The Communitarization of the Area of Freedom, Security and Justice: Why Institutional Change Does Not Translate into Policy Change

This article proposes an explanation as to why institutional change &#8211; understood as more competences for the European Union's supranational institutions &#8211; has rarely led to policy change in the Area of Freedom, Security and Justice (AFSJ). It draws attention to the constraints that newly empowered actors have faced in the wake of introducing the co&#8208;decision procedure. If the key principles of a given AFSJ sub&#8208;policy &#8211; its &#8216;policy core&#8217; &#8211; were defined before institutional change occurred, the Council (as the dominant actor of the early intergovernmental co&#8208;operation) has found it easier to prevail in the altered structural environment and to co&#8208;opt or sideline actors with competing rationales. The article compares the importance of the new decision&#8208;making procedure with two alternative pathways potentially leading to policy change, namely, the power of litigation and the impact of unexpected external events.


Introduction
The European Union's (EU) co-operation in the Area of Freedom, Security and Justice (AFSJ) touches upon core functions of statehood including the safeguarding of internal security, the control of national frontiers, and access of citizens and non-citizens (in particular migrants and asylum seekers) to justice and rights. Therefore, decision-making in this field affects the level of (actual or perceived) security of individuals in Europe as well as their rights and liberties (e.g. Mitsilegas et al., 2003).
Given the sensitivity of these issues, it comes as no surprise that the process of European integration has been characterized by a high level of political contestation. When policy-making on justice and home affairs (JHA) was still dominated by Member States, the Council of the EU (Council) shared more security-oriented positions and tended to disagree with the EP (European Parliament)more prone to put civil liberties forward. Often in conjunction with the European Commission (Commission), the EP positioned itself as a 'pro-migrant' actor and a defender of citizens' rights on issues such as data pro-tection (Kaunert, 2010b, pp. 142-3;Hix and Noury, 2007, p. 202;Argomaniz, 2009, pp. 123-4). The influence of the EU's supranational institutions, however, was limited. Virginie Guiraudon (2000) conceptualized the EU's early co-operation as a form of 'venue shopping', in which like-minded strategic actors (typically from ministries of the interior) sought venues of decision-making at the EU level in which they were protected from domestic veto players. This particular set-up raised questions around the eventual shift to the 'Community method' (Dehousse, 2011). The process of communitarizing the AFSJ, thus, became an object of study in itself (see, for example, Kaunert, 2010b;Wolff et al., 2011;Niemann, 2008;Monar, 2010;Peers, 2011;Carrera et al., 2013).
The article presents the findings of a collaborative research project that has examined in a systematic, comparative and theory-informed way the impact of this process of communitarization on the different sub-policies of the AFSJ, ranging from the migration field through organized crime to data protection (Trauner and Ripoll Servent, 2015). The comparative analysis proposes an explanation as to why institutional changeunderstood as more competences for the EU's supranational institutionshave rarely led to policy change. The article draws attention to the conditions that facilitate or hinder policy change and categorizes the different mechanisms that have led to policy stability and change in the AFSJ.
The article starts by discussing its contribution to the state of the art. The analytical framework developed in the next section builds upon the insights of two academic communities far too often disconnected from one anotherinstitutional scholars and experts on policy change. In the empirical sections, we focus on the behaviour of actorsin the present case EU institutionswhen confronted with a change in their structural environment. This perspective is not only of interest for the specialized literature on JHA but also for more general debates on the role of EU institutions and dynamics of decision-making (for example, Thomson et al., 2006;Farrell and Héritier, 2007;Beach, 2005;Bickerton et al., 2014). With the changing involvement and decision-making powers of the EU's supranational institutions, the AFSJ provides a quite unique setting for this discussion.

I. Beyond the State of the Art
Departing from pure intergovernmentalism, the Treaty of Amsterdam introduced a first major shift towards communitarization by transferring the fields of asylum, immigration, external border controls and civil law matters to the first pillar under Title IV. Still, for a transitional period of five years, the competences for the supranational EU institutions were constrained and unanimity voting was kept in the Council. The Treaty of Lisbon ended this institutional development by introducing the Community method also in police and judicial co-operation in criminal matters.
In the first period after the AFSJ's communitarization, scholars assumed a 'rightsenhancing' effect to be stimulated by treaty reforms. According to Léonard (2012, p. 1405), for instance, the strengthened role of the EU's supranational institutions 'has reinforced the liberal character of the EU asylum venue, which renders the adoption of more restrictive asylum provisions less likely'. From this perspective, the period when Member States could enhance their discretion and power was only transitory. The supranational EU institutions would now play a more powerful role and drive the evolution of different AFSJ sub-policies in a direction not necessarily foreseen or desired by Member States (see also Ette et al., 2011, p. 31;Roos, 2013).
These expectations have recently been nuanced for several AFSJ sub-policies, often in reaction to an analysis of the policy output achieved under the new co-decision procedure. For instance, focusing on the second generation of EU asylum laws, El-Enany (2015, p. 889) states that 'the asylum Directives contain broad scope for discretion in application and suffer from incomplete or mal-implementation, resulting in vast differences in practice across the EU in the level of protection afforded to asylum-seekers, as well as diverse rates of recognition of refugees'. In a similar vein, Peers (2013, p. 16) describes the changes in these revised asylum laws as 'being merely a "lipstick on a pig"'.
The fact that the EP has altered its positions after communitarization and/or has had limited impact in triggering change has been noted in other AFSJ sub-policies as well, including migration (Acosta, 2009;Lopatin, 2013) and data protection (Ripoll Servent, 2013). In contrast, the CJEU (Court of Justice of the EU) continues to be seen as a potential game changer. In the immigration field, the Court rulings are considered to have 'constrain[ed] the executive branch of Member States governments' (Acosta and Geddes, 2013, p. 179), in particular regarding decisions of expulsion and family reunification (see also Bonjour and Vink, 2013).
This article makes a twofold contribution to this debate. First, it seeks to go beyond the studies examining the role of a particular EU institution and/or the degree of change in a given AFSJ sub-policy. It aims to take a bird's eye perspective on this area as a whole, with the objective to overcome its increasing fragmentation and specialization and to grasp the nature of this evolving regime. The systematic analysis of all major AFSJ sub-policies demonstrates that there has been a high degree of policy stability in the wake of treaty reforms affecting the AFSJ. In this sense, the article confirms and extrapolates the findings that some scholars have made on individual sub-policies of the AFSJ. The second contribution is a theoretical one: the article argues that the absence of change in many AFSJ sub-policies after communitarization can be explained by focusing on the procedural and normative constraints actors face in a situation of institutional change.

II. Linking Institutional Change and Policy Change
For this project, we focus on the effects of treaty reforms as instances of formal institutional changealthough we recognize that this kind of change can come also from more informal sources (Farrell and Héritier, 2007;Shackleton, 2000). Given that the term 'institution' is used differently in EU terminology and the wider academic field (Stacey and Rittberger, 2003, p. 890), it requires a conceptual clarification. We label the European Parliament, the Council, the European Commission and the Court of Justice as 'EU institutions'. If the term 'institution' stands alone, it refers to a set of rules and practices that guide the interactions of actors in a given structural context (see March and Olsen, 1998, p. 948;North, 1990).
Our first expectations regarding the impact of the treaty reforms introduced in Amsterdam and Lisbon derive from the work of scholars interested in EU policy-making. The change of the decision-making procedure from consultation to co-decision has led to the introduction of new veto players into the law-making process. As a result, EU institutions are likely to adapt their behaviour to this altered opportunity structure (for example, Thomson et al., 2006). Given the new procedural rules, they have to develop new strategies in order to reach the necessary inter-institutional majorities. This may be achieved by building strategic alliances (for example, between the Council and specific EP party groupings) or by framing one's position as the most legitimate and appropriate one. The need to redefine the balance of powers may put into question existing beliefs (Sabatier and Weible, 2007, p. 204) or introduce alternative normative frames that need to be accommodated (Clemens and Cook, 1999). Therefore, policy entrepreneurs may actively foster the advent of new standards of legitimate behaviour or modify existing ones (Béland, 2009).
By building upon these scholarly insights, the article seeks to go one step further and explain why the Council has remained de facto an advantaged policy entrepreneur in the AFSJ regardless of the de jure empowerment of the EU's supranational institutions. We explain this phenomenon by focusing on the nature of a particular public policy as a factor influencing the opportunity structure that actors have in a situation of formal institutional change.

The Power of a Settled 'Policy Core'
Our argument is influenced by the work of Jenkins-Smith and Sabatier (1993), who differentiate between a 'deep (normative) core', a 'near (policy) core' and 'secondary aspects' of a public policy. The 'deep normative core' relates to the 'fundamental normative and ontological axioms' of the belief system of political elites that underlie a policy system and is very difficult to change. The 'near (policy) core' is defined as the 'fundamental policy positions concerning the basic strategies for achieving the normative axioms of deep core'. The 'policy core' is vital to the present research, given that this dimension reveals how actors seek to realize their deep normative beliefs. 'Secondary aspects' are moderately easy to change as they relate to 'instrumental decisions and information searches necessary to implement the policy core' (on these categories, see Sabatier, 1993, p. 31).
We argue that the existence of a settled 'policy core' has put particular actorsespecially those that were involved in its definitionin a more powerful position to lock in and legitimize the status quo in a situation of institutional change. If the newly empowered institutional rivals have pushed for policy change, the previously leading actorin our case, Member States in the Justice and Home Affairs Councilhave been capable of insisting on solutions close to the status quo. Non-agreement has automatically implied the maintenance of policies largely defined by the Council under consultation. This is not to say that policy change has been impossible to achieve after the introduction of new procedural rules. In principle, the 'core' of any given policy can be altered in the political process regardless of whether or not it has been set before institutional change takes place. Yet, a settled 'policy core' has constituted an advantage for the hitherto dominant actor, which has found it easier to sideline or co-opt the newly empowered actors in a situation of institutional change.
Policy developments in the AFSJ can certainly be influenced by factors other than treaty reforms. For instance, unexpected external events have often been seen as a major source of policy change in the AFSJ (for example, Monar, 2006). In order to evaluate the importance of these alternative pathways leading to policy change, the article also examines the impact of unexpected external events and court rulings.

Unexpected External Events as an Alternative Pathway of Change?
Many unexpected external events have triggered an enhanced level of activity at EU level. Yet, the long-term impact of these kinds of external events (not only terrorist attacks, but also particular conflicts at a national level or governmental changes in Member States) depends on how they are absorbed, used and interpreted by actors involved in the EU decision-making process.
We consider, therefore, that no external event unfolds an impact on its own. Pivotal actors such as Commission officials or the Council presidency may appeal to these events to shift the belief system and raise the sense of urgency and necessity for specific instruments such as the European Arrest Warrant or the Data Retention Directive (Kaunert, 2010a;Ripoll Servent, 2013). External events can therefore be seen as 'windows of opportunities' for actors to push for change that would be difficult to achieve under other circumstances (see also Kaunert, 2010b).

The Power of Litigation
The impact that the Court of Justice may have on different EU policies is well established in the literature (for an overview, see Stone Sweet, 2010). The Court may have 'strategic value as an instrument of European legislation' (Scharpf, 2006, pp. 852-3) if seen in conjunction with the Commission's competence to ensure compliance with European legislation (Wasserfallen, 2010). The interactions between the Court and the Commission can limit the margin of manoeuvre for other actors who wish to change the direction of a policy (Schmidt, 2000). Actors may use the jurisprudence of courts to prevent any deviation from the established judicial precedent or to fight for new competences at the EU level (Greer, 2006).
Beyond their immediate legal implications, actors may also discursively rely on rulings to enhance the legitimacy of their policy choices. Hence, our final expectation is that Court rulings may contribute to either policy change or stability depending on whether institutional actors make use of litigation to 'lock in' or to foster an alternative policy rationale. It is important to keep in mind that most CJEU rulings contain some legal ambiguity so their impact is never clear-cut. Legal ambiguity may result in a ruling having only a limited effect on a policy or, to the contrary, compel policy-makers to implement far-reaching changes (Blauberger, 2014;Schmidt and Keleman, 2012).

III. Policy Change and Stability after the Introduction of Co-decision in the AFSJ
Our comparative analysis demonstrates that most 'policy cores' of the AFSJ, once established, have kept a high level of stability regardless of the introduction of the Community method.
In asylum policy, for instance, the second generation of asylum laws negotiated under co-decision only marginally affected the core of the asylum regime defined by the Member States before 2005. The recast primarily fine-tuned and tweaked existing asylum instruments and laws (Ripoll Servent and Trauner, 2014). Similarly, the EU's early co-operation on EU border control put the field on a certain path. According to Roderick Parkes (2015, pp. 57-8), the 'aggressively exclusive trans-and intergovernmental cooperation founded between interior ministers [in the framework of the Schengen co-operation] laid the bases for later EU border policies'. In general, the measures adopted after incorporating the Schengen acquis into the EU's legal framework merely built upon and refined already existing laws and practices. In counter-terrorism, periods of relative inertia were followed by intense activity after terrorist attacks. Particularly after 9/11, Member States stopped treating counter-terrorism as a national concern and turned it into a European priority (MacKenzie et al., 2015). The core of the policy has remained unaffected by the Lisbon Treaty, even though the latter set co-decision as a standard.
Some AFSJ sub-policies do not possess a clearly defined 'core' yet. The field of data protection, in particular, has been regulated by a set of very diverse and disparate rules. The result is a patchwork of rules and regulations without an established 'policy core' (de Hert and Papakonstantinou, 2015). The EU has become increasingly aware of the problems resulting from this piecemeal approach and has sought to provide a more coherent framework. The 'package approach' taken to negotiate the General Data Protection Regulation (Regulation 2016/679) and the Police and Criminal Justice Data Protection Directive (Directive 2016/680) is a reflection of this quandary. The core of EU immigration policyaccording to Andrew Geddes (2015, p. 78) 'who enters, on what basis, for how long and in what number'has essentially remained in the hands of Member States. The EU primarily constitutes a new governance layer 'intervening' in how they frame and define their national policies. The EU's policy in itself has remained fragmented with policies on curbing irregular migration being more developed than those on admission (Geddes, 2015, p. 78).

Maintaining Policy Stability by 'Locking In' a Policy Core
How to explain this comparatively high level of policy stability after the introduction of co-decision? Given that communitarization has developed over a long period of time, those actors wishing to prevent future changes often had the chance to determine the short-to medium-term policy development before institutional change took effect. Member States sidelined potential institutional rivals by shifting decision-making to 'venues' in which they still had a dominant position. For instance, they used intergovernmental co-operation fora such as the 'G6' meetings or the 'Prüm Treaty', which reflects 'how member state executives continue to "escape to Europe" in an active search for more amenable venues' (Ette et al., 2011, p. 31).
Looking at the different sub-policy areas, one can detect several episodes where Member States attempted to sideline institutional competitors by pre-empting their empowerment. In EU police co-operation, for instance, the issue of who 'guards the guardians' (Wagner, 2006) has been particularly salient. The EP was eager to get a stronger grip on the work of Europol (European Police Office), yet it faced strong opposition from Member States (Trauner, 2012). The transformation of Europol from an intergovernmental body into an EU agency appeared to be a good occasion for the EP to improve its standing. However, the Council speedily adopted the Europol decision a few months before the Lisbon Treaty entered into force in order to avoid the enhanced EP's decisionmaking powers (Den Boer, 2015). In the 2009 Council decision, Member States only marginally enhanced the EP's control and consultation rights vis-à-vis the agency.
Another kind of sidelining consisted in preventing an institutional competitor from developing a different policy rationale. By insisting on a specific solution as the most legitimate one, actors managed to institutionalize a particular rationale and relegated alternative understandings into the background. This happened in EU civil justice, where the inter-institutional controversies revolved around the issue of the cross-border limitation of EU competences (Storskrubb, 2015). The main debate there hinged on whether this cross-border limitation in civil justice matters should be interpreted in a narrow or an expansive way. The Commission was clearly in favour of an expansive interpretation, yet a coalition of the Council and the EP favoured a narrower interpretation and constantly countered the Commission's efforts to expand the remit of civil justice co-operation. A factor explaining this relatively stable pattern of inter-institutional co-operation relates to the fact that the EP's legal affairs committee (JURI)and not the committee on justice and home affairs and civil liberties (LIBE), traditionally dealing with AFSJ-related mattersnegotiated with the Council. This committee was more experienced in negotiating under co-decision and had already internalized a consensusoriented norm of behaviour. 'This institutional norm has often appeared in the EP's justification why it agrees with the Council in the field of civil justice and has legitimated the more limited advances in terms of policy change and further EU integration' (Storskrubb, 2015, p. 213).
In some cases, the process of sidelining was only partially successful and instead led to a juxtaposition of policy rationaleseach aiming to maintain the preferred policy 'core' of an institutional actor. In the case of EU border policy, no inter-institutional coalition managed to effectively sideline its institutional competitors. The three EU institutions involved in the decision-making process have failed to reconcile their different approaches and visions of what purpose 'borders' are supposed to fulfil after communitarization (Parkes, 2015). As a result, their different agendas continue to compete and/or coexist rather than merge in an integrated way. Roderick Parkes refers to the reform of the Schengen Borders Code as an example of unfinished 'sidelining'. Agreed upon in May 2013, the reform was initiated in the wake of the increased migration flows triggered by the Arab Spring. As a result, 'whilst the Council acceded to the moves to supranationalise the system, it ensured that the decision to reintroduce borders remains in the hands of member States. Two very different visions of Schengen governance, one heavily supranationalised, one nationalised, now sit side-by-side in the reformed system' (Parkes, 2015, p. 66). This policy outcome is described as negatively impacting the effectiveness and democratic quality of EU border control policies.

Maintaining Policy Stability by Co-opting Competing Actors
While the EP and the Council have been formally on equal footing after the introduction of co-decision, the Council has been de facto in a better position in those AFSJ subpolicies where the policy core had already been settled. In these situations, the Council could more easily risk a failure of inter-institutional negotiations, since it enjoyed a first-mover advantage with which the other EU institutions struggled to catch up. In practice, this meant that the Council could often negotiate in a more uncompromising (and effective) way. The Council also fostered the advent of new, more consensus-oriented procedural norms, which led to the co-optation of certain EP actors, mostly from the centre-right and conservative groups.
This mechanism explains, for instance, the limited degree of policy change in EU asylum. In this field, the Council acted with a high level of internal cohesion. Member States felt no urgency to revise the first-generation asylum laws. They were more inclined to let the negotiations fail than to accept a change in the 'policy core' that had been defined under consultation up until 2005. The EP's position, by contrast, tended to fluctuate depending on the ability of political groups to form coalitions. Centre-left groups in the EP supported the Commission's attempt to draft new proposals that underlined the added value of more harmonized and liberal asylum rules. However, a coalition between the EP's left-wing and liberal groups was made more difficult after the 2009 parliamentary elections, which resulted in a conservative-led majority less prone to oppose the Council on security-related issues. In such a divided EP, the liberal group became a pivotal actor. The Council's success in convincing Commissioner Malmström to issue revised proposals in 2011 resulted in a de facto rapprochement of the two EU institutions. This co-optation helped to convince the liberal group to abandon the liberal-left coalition and join the EPP (European People's Party)-Council coalition. This breakthrough led to an agreement in June 2013 on an 'asylum package', where most of the Council's central demands were accepted. The behaviour of liberal parliamentarians was driven not only by the negotiation dynamics but also by new consensual practices that gained importance under the altered procedural rules (for details, see Ripoll Servent and Trauner, 2014).
This case illustrates a broader pattern. In the 2009-14 legislative term, ALDE frequently played a pivotal role in deciding the success or failure of coalitions within the EP. The Parliament's centre-right EPP tended to propose solutions close to the Council, which gave rise to an inter-institutional coalition that focused on co-opting the liberal political group. In several key decisions, liberal MEPs shifted gear and eventually accepted the positions and discourse of the EPP-Council coalition. One of the clearest episodes was the negotiations on a SWIFT (Society for Worldwide Interbank Financial Telecommunication) Agreement with the United States, which affected the protection of personal data. The liberal group shifted from a negative vote on the interim agreement in February 2010 to arguing in favour of a re-negotiated agreementeven though this second agreement was not substantially different from the one they had originally rejected (MacKenzie et al., 2015, pp. 107-9;Ripoll Servent and MacKenzie, 2012).
The same dynamic helps us to understand why the EP accepted an early agreement of the first (and highly controversial) EU immigration law negotiated under co-decisionthe EU's Returns Directive (2008/115/EC). Under pressure from Member States, the EP accepted the law as the 'least worse option', fearing that a rejection of the early compromise achieved in the trialogue negotiations would lead to a reopening of the package deal and result in even more restrictive solutions (Geddes, 2015;Acosta, 2009). Diego Acosta (2009, p. 39) notes that 'if the European Parliament does not have the possibility to go to the "second reading" because of its own incapacity or fear of consequences in the Council, then the co-decision process cannot be considered as a fair procedure among equal institutions'.

Triggering Policy Change: Legitimating a New Policy Rationale
EU criminal law has changed to a larger degree in the post-Lisbon era. Its core was defined in the post-Amsterdam period when the EU adopted a wide range of framework decisions ranging from trafficking in human beings to corruption in the private sector and racism and xenophobia. Still, post-Lisbon, new EU competences in the field of criminal proceedings (Article 82(2) TFEU -Treaty on the Functioning of the European Union) and a stronger focus on the rights of defendants and victims have weakened the overly security-driven rationale of this policy (Mitsilegas and Vavoula, 2015). Policy change in this case has happened due to a combination of factors including a high level of cohesiveness of particular institutional actors and the active use of alternative frames and norms to legitimize a shift in a policy rationale.
In EU criminal law, the negative side effects of mutual recognitiona principle that has dominated the field since the Treaty of Amsterdambecame eventually more visible. Flagship measures such as the European Arrest Warrant received increasing criticisms for lowering human rights standards and triggered numerous judicial contestations (Sievers and Schmidt, 2015). This enhanced the pressure to approximate substantive criminal law to mitigate the differences between EU Member States in key defendants' rights and standards, such as the access to a lawyer upon arrest (Wagner, 2011;Lavenex, 2007).
In this environment, the EP (often in conjunction with the Commission) acted as a cohesive entrepreneur and pushed for a more rights-based approach. A case in point was the EU's policy against human trafficking. Initially conceived as a means 'to enhance the crime-fighting tool box of law enforcement officials' (Mitsilegas and Vavoula, 2015, p. 148), the EU eventually focused more on the protection of victims. The Directive on Human Trafficking (Directive 2011/36/EU)the first law in this field negotiated under co-decisioncontains detailed provisions for the assistance of victims, particularly children, and establishes an EU anti-trafficking coordinator. The Parliament was successful in inserting these provisions by achieving an oversized internal majority that included both centre-right and centre-left groups. It also referred back to certain rulings of the European Court of Human Rights (ECtHR), which exerted judicial and normative pressure on Member States by asking them to better protect women against trafficking and exploitation in the sex industry (for example, case Rantsev v. Cyprus and Russia, see Mitsilegas and Vavoula, 2015). By framing its solution within the understanding of the ECtHR and giving it overwhelming support from all ideological positions, the EP imbued its solution with greater legitimacy and was in a position to force a change in the policy rationale that Member States would be hard-pressed to oppose.

IV. The Impact of Unexpected External Events
In the AFSJ, the terrorist attacks of 11 September 2001 were exceptionally influential; they contributed to developing a new threat perception and, thereby, a new understanding of the EU's role in 'high politics' (den Boer and Monar, 2002). In turn, actors used this new framework to develop some AFSJ sub-policies with a security-driven focus and/or with a stronger EU component. For instance, the Commission was particularly keen to seize the terrorist attacks as a 'window of opportunity' to push for a stronger EU involvement in counter-terrorism (MacKenzie et al., 2015). Similarly, after 2001, the fight against terrorism was used to enhance the processing of personal data for security purposes. This development revealed major division between those 'willing to sacrifice a level of data protection … for increased security within the EU and those with pro-data protection stances' (de Hert and Papakonstantinou, 2015, p. 185).
The 9/11 terrorist attacks were indeed at the source of EU actors' efforts to develop the AFSJ. However, external events have had less impact on the EU institutions' efforts to change a particular AFSJ sub-policy. Most of themno matter how much media attention they initially receivedonly had limited influence in terms of triggering far-reaching policy change. In many instances, the Council used a perceived situation of crisis to counter efforts of actors aiming to introduce a less security-oriented rationale. In border policies, for instance, the Council used crises 'to bounce Parliament into first-reading agreements and browbeat it into "behaving responsibly"' (Parkes, 2015, p. 65).
It is important to mention that it has been too early for this research to assess the impact of the unprecedented influx of migrants arriving in the EU starting in 2015 and the November 2015 terrorist attacks in Paris, where Islamist terrorists killed 130 and wounded more than 350 people. Given the severity of these events, the EU has sought to come up with new approaches. This has often happened in intergovernmental settings (such as the 'leaders' meeting on refugee flows along the Western Balkan route') marginalizing the Commission and, especially, the EP. In EU counter-terrorism, the Council (2015) quickly agreed on a range of measures such as increased checks on EU citizens travelling abroad and more control of the circulation of illicit arms. In reaction to the migration crisis, the EU adopted a legally binding relocation scheme for 160,000 migrants within Europe and started to negotiate a standing European border force.
While these measures seem to indicate the advent of a higher order of change in EU asylum and border control policies, a note of caution is needed. By early 2016, the EU still refrained from touching the key pillars of its asylum policy, notably the Dublin regime. The relocation scheme, in combination with the 'hotspot' approach, is meant to provide an additional layer of instruments to the Dublin regime. The aim has been to make the Dublin system compatible with a situation of crisisbut not to replace it (for details, see Trauner, 2016). Furthermore, some Member States, in particular in eastern Europe, have continued to challenge politically and judicially the mandatory relocation quota and oppose a further transfer of national powers on border control (FT, 2015). Given these cleavages amongst Member States, it remains to be seen to what extent the core of EU asylum and border control policies will be altered in reaction to the migration crisis.

V. Litigation: 'Locking In' Stability or Legitimizing Change
Litigation may lock in a particular policy rationale. This happens if a series of court rulings contributed to settling a specific policy corewhich made it more difficult for rival actors to change it. The most evident judicial 'lock-in' in the AFSJ took place in EU citizenship.
The field of EU citizenship is exceptional, since it has always been decided under the Community method. The controversies in the field concerned primarily how many 'rights' should be linked to 'being' an EU citizen and who decides on these rights. Building upon the 'skeleton of provisions' (Kostakopoulou, 2015, p. 159) defining EU citizenship in the Maastricht Treaty, the CJEU's case law steadily enhanced the rights linked to the status of EU citizens. The process raised substantive yet ineffective criticism from Member States about the 'quasi-legislative role played by the Court and its relation to Member State autonomy' (Kostakopoulou, 2015, p. 160). The Commission eagerly followed up this case law by proposing rights-enhancing legislation and framing the debate in a way that resonated with the Court's rulings. In other words, the Commission used the case law of the Court to stabilize and deepen a rights-enhancing rationale in EU citizenship. While litigation can be used in a 'lock-in' mechanism, it can also serve to legitimate alternative policy rationales, leading thus to policy change. This dynamic was of importance in the early days of EU criminal law (Mitsilegas and Vavoula, 2015, p. 140). The Council initially opposed any European Community/European Union involvement in the field of criminal justice. By referring several cases to the Court, the Commission contested this view and fought for the EU's right to adopt criminal law measures even in the absence of express treaty competence (Mitsilegas and Vavoula, 2015, p. 140). In cases such as the environmental crime case (C-176-03) and the ship-source pollution case (C-440/05), the Court sustained the Commission's point of view and allowed for the adoption of EU criminal law measures if the latter were necessary to achieve other Community objectives. Member States eventually accepted a need for stronger EU involvement in substantive criminal law and formalized the new competences in the Lisbon Treaty (Article 83(2) TFEU).
In EU immigration policy, individuals and migrant entrepreneurs have taken the lead in resorting to the CJEU's enhanced powers of the post-Lisbon era. The Court's decisions have been particularly relevant in the fields of expulsion, family reunification and integration, and have challenged some (restrictive) practices of Member States. According to Andrew Geddes (2015; see also Acosta and Geddes, 2013), this kind of dynamic has been discernible at the national level for several years already, with the EU now catching up. 'In post-Lisbon Europe there have also emerged social and political spaces linked to the role of courts that has been an important component of the politics of immigration in Europe since the 1970s and now has an EU dimension' (Geddes, 2015, p. 89).
There are, however, a range of AFSJ sub-policies in which the rulings of the Court have not been used to embark on far-reaching change. EU asylum policy is such a case. Certain practices, such as Dublin transfers of asylum seekers to other Member States, considered to have an ill-functioning asylum system have been much contested before national and European courts. As El-Enany (2015, p. 889) argues, 'Despite the CJEU's willingness to interpret EU asylum provisions in a manner that protects refugees, uniformity and complete application of the asylum Directives have not been achieved through the Court's exercise of its jurisdiction.' This shows that the existence of jurisprudence is not a sufficient condition for policy change. It needs to be picked up and actively incorporated into EU policy-making for it to have a long-term effect and become a legitimate rationale shared by a wide range of institutional actors (see Table 1).

Conclusions
This article has sought to establish whether and, if so, how institutional change has contributed to triggering policy change in the AFSJ. We argue that the Council has tended to be in a stronger position in the wake of treaty reforms affecting the AFSJ due to the fact that Member States had, by then, already defined the constitutional pillars of a whole range of AFSJ sub-policies. The newly empowered supranational EU institutions were often compelled to satisfy themselves with changes of secondary importance that had little impact on the constitutional pillars of a given AFSJ sub-policy. Furthermore, EU institutions adapted to the new procedural rules introduced by treaty reforms and changed their institutional practices. This has concerned first and foremost the EP, which has developed a more consensual behaviour vis-à-vis the Council and a feeling of shared responsibility for policy outcomes.
If policy change has been triggered, this has often been the result of a combination of factors. In EU criminal law, for instance, the introduction of a less security-oriented rationale was facilitated by a high level of unity within the EP in its ambition to develop a more rights-based approach and the possibility to legitimize its position by appealing to broader criticisms regarding the implications and negative side effects of the mutual recognition instruments adopted thus far. More broadly, the question about how EU actors absorbed and incorporated external events or court rulings into the EU's political process has been of high relevance. These 'external' influences have provided policy entrepreneurs with windows of opportunities to press for changeor to avoid change in a particular direction. However, as the empirical section has shown, in the last decade, litigation has been a more frequent and successful path than the use of external events to either lock in or change a policy rationale.
What are the implications of these findings? Although the dynamics of supranationalism have become more discerniblenotably due to the activities of the Court -Member States in the Council still appear to remain privileged policy entrepreneurs in the AFSJ. Their dominant role at the time when most 'cores' of the AFSJ sub-policies were defined and their capacity to present a united front when confronting the other EU institutions has allowed them to shape the policy debates and set standards of legitimacy that have proven difficult to modify for the newly empowered supranational EU institutions. A (tentative) look of the EU's reaction to the November 2015 terrorist attacks and the 'refugee crises' seems to confirm these findings: Member States have remained key in determining the depth and scope of policy change in the Area of Freedom, Security and Justice.