The changing character of EUFOR Althea: power politics or learning?

Abstract States have different strategic cultures when it comes to legitimating the use of military force and its relation with other foreign policy instruments. However, increasingly, military operations are conducted in multilateral forums; EU military operations are one of the most notable examples of this development. While some claim that these operations reflect power relations between nations with different strategic cultures, others argue that these common missions involve states in a process of collective learning and convergence of interests. Drawing upon an advocacy coalition approach, this paper confronts the competing hypotheses in the case of European Union Force (EUFOR) Althea in Bosnia-Herzegovina (BiH), as the EU’s longest running military operation (since 2004). On the basis of policy documents and semi-structured interviews with policymakers and politicians, this paper concludes that the evolution of EUFOR Althea has been primarily the result of the power politics of different coalitions, but there have also been a few instances of learning.


Introduction
How and why have EU military operations changed? Does ten years of EU military operations reflect a process of collective learning? Or has it primarily been driven by the power politics of particular coalitions? This debate is of great importance as the EU's collective use of military force is closely related to the debate on the identity of the EU as international security actor. It also helps in assessing the potential for the further development of the EU as an international security actor.
Since 2003, the EU has launched eleven military operations as part of its Common Security and Defence Policy (CSDP). The position of the EU's military operations within the EU's foreign policy has been subject of debate ever since the first military operation was launched, Operation Concordia, in FYR Macedonia. When the EU still was an actor without military power, it received many labels, like 'civilian power' and 'normative power' (Duchêne 1972;Manners 2002). Also, in the policy discourse of the EU the absence of military means has been claimed as a strength of the EU's distinctive actorness (see Kurowska 2008). Hence, the EU's military operations have raised questions about the EU's identity as an international actor.
The EU's military operations reflect underlying struggles about the development of an EU strategic culture-that is, how and when to use military force. Under the guise of apparent unanimity, the actors involved in the decision-making are divided on what CSDP should be and do in practice (Barrette 2014).
This article focuses on the case of European Union Force (EUFOR) Althea in Bosnia and Herzegovina (BiH). Operational since 2004, it is the EU's longest running military operation, covering almost the whole period of CSDP. While the EU's involvement with BiH has received considerable attention from different (theoretical) angles, and the change of the EU's involvement in BiH over time is recognized (Juncos 2013;Noutcheva 2009;Bassuener and Ferhatovic 2008;Sebastian 2009;Sebastian-Aparicio 2014), it has not been the subject of sustained longitudinal analysis.
The question, thus, is whether we find in the evolution of EUFOR Althea evidence that a process of collective learning has taken place or whether any shifts in the mission are best explained in terms of power politics. To this end I propose an advocacy coalition approach as the overarching analytical framework. This approach is well suited to analyse struggles at the discursive level on the position of the military instrument, and allows us to critically evaluate the competing expectations. The paper concludes that the development of the character of EUFOR Althea is best understood by power politics, while there have been some instances of collective learning as well.

Theorizing the evolution of the CSDP
This paper aims to identify the mechanisms that drive the activation and organization of military operations by the EU. While the actual analysis focuses on the evolution of EUFOR Althea as the longest standing EU military operation, the assumption is that the mechanisms that are identified there are reflective of the wider development of CSDP over its first decade of existence. Hence, the theories adopted-institutional learning and power politics-are taken to operate on both levels: individual operations and the CSDP as a whole.

Learning or power politics?
Existing research approaches the development of CSDP operations from mainly two angles. First, studies that focus on processes of institutional learning suggest an increasingly shared/common understanding of the role of the military instrument as part of the EU's overall foreign policy (Faleg 2012;Smith 2012). Central to the institutional learning literature concerning CSDP is that the accumulation of operational experiences leads to a 'deliberate, pro-active, transparent, collective/social, policy-relevant and progressive' process of reform (Smith 2015, 118). Moreover, the role of epistemic communities and professional expertise is emphasized (Smith 2012;Faleg 2012). The underlying conflicting ideas on CSDP operations (that is, the relationship between the military instrument and other EU foreign policy instruments, and its justification) and the EU's development as an international security actor are depoliticized.

The changing character of EUFOR Althea 69
Second, research on national strategic cultures emphasizes the continuing relevance of national views (Haine 2011;Schmidt 2011). Haine (2011, 587) argues that no overarching EU security culture is emerging, but that the 'prism of national experiences is dominant', highlighting the differences among the Big Three. Indeed, Haine (2011, 586) claims that socialization and learning processes may actually have become weaker since the creation of CSDP in 2003. Similarly, Biehl et al (2013, 396) conclude in their edited volume in which country experts assess the strategic culture of all EU member states that 'persistent difference is just as, if not more, frequent'. Hence, they expect CSDP operations to emerge from ad hoc coalitions, rather than being the result of an institutionalized EU consensus.
A more nuanced position is taken by Meyer (2006Meyer ( , 2011Meyer ( , 2013 who observes a considerable ideational convergence among the EU member states, while also noting a lack of fundamental convergence. To account for this contradictory evidence, he refers to differences in material resources, institutions, and legal-constitutional factors that 'dilute the influence of new discourse coalitions advocating changes in norms and ideas about the use of force as well as limiting the extent to which they can be put into practice' (Meyer 2011, 678).

Advocacy coalitions
Building upon Meyer's notion of different coalitions, this paper proposes to adopt an advocacy coalition framework. This encompassing analytical framework allows us to arbitrate between the different hypotheses as it takes into account both continuing conflict and power dynamics as well as processes of learning and institutionalization.
The advocacy-coalition framework (ACF) was originally developed by Paul Sabatier (1988) to assess processes of policy change. Note that ACF, which makes us sensitive to coalitions, is used here as a heuristic device to test competing arguments on policy change in the field of CSDP-that is, it is not about testing specific ACF hypotheses. Different advocacy coalitions are situated in a context where resources and (institutional) constraints affect the potential for success of the different coalitions. Resources vary by coalition and relate to both material capabilities, institutional and discursive resources. Moreover, the extent to which coalitions actually coordinate may also be a factor that influences the degree of effectiveness of coalitions.
While resources are located at the level of the coalitions, (institutional) constraints are the same for the whole policy field. Nevertheless, their impact on particular coalitions may differ. For example, while the economic crisis affects all coalitions, this can be more or less constraining depending on a coalition's ambition for CSDP.
In contrast to more rational institutionalist approaches (Klein 2010;Dijkstra 2013) which focus on formal powers and assume that actors pursue material self-interests, an advocacy coalition approach is open to the inclusion of a variety of (non-state) actors and does not take the preferences of actors as given (Sabatier and Weible 2007). In line with this more open approach, this study does not limit the analysis to the "big three" (Germany, France and the UK), which is quite common in studies on CSDP. Rather it starts from the different belief systems that are out there concerning the use of military force in the EU context. These different "coalitions" involve a range of national governments, but also include EU institutions (European External Action Service, Commission, Parliament). Since this paper focuses on EU-level decision-making, it does not differentiate between different branches of governments within national governments. However, I do take these functional differences into account by distinguishing between different branches in EU institutions. 1 Since non-governmental actors (for example, journalists, researchers and policy analysts) are not key to advocacy coalitions in CSDP (cf. Barrette 2014), this analysis only includes governmental actors.
The advocacy approach yields two distinct mechanisms of policy change, which correspond to the positions taken in the aforementioned debate on the development of CSDP: either the dominant coalition is replaced (for example, power dynamics), or there is a win-win situation in which all major coalitions agree that a policy change is necessary (for example, learning) (Sabatier and Weible 2007;Sabatier 1998 In line with Levy's (1994) well-known account of learning in foreign policy, I emphasize that learning is about a change of beliefs as a result of experience. Moreover, since there are different understandings of learning (see Bennett and Howlett 1992), two issues need explanation: who is learning and what is being learned. Concerning the subject, we can distinguish between individual actors that learn, and a learning process within coalitions and across coalitions. Institutional learning focuses on the collective learning processes. Regarding the object of learning, there is a limited degree of learning when it is restricted to procedures and organizational changes. In contrast, ideational learning is most fundamental, resulting in changing objectives and goals (a paradigm shift). The level of learning between these two extremes is instrumental learning (see Hall 1993, 278). Hence, an ideational learning process across coalitions is the highest degree of learning, while an individual actor learning about procedural and organizational issues is a very limited degree of learning. As the hypothesis formulated above shows, this case study understands institutional learning as a collective process and as involving more than a change in procedural and organizational issues-that is, it is a change in the character of EU military operations.
Actor constellations hypothesis: The changing character of EU military operations results from a change in the balance of power between different (advocacy) coalitions, following shifts in material resources and institutional factors.
For both historical and analytical reasons, the actor constellations hypothesis claims primacy over the institutional learning hypothesis in the sense that we would assume the CSDP to be driven initially by the power constellation among the member states. In that sense, the burden of proof is on institutional learning to take place as the CSDP (and the operation) evolves.
Furthermore, it may be clear that the two hypotheses do not logically exclude each other; that is, the power politics among different coalitions may co-exist and even impact on institutional learning (and vice versa). Yet analytically they are distinctive. In particular, institutional learning would entail a change in ideas to bring about change, while changing actor constellations may bring about change irrespective of a change of ideas of actors. As such, the two mechanisms relate to different empirical observations (see section 2) and also have fundamentally different implications for the development of CSDP at large. Clear evidence of institutional learning signals a more systematic attempt at shaping CSDP, and a strengthening of a particular CSDP identity.

Research design
I develop an analytical, theory-driven case study to assess whether EUFOR Althea has changed as the result of a process of learning or whether it was primarily a process of changing coalitions (Blatter and Haverland 2012, 144-204).

Case selection
The development of the EU as an international security actor cannot be separated from the Yugoslav crises of the 1990s.
Hence, the EU's involvement with BiH is of particular interest to assess the position and role of the military instrument in the EU's foreign policy. In 2004 the EU launched EUFOR Althea, and this operation is still ongoing. Hence, the case of the EU's involvement with Bosnia in general and the deployment of EUFOR Althea in particular runs parallel to, and is part of, the EU's development as an international security actor. That is not to say, of course, that all changes in Althea necessarily reflect broader developments. Still, as the EU's longest running operation, we expect this case to yield evidence of the mechanisms that have driven the evolution of the EU's military operation at large.

Althea in context
From 1992 to 1995, the Bosnian War took the lives of a hundred thousand people (RDC 2007). While the European Economic Community, together with the United Nations, was involved in several diplomatic attempts to find a constitutional solution to the conflict (Cutileiro plan, Vance-Owen Plan, EU action plan), North Atlantic Treaty Organization (NATO) airstrikes eventually led to the signing of the Dayton Agreement in 1995 (Sebastian Aparicio 2014). NATO also deployed troops (Implementation Force and Stabilization Force) to ensure the implementation of this peace agreement. The role of the European Union was limited to the provision of humanitarian assistance, and financial aid for reconstruction and development (Friesendorf and Penksa 2008). This changed in 2004 when the EU took over from NATO with Operation Althea.
This had not been obvious from the start, as the US in 2003 had blocked a proposal of the European Council (2002) to take over from NATO's SFOR, because of a fear of EU autonomy in the field of security and defence (EU Observer 2003; see also Dijkstra 2013;Pohl 2014).
Eventually, an agreement was reached on the EU's use of NATO assets and capabilities-the Berlin Plus arrangement-which was put into practice in 2003 with the launch of the EU's first military operation in Macedonia: Concordia. This provided not only the EU but the US as well with the confidence that the EU would be able to live up to the expectations in taking over the BiH military operation from NATO. NATO remained present with a military headquarters involved with defence reform and counter-terrorism.
Research on the relationship between NATO and the EU points at both instances of interorganizational learning (in terms of institutional set-up) and efforts to make the EU's operations distinctive from those of NATO (Varwick & Koops 2009). In this paper, I focus on the internal EU dynamics. In this regard, the fact that EUFOR Althea took over from a prior NATO operation using NATO assets (Berlin Plus arrangement) acts as an important "constraint" affecting different coalitions differently. On the one hand, it provides room for institutional learning as there is a common preceding experience. On the other hand, it may fuel power politics among different coalitions, as the relationship with NATO is one of the crucial issues of contention between them (see section 3.1).

Data collection
The research consists of a range of data sources, including interviews, policy documents, parliamentary debates, newspaper articles and secondary literature. Regarding the interviews, in May/June 2013 approximately twenty-five semi-structured expert interviews were conducted-primarily in Brussels. The interviewees cover different national diplomats, EU institutions and time frames. The policy documents consist primarily of the different Council Joint Actions and reports to the United Nations.

Structure of the analysis
The analysis is organized around the critical decisions that have affected the character of EUFOR Althea (explanandum). Logically, the analysis departs from the year that the decision (Council Joint Action) to launch EUFOR Althea was taken, 2004. The circumstances surrounding the launching decision help to identify the different coalitions at the time and the main issues of contention. Thus, it serves as a baseline for the rest of the analysis, which is organized around the main critical decisions that have changed the character of EUFOR Althea in terms of its objectives, the institutional relationships and the number of troops.
While the 2004 Council Joint Action has not been officially amended, there have been two moments at which the character of the mission has been significantly revised. Firstly, in 2007 the initial number of 7000 troops was decreased to 2500 and the position of the EU Special Representative was strengthened. Secondly, in 2010-2012, non-executive tasks were included in the operation and the number of troops was further decreased. For each of these critical decisions, I examine whether it was primarily the result of the changing power constellations between political actors or whether it was primarily driven by processes of learning.
Empirical observations that are in line with the learning hypothesis are: processes of lessons learned, procedures of evaluation and review, and references to The changing character of EUFOR Althea 73 past experiences. Moreover, a change in response to changes "on the ground" in BiH would count as evidence for learning as well. Empirical observations in line with the actor constellation hypothesis refer to changes in the resources of particular coalitions: changes in institutional relationships, actors changing from coalitions. Moreover, requests for (or lack of) support from external actors (US, UN) impact on the positions of different coalitions.

Mapping the advocacy coalitions
To map the variety in positions on CSDP, and in particular on EUFOR Althea, I organize the main actors (member states and EU institutions) in four advocacy coalitions, building upon the work of Vennesson (2010) and Meyer (2006). Following Sabatier's (1988;1998) understanding of advocacy coalitions, these coalitions are taken to share a common belief system-that is, a set of value priorities and causal assumptions about how to realize them. As these core beliefs are deeply engrained, the coalitions are expected to be relatively stable in terms of membership. A country's affiliation is not expected to change because of a simple change of government. However, in a way somewhat akin to the strategic culture approach, a country's value position may change/differ due to specific experiences it has in its foreign engagement.
The key issues that divide the different coalitions are the relationship of EU military engagement with NATO and the scope of the military mandate. These are essentially key issues that run throughout the development of the CSDP. In the context of EUFOR Althea, the particular issue of contention concerning the scope of the military mandate involved the question of whether or not to include the fight against organized crime in EUFOR's mandate.
For each of these coalitions I discuss its basic principles, key members and position on EUFOR Althea. Let me note that the fact that members of a coalition share a basic value position need not preclude disagreements among them as regards more operational decisions. Also, the differences between the coalitions are not absolute in character. For example, the human security coalition does not have a monopoly on a value-based justification of military force (further discussed below) but, compared to the other coalitions, it is a distinctive mark of this coalition.
3.1.1. Euro-Atlanticist coalitionThe heading text "311 Euro Atlanticist" has been changed to "311 Euro-Atlanticist coalition" Please check the change conveys the intended meaning or amend. The Euro-Atlanticist coalition consists of actors that share the conviction that the military instrument performs an important role in foreign policy, primarily in the context of NATO. EU member states that were at home with this coalition are: the United Kingdom, the Netherlands, Greece, Italy, Poland, Portugal, Hungary, Lithuania, Romania and Slovakia (Vennesson 2010;Biehl et al 2013). The Euro-Atlanticist group welcomed EUFOR Althea to underline the EU-NATO complementarity and, hence, emphasized the continuity with the preceding NATO operation. However, on the inclusion of the fight against organized crime in EUFOR's mandate, the UK and Italy took a different position from the Netherlands. Whereas the Netherlands did not see crime fighting as part of EUFOR's mandate, the UK and particularly Italy were in favour of it (Dutch Ministry of Defence 2004).

Global Power EU.
In the Global Power EU coalition the military instrument is of great importance in the foreign policy toolbox. Members of this coalition are united in their belief that military operations are an indispensable instrument, for which European citizens should be willing to make a financial and human effort (Barros-Garcia 2007). Moreover, this coalition is ready to use the military instrument for a whole range of purposes, including more geostrategic and national interests. The crucial distinction from the Euro-Atlanticist group is that the Global Power EU coalition strongly emphasizes the autonomous EU security identity, distinct from NATO. The most prominent members of this coalition were France and former High Representative Javier Solana (Meyer 2006). Spain and Belgium also fit with this group in light of their vocal support of the EU as a security actor, although their risk averseness makes them less ambitious (Arteaga 2013;Biscop 2013).
The Global Power EU coalition welcomed the initiation of EUFOR Althea to succeed NATO's SFOR, albeit for different reasons than the Euro-Atlanticist group. Central to the support of this group is the emphasis on the distinct character of EU security policy, which by implication would also apply to operation EUFOR Althea. While Solana has been one of the key actors promoting EUFOR Althea, France took a more reserved position.

Human Security.
In the Human Security coalition the use of military force is narrowly circumscribed; that is, it is only legitimated for the protection of civilians. Members of this coalition prefer the use of non-military instruments and aim to embed the EU's military operations within the EU's broader foreign policy framework. However, when the members of this coalition are convinced that lives of innocent civilians are at stake, and the deployment of military force is the lesser evil, they are willing to take risks (that is, deploy their troops) in defence of those values. This coalition prefers the EU over NATO and puts great emphasis on the UN as an indispensable source of legitimacy (Wunderlich 2013).
EU member states that fit this coalition are 'the neutrals' (Ireland, Finland, Sweden), with a leading role for Sweden, which conceives of itself as the 'world's conscience' and 'moral superpower' (Wunderlich 2013, 277). Additionally, the European Parliament is part of this coalition. Although lacking any formal powers, it has developed a "distinct" foreign policy identity, with an emphasis on human rights and democracy promotion (Zanon 2005). The Human Security coalition supported EUFOR Althea taking over from NATO's SFOR. In particular, the European Parliament (2004) turned out to be a vocal supporter of an operation in line with Solana's proposals. However, the EU member states in this coalition were hesitant regarding fighting organized crime as part of the mandate of the EU's military operation (Friesendorf and Penksa 2008, 688).
The changing character of EUFOR Althea 75

Bystanders.
The key characteristic of the least cohesive, Bystander coalition is the unwillingness to take substantial risks to defend European values by the use of force. The reluctance to do so has different reasons, ranging from a strong developed security of risk avoidance (Germany; Geis 2013), a strong focus on domestic defence (Austria, Bulgaria, Czech Republic, Cyprus, Estonia, Latvia, Malta and Slovenia; Biehl et al 2013), and a fear of being sidelined (European Commission). They are not a driving force behind CSDP operations, and their support has to be sought time and again.
The Bystander coalition was not fiercely opposed to EUFOR Althea. Germany positioned itself as a moderate supporter, while the Commission emerged as the most critical actor (Dijkstra 2013;Bundestag 2004).

The outcome: winners and losers
EUFOR Althea was given the mandate to provide deterrence, continued compliance with the responsibility to fulfill the … General Framework Agreement for Peace in BiH and to contribute to a safe and security environment in BiH, in line with its mandate, required to achieve core tasks in the Office of the High Representative's Mission Implementation Plan and the

SAP [Stabilisation and Association Process]. (Council of the European Union 2004a)
Obviously, the driving advocacy coalitions behind EUFOR Althea were the Euro-Atlanticists and Global Power EU. Specifically, the most notable advocates of EUFOR Althea were Solana and the UK, albeit for very different reasons. Whereas Solana, supported by the European Parliament, aimed at making EUFOR Althea a showcase of the EU's distinct security and defence policy, the UK insisted that the EU operation would be closely aligned to NATO, and hence the US. The fact that Solana was a former NATO Secretary General may have helped his ability to cooperate with the UK. A general shared "feeling" to set the record straight after the inability to effectively deal with the Balkan crises of the 1990s was also shared by the Bystanders. In the end, with the exception of the Commission, the operation met no significant opposition.
Turning to troop contributions, we can observe that with important contributions from different advocacy coalitions, EUFOR Althea was not dominated by one particular coalition. Most troops were provided by Germany, the UK and Italy, followed by considerable contributions of the Netherlands, Spain and France (Table 1). All member states contributed to EUFOR Althea except Cyprus, Malta and Denmark. And while the transition from NATO's SFOR to an EU operation While the differences between the advocacy coalitions did not prevent the launch of the operation, they were reflected in different views on what the operation should look like. In line with the Euro-Atlanticist position, the mandate and number of troops of EUFOR remained similar to those of SFOR. Moreover, NATO remained present with a more robust operational capability than just an advisory or support presence (Kim 2005). However, with the reference to the SAP a clear link to the EU's broader engagement with BiH was added, which reflects the preferences of the Global Power EU and Human Security coalitions.
Yet this linking of EUFOR to the SAP was against the wishes of the European Commission. Indeed, the Joint Action clearly subordinated the European Commission to EUFOR, stating: 'the intention of the Commission [is] to direct, where appropriate, its action towards achieving the objectives of this Joint Action' (Council of the European Union 2004a; see also European Council 2004). This would mean that SAP, an instrument of the European Commission, would follow EUFOR rather than the other way around. The fear of the military logic taking the upper hand was further fuelled by the fact that the function of the crisis response coordination teams, in which the European Commission was present, was limited to being a forum where the European Commission was "informed", rather than being an equal partner (Schroeder 2007).
As indicated, another contentious issue concerned the importance of the fight against organized crime as part of EUFOR's mandate. Eventually, the fight against organized crime was not included in EUFOR's mandate but ended up as a key supporting task in the concept of operations, as drawn up by Operation Commander Leakey (Dutch Ministry of Defence 2004; see also Dijkstra 2013, 115). In practice, EUFOR's proactivity regarding the fight against organized crime would differ per geographic area of the military operation; that is, the UK-led area did contribute to this key supporting task, while in the Finnish-and Spanish/Frenchled areas EUFOR's mandate was understood more narrowly (Dutch Ministry of Defence 2004, 10).
Finally, Solana proposed strengthening the position of the European Union Special Representative (EUSR). However, this did not make it into the Council Joint Action, which limited the role of the EUSR to giving 'political advice' (Solana 2004).
In sum, while the Euro-Atlanticst and Global Power EU coalitions were successful in initiating an EU military operation, they did not manage to get the fight against organized crime written into the mandate. This was only included as a key support task in the Concept of Operations. The explanation for this may be found in the fact that, next to the Bystanders and the Human Security coalition, important actors in both coalitions were more hesitant towards EUFOR's activism. By including it as a key supporting task the different actors could "agree to disagree", allowing the different contributing countries to decide themselves upon their activism on this task. Thus, while a small number of actors have been active in pushing the operation, the particular character of EUFOR Althea shows that there was a deal between different coalitions, taking it beyond the lowest common denominator outcome.
The changing character of EUFOR Althea 77

The 2007 changes of EUFOR Althea: reducing troops and strengthening the position of the EUSR
While the Council Joint Action that launched EUFOR Althea in 2004 has not been officially amended, there have been considerable changes in the character of the operation. I zoom in on two periods of change: 2007 and 2010-2012. In this section, the changes in these two periods are examined as well as the internal and external factors that occasioned them. Thus, we can assess whether these changes are best understood as reflecting shifts in the power balance between the coalitions or rather as processes of collective learning.
By the end of 2005, and coinciding with the succession of Force Commander Leakey by the Italian Commander Chiarini, the emphasis of EUFOR Althea on fighting organized crime disappeared (Dijkstra 2013;Friesendorf and Penksa 2008). Subsequently, in 2007, two substantial changes were adopted. First, the number of troops was reduced to 2500, accompanied by a change in the configuration of troops. The Netherlands and the UK withdrew their involvement, while Italy and Spain stepped in as main contributors, followed by Germany, Poland and Austria. 2 Second, the position of the EUSR was strengthened. Whereas the position of the EUSR vis-à-vis the EUFOR Commander initially was described as "giving local advice", this changed by the end of November 2007 into "providing political guidance" (Council of the European Union 2004b;2007a). These two changes were complemented by an important non-change: the mandate continued to underline the peace-enforcement character (Council of the European Union 2007b).

Troop reductions: external demands and unilateral withdrawals
The reduction of troops was first of all justified for external reasons, namely the rather positive evaluation of developments in BiH. Yet the way the reduction was executed and the actual reconfiguration that took place suggest little of a collective learning process, but rather the strategic behaviour of particular coalitions.
For sure, there were clearly positive developments on the ground. Thus, the EUMC could conclude that the security situation in BiH had become stable enough for a "reposture" of the number of troops (EU Military Committee 2007). Already, in the first years of EUFOR Althea, the EU reported considerable progress in BiH, and the EUSR noted that 'BiH authorities are assuming increasing responsibility' (Council Secretariat 2006). On this basis, negotiations for a Stability and Association Agreement (SAA) were started in November 2005, and were concluded in June 2008. Similarly, the Six Month Review of September 2008 stated that EUFOR had accomplished its mission to a great extent (Dutch Ministry of Defence 2008).
Yet, in practice, the EU Military Committee (2007) noted that the reduction took account of 'Member State aspirations', which reflects the fact that the troop reductions were primarily motivated by instrumental reasons. In particular, the UK referred to its 'overstretch' in Iraq and Afghanistan, to justify a reduction of troops in BiH (Oliver 2007). This was perceived by France and Germany as 'rather brutal ' (cf. Pohl 2014). In 2009/2010, France and Finland reduced their troops to a minimal contribution as well, which was qualified by the Netherlands as a 'unilateral withdrawal' (Dutch Ministry of Defence 2008).
Still, despite the withdrawal of troops of its main proponents, the UK and the Netherlands, the operation largely retained its main characteristics. Most notably, the mandate of the operation was left unchanged-that is, the executive mandate was neither removed nor expanded.

Strengthening the position of the EUSR: internal learning
In contrast to the logic underlying the decision on troop reduction and reconfiguration, the strengthening of the position of the EUSR did seem to reflect an internal learning process. While High Representative Solana had not been able to strengthen the position of the EUSR from the start, he continued pushing the issue. It figured prominently in a joint report that he published in 2006 together with EU Commissioner Rehn, Reinforced EU presence in Bosnia and Herzegovina (Solana 2006). Moreover, strengthening the position of the EUSR was one of the main recommendations that resulted from a case study prepared for the Council on the 'coordination and coherence between EU Special Representative, the EU military operation and the EU Police Mission in Bosnia and Herzegovina' (Council of the European Union 2006a;2006b;Politico-Military Group 2006). The evolution of positions on the strengthening of the EUSR was most visible in the changing position of the European Commission, which now actively supported the strengthening of the EUSR, as is indeed visible in the joint report by Solana and Rehn. In sum, the troop reduction and reconfiguration were primarily an issue of strategic calculations of the different coalitions rather than a process of collective learning, although the consensus on the positive development in BiH provided a common ground for discussion. In contrast, the strengthening of the position of the EUSR was a clear instance of internal learning.

The 2010-2012 changes of EUFOR Althea: another reconfiguration of troops and the inclusion of non-executive tasks
While the EU has not amended the 2004 Joint Action that laid out the mandate and objectives of EUFOR Althea, in 2010 the executive mandate was extended (not replaced) by including 'non-executive capacity-building and training support' for the BiH authorities (Council of the European Union 2010). Also, in 2012, another troop reconfiguration reduced the number to 600 and put even more emphasis on capacity building and training (Council Secretariat 2012;Politico-Military Group 2013). The 2012 reduction in troops left Austria and Hungary as the main contributors. These changes cannot be separated from the broader debate over whether or not to terminate EUFOR Althea, which had already started in 2008 (NRC Handelsblad 2008).
How come the extension of the mandate had not been possible in 2007, but it was possible in 2010/2012? Had there been a process of learning? Or did the balance between the different coalitions change?
The changing character of EUFOR Althea 79

Another troop reduction and reconfiguration: limits on learning
The troop reduction and reconfiguration of 2012 took place in a considerably different context from that of 2007. The economic crisis made it difficult to maintain. Moreover, the developments in BiH itself were increasingly evaluated negatively (Council Secretariat 2010a;2010b;Politico-Military Group 2011). Just as in 2007, the troop reductions followed primarily reflected strategic calculations of different coalitions and cannot easily be understood as a coordinated response.
The Euro-Atlanticst coalition understood the lack of political progress in BiH as a confirmation of the need for continuing Operation Althea. Notably, this was also the coalition with which the main contributors to EUFOR Althea in this period, Austria and Hungary, had most affinities (Póti and Tálas 2004). 3 However, the Euro-Atlanticist reading of the situation in BiH was challenged by others (primarily the Human Security and Global Power EU coalitions) who rather came to the view that the military operation itself actually inhibited political progress. 4 To the extent that this view can be seen as a form of learning, it notably took place within the boundaries of specific coalitions rather than being a collective process.
Another development that casts doubt on collective learning is the completion of the EU's Police Mission in BiH in 2012. The usual trajectory for the EU to phase out its external involvement is to have a military operation succeeded by a police mission. In that light, the termination of the EU's police mission in BiH, while at the same time keeping the military operation, gives contradictory messages and suggests an awkward settlement between the different advocacy coalitions, rather than a collective learning process based on the developments on the ground.
In sum, the troop reduction and reconfiguration of 2012 shows how a common evaluation of the lack of progress in BiH still allowed very different views on the added value of EUFOR Althea. This finding casts doubts on the ability of EU actors to engage in a process of collective learning.

The inclusion of non-executive tasks: a negotiated settlement
The discussion on changing the mandate of EUFOR Althea cannot be isolated from the overarching debate on whether or not to terminate the operation. The Global Power EU and Human Security coalitions were ready to end the EU's military presence in BiH.
After the informal EU meeting of defence ministers during the French Presidency in 2008, French Minister of Defence Morin indicated: 'it makes sense that European states at some point give the signal to civilians that when we are able to start a mission, we are also capable of closing one' (NRC Handelsblad, 2008). Following up on this, the Swedish Presidency in the second half of 2009 launched a discussion on EUFOR Althea, proposing to turn it into an advisory military mission. The position to turn the operation into a non-executive mission was also endorsed by Germany (Bundestag 2011). Most notably, even the (Italian) EUFOR Force Commander at that time recommended not to ask for a new executive mandate, but to transform Althea into a smaller training mission without an executive mandate (US Embassy Sarajevo 2009).
Yet, despite the fact that a majority of EU member states were in favour of closing down EUFOR Althea, the need for consensus has prevented this from happening (Sweden Parliamentary Committee 2013; Bundestag 2010). So, while in 2010/2012 the focus of EUFOR Althea changed to training and capacity-building, it remained under the mandate of a military operation.
Essentially, it is the Euro-Atlanticist coalition that insists on keeping the executive mandate and prevents the ending of EUFOR Althea. The crucial reason for this insistence is that the executive mandate serves to 'keep Berlin Plus alive'that is, to maintain the visible ties of CSDP with NATO and the US: The EU military operation in Bosnia. We know that it has to be transformed. We know that this is not the right answer to what Bosnia is facing right now. Since five years we have been saying so, but we cannot close the mission: it is the only and last EU mission with NATO assets. 5 CSDP will remain important for the EU. Althea's executive mandate is an example of this. … It is the only "real" military operation that is left. 6 While it was decided already by the end of 2009 to transform the operation into a training mission, the Euro-Atlanticist coalition was able to delay the timeline for doing so, with a reference to the Bosnian general elections of 2010. The UK played a central role in this, supported by the member states closest to Bosnia-Herzegovina. Moreover, by 2009, the proponents of a more activist military policy saw Solana leave office. The policy of his successor, Ashton, turned out to be much more aligned with the Bystander coalition.
In sum, the shift in the focus of EUFOR Althea to capacity-building and training indicates a negotiated settlement rather than a process of learning. While the Human Security and Global Power EU coalitions responded to the disappointing developments in BiH by pushing termination of the EU's military presence in BiH, the Euro-Atlanticist coalition does not want to close the operation. The different readings hamper a collective learning process. The unanimity that is required 5 Interview with former member of Solana cabinet A, 2013. 6 Interview with PMG-member B, 2013. for an exit of EUFOR Althea favours the position of the Euro-Atlanticist group. So, while the main proponents of the Euro-Atlanticist coalition that want to keep EUFOR Althea alive, the Netherlands and the UK, are not putting their money where their mouth is, the Euro-Atlanticist coalition is still the key driver behind the continuation of the operation.

Discussion
In analysing the evolution of EUFOR Althea since its launch in 2004, evidence of both power politics and collective learning has been found. However, on the whole, power politics appears to have been the main driver (Table 2). Changes have been primarily informed by national strategic interests, shifts in actor engagement and entrapment. From the start of the operation, the Euro-Atlanticist coalition, which favours an activist use of the military instrument, was able to put its mark on the EUFOR Althea. As troops were subsequently reduced, these reductions were primarily driven by a decreasing support for keeping EUFOR Althea alive (for example, Finland/France) and by competing external demands (the UK). And even when the main initiators of the Euro-Atlantics coalition, the UK and the Netherlands, withdrew their troops, they were able to prevent termination of EUFOR Althea because of the unanimity requirement for doing so. The outcome was to keep the 2004 military mandate, while shifting the focus to capacity-building and training. This points at a negotiated settlement with the other coalitions which, led by France, had sought to close down EUFOR Althea. More generally, the study shows that the initial mandate has a strong lock-in effect, making it difficult to change, or end, the mission after it is actually launched. Despite the active involvement of actors aiming at changing the character of EUFOR Althea, this proved to be very difficult. This is not to deny that there have been instances of a learning process. First, the strengthened position of the EUSR did not immediately affect EUFOR Althea. However, it provided a message that EUFOR Althea would be more strongly embedded in the EU's overall foreign policy. Second, the different coalitions agreed on their evaluation of the lack of progress in BiH, but their assessment of what that would mean for EUFOR Althea differed. This suggests that learning takes place primarily within the bounds of coalitions.
Finally, the study yields that Althea has informed cross-case learning: Althea, and the preceding NATO operation, have proven that that is the way how not to do it. We have been in Bosnia for 15 years without much progress. And the problems in Bosnia, for years, are no longer military. They are political, ethnical, and economical-that is where we should act. Because we did not handle this well from the beginning, we are now running into huge problems. … With Althea in mind we want that Somalis and Malians are able to do that. We don't want to be there for another 15 years. 7 Notably, the mandates of EU military operations after EUFOR Althea have always included an end date. Such signs of cross-case learning point at the potential of the EU to develop as an international security actor on the basis of past experiences, but this potential appears severely constrained by deeply rooted strategic cultures. This implies that the EU's use of force remains essentially contested and that the launch of any military operation remains to be decided by the balance of power between the different coalitions and their resources.

Conclusion
This paper departed from two theories of policy change in the CSDP: institutional learning and power politics. The existing literature either emphasizes learning and convergence among the main actors in CSDP or draws attention to persistently different strategic cultures. Whereas in the former "school", conflicting ideas and interests appear underestimated, in the latter these different ideas and interests seem impervious to interaction. To advance this debate, I adopt an advocacy coalition framework to assess the relative value of the two approaches in accounting for the changing character of EU military operations in the last ten years.
The analysis of EUFOR Althea shows a predominance of power politics, with a limited degree of learning. As such, this case study does not support the argument for an emerging common EU strategic culture. Nor does it provide a lot of optimism concerning the EU's ability to rise to the challenges arising from the ring of instability surrounding its borders.
While the advocacy coalitions are expected to be relatively stable and the actors to stay within the boundaries of the core beliefs of their coalition, this case study points out that these assumptions need some nuance. Most notably, while Austria's strategic culture fits with that of the Bystanders, in this particular case of Bosnia-Herzegovina proximity trumped its reluctant and hesitant position concerning the use of military force. This case study points at a prominent role for the Euro-Atlanticist coalition, and particularly the UK. With Brexit the actor constellations in CSDP change dramatically. While Brexit does not exclude the UK from participating in EU military operations, it will no longer have the strong position it has today, allowing it to veto the termination of Althea. Since the UK is the only country left that is standing firm on this issue-as it is in blocking an EU Operational Headquarters-a Brexit may result in a changing Althea into a non-executive training mission.

Disclosure statement
No potential conflict of interest was reported by the author.

Funding
This work was conducted at Vrije Universiteit Amsterdam and supported by a Research Talent Grant from the Netherlands Organisation for Scientific Research (NWO) [grant number 406-11-027].

Notes on contributor
Trineke Palm is postdoctoral researcher at the project "Blueprints of Hope: designing post-War Europe" at Utrecht University.