 232). The failure of the UK and the EU to reach an agreement on security and defence is therefore puzzling. Much of the early literature on Brexit and security operated on the explicit assumption that an agreement in security and defence made strategic sense for both sides, and that this would make it more likely that an agreement would be reached (Biscop, 2016;Sus and Martill, 2019). Heisbourg, for instance, has noted that whilst in 2016 'it was reasonable to expect that governments sharing the same basic defence and security interests would behave rationally and work together to limit the negative consequences of Brexit', two years later and, 'in the wake of postreferendum fecklessness on both sides, it would in fact be reckless to assume that good sense will prevail in the coming months or years ' (2018: 24). And What is surprising to European security observers is also surprising from the perspective of existing theoretical explanations of international cooperation, given the clear presence of key factors thought to underpin successful cooperation, including shared strategic threats (Wivel, 2008), opportunities for mutual gains (Snidal, 1991), and the development over time of shared practices and a common security community (Adler, 2008).

The politics of withdrawal
In the remainder of this article, we aim to show how various aspects of the politics of institutional withdrawal have undermined the likelihood of an agreement in security and defence, in spite of the a priori strategic interest in cooperation. While we acknowledge the considerable interdependence between the concepts of politics and strategy, we distinguish them practically by defining politics in terms of the domestic and institutional interests faced by decision-making actors in London and Brussels, in contrast to the strategic imperatives established by the degree of external threat and the shifting balance of international power. Our claim is that, while strategic imperatives should -and initially did -push both sides toward an agreement, the political fallout from the referendum vote has gradually pulled London and Brussels further apart. By theorising the underlying processes responsible, we aim to explain not only why a potentially simple agreement became progressively more difficult over time, but also how the dynamics of withdrawing from pre-existing institutional arrangements produce distinct outcomes from scenarios in which states agree on the terms of collaboration from a blank slate.
The role of 'exit' in world politics has received considerable attention in the context of institutional reform ever since regime theorists sought to adapt Hirschman's triptych of exit, voice and loyalty to explain how states seek to influence international organizations (Keohane, 2020: 12;Lavelle, 2007: 373-375). The key insight from this literature, that the threat of exit can increase a state's influence, is also prominent in the two-level-games literature, which has drawn attention to the significance of credibility, alternative options, and the impact of the domestic environment in relation to exit (Putnam, 1988). More recently, a specific literature on the politics of exit in relation to the EU has emerged, focusing in particular on the prospect of European disintegration (Jones, 2018;Vollaard, 2014). Yet this diverse body of scholarship cannot fully capture the political dynamics of Brexit. The literature on exit and voice, for instance, has focused more on the threat of exit than on the consequences of realizing it, and has in any case focused on institutional environments that are less dense than the modern-day EU (cf. McEwen and Murphy, 2021). Meanwhile, the more specific literature on European disintegration and institutional obsolescence fails to capture the dynamics inherent in the departure of a single member state and the specific institutional changes which this may unleash.
The difficulty of accounting for these dynamics has led scholars working on Brexit specifically to consider a variety of new perspectives, including research on differentiated disintegration (Schimmelfennig, 2018), the politics of divorce (Winn, 2018), and processes of de-institutionalisation (Wright, 2020) and de-Europeanisation (Rosamond, 2019). Our article aims to contribute to this task by unpacking those aspects of withdrawal which undermined the prospects of an agreement on security and defence. Specifically, we identify four factors in the politics of withdrawal, each of which contains multiple identifiable dynamics: (a) First, we show that problems of institutional complexity can arise, creating concerns about institutional autonomy, problems associated with selective participation, and the risk of incentivising further demand for exit; (b) Second, we highlight dynamics produced by the negotiations, including the hardening of preferences, politicisation of issues, and a tendency toward centralised and in-group decision-making; (c) Third, we show that withdrawal itself spurs changes which need to be factored into the equation, notably by establishing incentives to demonstrate independence and by facilitating new institutional developments; (d) Fourth, and finally, we demonstrate that withdrawal can re-establish distributive dynamics over the spoils of departure, the role(s) of the leaving state, and the choice of institutional framework(s) for managing future relations.

The politics of institutional complexity
The UK's departure from the EU raises a host of questions regarding institutional complexity and the problems of non-membership, since existing security and defence arrangements are tied to the Union's institutional structure, even as they are distinct from other policy areas. This raises a host of problems for both sides.
A first issue concerns institutional autonomy, a claim both sides have insisted on, but which are incompatible together (cf. James and Quaglia, 2021). The EU is keen that the Union's decisionmaking autonomy -its ability to govern itself -is not undermined by the inclusion of non-members in decision-making processes. Third countries, even those like Norway which opt-in to a number of EU policies, are thus excluded tout court from EU decision-making, although in practice both formal channels for consultation and informal forms of influence are available to nonmembers (Svendsen, 2020: 529-531). Even if the nature and aims of EU security and defence policies differ from other EU policy areas, they must still respect this boundary condition. Proposals which would see the UK participating, even if informally, in decision-making were rejected on the basis that to allow a non-member such privileges would therefore undermine the Union's decisionmaking autonomy (European Parliament, 2018: 10-12). And yet, perhaps understandably, the UK rejected the notion that, as a major security and defence actor, it could be bound up in arrangements over which it would have no say. The UK would not accept being placed at the bottom of the decision-making hierarchy nor committing troops and military equipment to a CSDP mission it had no role in shaping, owing to concerns about its own autonomy.
A second issue concerns the problem of institutional selectivity, and what has come to be understood as 'cherry-picking'. Allowing the UK to opt-in to policy areas it regards as beneficial is seen as undermining core aspects of the logic of integration, which would ultimately collapse if members were free to opt out of any policy areas they regard as unbeneficial. Such a move would also undermine the solidarity of the Union which, although differentiated in some respects, is built on the principle of a common core shared by all members, with the foreign and security policy apparatus viewed not simply as an optional area of integration, but as a privilege of membership and as a means of expressing the common external goals of the members of the institutional core. Allowing the UK to participate as a non-member, from this perspective, would be to confer benefits of membership which it does not deserve, to establish disparities with existing non-members (such as Norway or Turkey), to alter the underlying logic of the connection between the CSDP and the Union, to foster the proliferation of new and complex forms of association, and to introduce a potential spoiler into the EU camp whose commitment may be questioned (European Parliament, 2018: 14-15).
Third, the act of UK withdrawal established a risk of contagion for the EU. Specifically, Brexit threatened the credibility of the European project and raised the prospect of existing members also seeking to depart, especially had the UK been seen to obtain favourable terms. This reality had a number of implications for security and defence collaboration. To begin with, and as we discuss in detail below, it led the Union to prioritise reforms in security and defence policy which de facto diminished London's ability to participate (Bérard-Sudreau and Pannier, 2020: 5-6). Moreover, forestalling contagion all but required Brussels to preclude offering the UK a generous deal, including access to elements of EU policymaking on an a la carte basis, or without the trappings of membership. The EU's response to the UK's initial proposals made it clear that these policy domains are reserved for members, and that the UK cannot simply pick and choose (European Parliament, 2018: 9). To this extent, the UK has become, in Beaumont's (2019) words, an 'unwitting martyr for the cause', since the troubles of Brexit ultimately play into EU narratives regarding the value of membership. By precluding a scenario in which the UK gets a bespoke deal, Brussels prevented the establishment of a broader 'British model' of association which other member states might have found attractive.

The politics of negotiation
Britain's exit has also required both sides to engage in a complex process of negotiations over the terms of withdrawal and the nature of the future relationship, since the UK's near five-decade-long membership of the Union established myriad institutional linkages which needed to be unpicked; and the dynamics of the negotiations, in turn, affected the prospects of a security and defence agreement.
First, the negotiations contributed to a hardening of preferences on both sides. Inflating the UK's asks was a key component of Theresa May's bargaining style, while the EU offered in return only existing models of association, a position to which it stuck steadfastly through much of the talks (Martill and Staiger, 2020). British hard bargaining also contributed to a decrease in the residual trust between both sides, and to a souring of political relations, lowering the willingness to compromise across the board and, crucially, decreasing the EU's propensity to take the UK at its word (European Parliament, 2018: 14). The negotiations on security and defence also became hostage to other aspects of the talks in which agreement was always going to be more difficult to reach, with early UK efforts to leverage the security commitment undermining the UK's sincerity in such matters (Hill, 2018), and with Brussels coming to regard the risk of cherry-picking in security and defence as equivalent to that associated with such practices in the economic domain (European Parliament, 2018: 14).
Second, the course of the negotiations has been accompanied by political changes at the domestic level in the UK and within and between the EU's member states which made the prospect of agreement less likely. In the UK, the talks presented political opportunists with an obvious means of out-foxing May on the right, allowing hard Brexit supporters to push the government closer to their minimalist view of the post-Brexit relationship. This was reinforced by May's own Eurosceptic rhetoric and her emphasis on the possibility of 'no deal', as well as by the revolutionary moment created by Brexit. As a result, areas of the negotiations once assumed to be under the political radar ceased to operate as such, and the salience of the security and defence relationship increased significantly, with pro-Brexit detractors complaining the UK was being 'subjugated' in this domain (Martill and Sus, 2018). The Conservatives' move to the right augured against a more comprehensive partnership with the EU, and saw the UK -under May, and then Johnson -gradually embracing harder designs on Brexit (Martill and Staiger, 2020). While the EU is a more complex actor, it is notable that these dynamics are not confined to the UK, and analysts have identified similar trends within major member states, with a gradual disillusionment with the UK and calls for harsher treatment by Brussels, alongside growing support for a more autonomous role in EU security and defence (Eurobarometer, 2019).
Third, the conduct of the negotiations was characterised by centralising dynamics on both sides which to some extent served to decrease the flexibility involved. In the UK, the government eschewed cross-party or cross-regional membership of the negotiating team and May worked hard to centralise the process within the Cabinet Office. This arguably made it more difficult for actors favouring a soft Brexit to get their voices heard, and it made the UK government dependent on a number of pro-Brexit veto players. The EU, in contrast, worked hard to foster unity among the EU27 in order to head off British attempts to 'divide-and-rule' and established a clear line of delegation from the European Council to the Commission's Taskforce 50 as a means of streamlining the talks and keeping the member states together (Laffan, 2019). The upshot of the EU's conduct in this regard -and it is not clear why they would have done otherwise -has been the prioritisation of the collective EU interest over individual member state interests, not least over and above those who would have wished to see the UK more involved in EU security and defence. Indeed, it has been suggested informally that one of the key reasons the UK sought to eke out a comprehensive agreement on security was because individual member states had shown considerable enthusiasm for just such a prospect. But this was simply not compatible with the levels of unity and centralisation required of the EU's conduct of the broader negotiation process. While member states also had an incentive in the EU holding the line, this does not mean that many would not have preferred London to be more involved than Brussels would have, and the trade-off between these interests was the product of a careful and deliberate effort to maintain unity in spite of divergent positions.

The politics of institutional change
Brexit has also induced important changes in the priorities of both the UK and the EU in ways that have implications for the viability of any future security and defence relationship. In other words, it does not make sense to view Brexit solely as an institutional rupture between two actors at a single moment in time, since the decision unleashes subsequent changes which will likely affect the prospects of future cooperation.
First, Brexit has pushed both sides to find ways of demonstrating meaningful divergence in their foreign and security policies, even as the underlying incentives push for closer collaboration. On the UK side this has taken the form of the Global Britain discourse, a vision of the UK as a renewed power on the international stage with a global reach obtained by breaking free from the 'shackles' of European integration (Hill, 2018). The slogan was first articulated by May in her Lancaster House speech, but has continued to gain purchase among Brexit supporters, and has been modified by Johnson into a focus on 'unleashing Britain's potential', though the substance remains largely the same. The EU, for its part, has stepped up discussion of 'strategic autonomy' in the wake of Brexit, the aim being to reinvigorate the Union's drive as an actor and portray its ability to eschew reliance on third countries, including the US and the (post-Brexit) UK. It is instructive that both sides have articulated novel foreign and security policy discourses following the Brexit vote, a product of the close connection between national security questions and the much valued -yet conceptually empty -signifier of 'independence'. But these discourses also distort, to some extent, the underlying connection between globality/autonomy and the question of UK-EU security and defence collaboration. While they are intended to highlight the value of independence for both sides, analyses of the UK's globality generally show this to have been a function of the multiplier effect of EU membership (Hill, 2018), while similar assessments of strategic autonomy posit that close collaboration with the UK would be a necessary facilitator of a more independent European posture (Biscop, 2016;Howorth, 2019;Martill and Sus, 2019).
Second, Brexit has facilitated a number of key reforms on the EU side which have emerged gradually as the negotiations have proceeded. While the initial motivation behind these initiatives came from the prior recognition of Europe's weakness in an increasingly insecure world (European Commission, 2013: 3), Brexit helped bring about a reform moment within the EU, forcing the Union to address the capability gap brought about by Brexit and to show that further integration was still viable (Bérard-Sudreau and Pannier, 2020: 6), whilst simultaneously removing the UK as a veto player in this domain (Euractive, 2011). The significant initiatives which have emerged from the EU's momentum in security and defence include the creation of the EU military headquarters, the Military Planning and Conduct Capability (MPCC), the launch of Permanent Structured Cooperation (PESCO) which facilitates binding cooperation on defence projects between specific groups of interested member states (Sweeney and Winn, 2020: 226), the establishment of a European Defence Fund (EDF) through which community funds will be allocated for technological innovation, defence research and technology (Koenig and Walter-Franke, 2017), and the instigation of a Coordinated Annual Review on Defence (CARD) to identify possibilities for the further pooling of resources. More recently, the EU has articulated the 'Strategic Compass' concept, initiated by Germany, to guide the developments in this area and move toward a common strategic culture between member states (European External Action Service, 2021). These institutional developments on the EU side make it more difficult to envisage a close UK-EU security partnership: on the supply side they raise the costs for the EU of introducing an Atlanticist 'spoiler' into their new initiatives, and make the EU less interested in harnessing UK capabilities; on the demand side, they make it so that London's participation would involve more onerous commitments, while also making the EU something of a moving target in security and defence.

The politics of distribution
Finally, Brexit establishes conditions for increased competition between actors over some of the distributive elements of European security and defence provision. Key areas of distributive wrangling have involved competition over the spoils of withdrawal and the UK's former roles, as well as institutional wrangling over the most appropriate format for maintaining collaborative ties with London.
First, although there was desire for a security and defence agreement on both sides, this was always accompanied by clamouring between a number of the remaining member states for the spoils of Brexit. Most notably up for grabs were the specific UK commitments within the CSDP, including the hosting of the OHQ for Operation Atalanta, the command, under a British general, of Operation Althea, and the UK commitment to lead an EU Battle Group in 2019. Each of these responsibilities the UK sought to maintain as a means of maintaining closer ties and demonstrating its commitment to CSDP, and each of these were subsequently removed from the table, owing to a combination of the UK's imminent departure and the desire of other member states to take charge of these operations. Spain and France, for instance, undertook to jointly host the OHQ for Operation Atalanta, while the Commission stated that the UK could not host operational missions, nor host a battle group in 2019 (European Parliament, 2018: 15), although it was London which formally withdrew from this arrangement in 2019, citing a lack of capacity. For the EU it made sense to deprive the UK of the benefits of membership and allow other member states to take on an increased role in CSDP, but such rebuffs were difficult to understand in London given that such offers had been couched as concessions to the EU.
Second, Brexit establishes new forms of competition between European security actors, since the roles played by the UK in the European security milieu are now, to a certain extent, open to external contestation. As an important EU member, the UK played a number of specific roles within the foreign and security policy domain, including those of European security 'leader' and of 'transatlantic bridge' between the US and the EU (Ewers-Peters, 2020). While the UK will continue to be a major player in European security whatever happens, its claim to leadership in this role has been dented by the Brexit process and by London's having removed itself from the EU's institutional framework. Moreover, the UK's role as a transatlantic bridge has been diminished by Brexit, creating opportunities for other actors to fill this role: While Germany, as the 'reluctant hegemon', has begrudgingly committed to becoming more active in security and defence, France has embraced its new position as the major security and defence actor in the EU, and as a potential interlocutor between Washington, London and Brussels. Indeed, the most notable innovation in French security policy in recent years, Emmanuel Macron's European Intervention Initiative (EI2), aims to offer a flexible and effective non-EU format for rapid intervention (Pannier, 2017), but is also motivated by a desire to keep the UK aligned to EU partners and to demonstrate -and enhance -French leadership in European security (Martill and Sus, 2018: 858).
Third, Brexit has re-ignited institutional competition within the field of European security by posing the question of which framework will ultimately be utilised to keep the UK on board. While both the UK and the EU, as we noted above, saw value in a bilateral partnership, there are a multitude of other institutional means through which the UK may pursue closer collaboration with European partners, including EU member states. Options include the strengthening of bilateral and minilateral arrangements along similar lines to the Franco-British Lancaster House arrangements, the UK-initiated Joint Expeditionary Taskforce in the Nordic region (Howorth, 2019) or existing E3 arrangements between the UK, France and Germany (Billon-Galland et al., 2020); enhanced forms of EU-NATO cooperation (Ewers-Peters, 2020; Howorth, 2019) and a revisiting of the European 'pillar' idea within NATO; or engagement through new non-EU formats, including Macron's EI2 initiative (Pannier, 2017). The plethora of available options makes the task of securing agreement between the UK and the EU much more complex than a simple coordination game, since both sides must not only consider which format is best suited, but also how these different initiatives will fit together. It also offers more options to the UK, allowing London to engage in 'forum shopping' to a limited extent, although not all options are as viable as each other, and many ultimately work to complement EU structures. Nonetheless, the presence of alternative formats is thought to be a major contributing factor in the decision of the Johnson government to shelve the idea of a comprehensive UK-EU security partnership (Whitman, 2020).

Conclusion
In spite of the shared strategic interests between the UK and the EU, it has proved far more difficult than expected to agree upon the terms of continued security and defence collaboration. Why agreement proved so difficult is puzzling for analysts of European security, who expected agreement to be reached relatively easily in this area, but also for existing theories of international cooperation, which identify shared threats, mutual interests and prior engagement, all of which pertain in the EU-UK relationship, as key drivers of cooperation.
We argued in this article that the failure of both sides to reach agreement was the product of political constraints resulting from the withdrawal process itself. First, we highlighted problems of institutional complexity, including the incompatibility of mutual demands for institutional autonomy, problems of institutional selectivity, and the consequences of efforts to head off contagion. Second, we showed how the negotiation process made agreement less likely by pushing both sides toward a harder position, increasing the salience of security and defence issues, and contributing to the centralisation of decision-making on both sides. Third, we examined how processes of change affected the chance of agreement, with the UK and EU setting out highly political (and stylised) narratives of autonomy, and with the EU undertaking security and defence reforms that made UK participation potentially more costly and more complex. Finally, we discussed the re-emergence of distributive issues, noting how competition over the spoils of the UK's exit, London's former leadership and brokerage roles, and the institutional framework for keeping the UK plugged-in contributed to making an agreement less likely.
By highlighting the distinct dynamics of exit, we are able to explain why an agreement which may have appeared relatively simple at the beginning of the Brexit process came to be mired in controversy and eventually abandoned as British withdrawal got underway. The argument also contributes to our understanding of the politics of exit more generally by helping us to identify the distinct mechanisms associated with the politics of withdrawal, and showing how these can work against more traditional determinants of international cooperation. As the Brexit example shows, the very process of exiting can unleash path-dependent processes which can serve to preclude vestigial agreements, even where both sides initially support a deal and where the conditions for continued cooperation are propitious. As challenges to the liberal international order mount, and as populist leaders take aim at the architecture of global and regional governance, it is vital that we understand the dynamics unleashed by institutional exit and the various ways in which these mitigate against continued collaboration even after withdrawal has taken place.
Where does this leave the European security architecture and the future UK-EU relationship? Since both sides share many of the same interests, forms of informal collaboration will likely develop over the months ahead, not least during times of crisis or significant perceived need. In other words, the failure to reach an agreement does not mean diplomatic relations will collapse, nor that shared interests will not continue to motivate cooperation. But the absence of an agreement may well make things more difficult and will prevent both sides from realising the kind of European strategic autonomy which may be required in the long run to head-off emergent threats and cope with the long-term restructuring of the American commitment to Europe. Whether or not the kind of partnership initially envisaged will emerge in the years ahead depends largely on whether the conditions brought about by the Brexit process itself can be reversed or mitigated as the acrimony of withdrawal recedes into the background.