What is intergovernmental about the EU’s ‘(new) intergovernmentalist’ turn? Evidence from the Eurozone and asylum crises

Engaging with recent claims of increased intergovernmental dynamics, this article asks what exactly is intergovernmental about the EU ’ s major crisis-induced reforms. Drawing on central claims of both New Intergovernmentalism and Liberal Intergovernmentalism, it is demonstrated that the Eurozone reform and the asylum reform differ significantly regarding the role played by the European Council (NI) and the role of institutional expertise provided by supranational actors (LI). While the European Council played a central facilitating role in the Eurozone crisis and worked effectively with the Commission, which provided important technical expertise, expertise in the area of asylum still largely lies with the member states. The Commission therefore acted as a political stake- holder, thus estranging the European Council that subsequently acted as a reform blocker. This article is a first attempt to assess empirically the micro-level foundations of different types of intergovernmentalism and to nuance claims on the weakened role of supranational institutions.

2. The different role of supranational expertise, or institutional ideas and information, as a power differential in these reform processes.
By means of these two differences, we engage with intergovernmentalist theorising, both 'new' and 'old'. The first element, the central role of the European Council is at the heart of new intergovernmentalist (NI) theorising of EU decision making (Bickerton et al. 2015;Puetter 2014). According to NI, intergovernmental coordination and deliberation between the Heads has replaced the Community Method as the central node of the decision making. LI explicitly engages with the second element, the role of expertise as a power differential between intergovernmental and supranational actors (Moravcsik 1999: 268).
Regarding the different role of the European Council, in the Eurozone crisis, specifically the processes of setting up the European Stability Mechanism (ESM) and banking union, we witness a rather effective interplay between the European Council and the Council, which meant that the dossier moved relatively smoothly between the different levels. If an issue was deadlocked at Council (i.e. Eurogroup or Ecofin) level, it was taken up to the European Council, which would provide guidance and set deadlines for the ministers. In the asylum crisis, however, involvement of the European Council was rather perceived as a nuisance, if not hindrance, of the 'regular' decision-making channels. Repeatedly, the Justice and Home Affairs (JHA) Council was unable to make progress, as political leaders were due to touch upon the same issues (e.g. relocation or asylum, specifically border, procedures) at an upcoming summit.
Regarding the different role of institutional actors and expertise: in the Eurozone crisis, the member states relied heavily on expertise, ideas and information, provided by the institutional actors, specifically the substantive input of the Commission and the European Central Bank (ECB), in order to be able to come up with European level solutions. In the asylum crisis, on the other hand, expertise lay more with the member states, who have experience with processing asylum applications and managing borders, and who have the necessary legal precedent. The Commission was therefore perceived less as a 'technocratic' advisory and executive body and more as a stakeholder, taking up highly contentious and ambitious reform positions, especially in the immediate crisis response. This negatively affected its impact later in the asylum reform.
We proceed as follows: in the next section, we pinpoint the current inter-and intra-theoretical debate about the recent rise of intergovernmentalism. We then proceed to the comparison of the management of the Eurozone and asylum crises. We discuss the arguably most important reforms in both crises: the European Financial Stability Facility (EFSF), the ESM, and the banking union for the Eurozone crisis and the relocation mechanisms, the EU Turkey deal, and reform of the Common European Asylum System (CEAS) for the asylum crisis. We highlight crucial differences between the two crises that explain variation in the immediate and long-term crisis response. In the conclusion, we suggest how we can take forward intergovernmentalist theorising, by considering these differences regarding the timing and type of European Council involvement and the different role and type of institutional expertise.
The gap: intergovernmentalism in major crisis reform Few insiders or outside observers will deny that intergovernmentalism is on the rise in EU decision making. Nevertheless, intergovernmentalism as a theory is currently challenged from the outside as well as from within. For different reasons, neo-and post-functionalists question the applicability of intergovernmentalist theorising, specifically the predominance of member states bargaining, based on domestically determined (sectoral) interests, and the overall limited impact of EU level actors. Neofunctionalists highlight the continued (or renewed) relevance of functional interdependencies and spill-over in Eurozone and migration policies as driving forces in the EU's crisis responses Niemann and Speyer 2018). Post-functionalists point to the increased politicisation of EU decision making, which takes precedence overand can run counter tosectoral, economic interests (Hooghe and Marks 2009;Kleine and Pollack 2018).
Within intergovernmentalist theorising, 'liberal intergovernmentalism' (LI) (Moravcsik 1998) has been confronted by several challengers, the most prominent being 'new intergovernmentalism' (NI) (Bickerton et al. 2015). Within NI, we again find various sub-types, including 'deliberative intergovernmentalism' and 'intergovernmental union' that produce quite different expectations of the process and outcome of these crisis reform processes (Fabbrini 2013;Puetter 2014). The concept of 'intergovernmental union' (Fabbrini 2013: 589) refers to dominance, hierarchy of the creditor member states (primarily Germany) and a centralization of the policy regime, which stays very close to LI. The concept of 'deliberative intergovernmentalism' (Puetter 2014(Puetter , 2016, on the other hand, runs counter to LI in several important ways. By introducing the, arguably constructivist, elements of consensus and deliberation, it paints a different picture of member states negotiations, in which the Heads are less constrained by exogenous factors, more specifically domestically determined interests and preferences, that are central to LI (Puetter 2016: 611). This brings NI closer to 'supranationalist institutionalist' views, in which the European Council is portrayed as a forum for deliberation in which leaders seek consensual outcomes and are able to downplay political controversies, more than representatives of line ministries would be able to do in the lower level negotiations (Kleine andPollack 2018: 1499;Schmidt 2019Schmidt : 1550.
Moravcsik responded to these challenges by putting forward a 'new' or at least 'updated' version of LI ('LI 2.0 0 ), that can account for major crisis reform processes that do not touch directly upon economic interests, like the 'refugee crisis' (Moravcsik 2018;Zaun 2018). While downplaying the role of sectoral interests, it maintains the view that EU crisis responses were primarily determined by domestically determined interests, (asymmetrical) interdependence, and with little autonomous role for institutional actors, resulting in minor, lowest common denominator steps forward to deal with 'functional challenges' posed by the crises (Moravcsik 2018(Moravcsik : 1661(Moravcsik -2, 1670. This functional(ist) element has led other scholars to suggest that the crisis reforms, in fact, reflected a mix of intergovernmental and supranational elements, in which the intergovernmental actors and action channels were only able to produce series of sub-optimal solutions, that immediately set the stage for the next round of reform debate (Jones et al. 2016(Jones et al. : 1012, thereby putting endogenous change towards more integration back on the table. This proliferation of different types and sub-types can be considered problematic. Hooghe and Marks (2019: 1) are correct to note that we should not treat the various perspectives as competing theories, but rather as 'flexible bodies of thought that resist decisive falsification'. There are few scholars who still cling to the idea 'gladiator-like tests', in which two theories enter the arena and only one steps out (Beach and Pedersen 2019: 56). However, what is problematic is that both within and beyond the intergovernmentalist school of thought, interpretations of what makes a particular decision-making process (more) intergovernmental vary, which makes it difficult to compare and evaluate the different perspectives and the extent to which they can be considered complementary or competing on an empirical level (King et al. 1994: 28-9). With this article we do not intend to determine who is winning, but rather aim to clarify the grounds on which the different perspectives are competing. In the remainder of this article we intend to clarify the 'competing (and complementary) observable implications' regarding the role and influence of the European Council (NI) and the role and influence of institutional expertise (LI).

The role of the European Council
In this section, we first discuss the role of the European Council in theory, thereby engaging with NI. After this we compare the (quite effective) role of the European Council in the reform the Eurozone to the (less effective) role of the European Council in the reform of asylum policies.
The European Council in theory: engaging with new intergovernmentalism NI starts from the supposition that the centre of gravity in EU decision making has shifted, definitively instead of just temporarily, towards the European Council, and to a lesser extent to the Council (of Ministers). In this new system of 'European Council centered governance' (Ludlow 2016), it is the Heads themselves that exercise leadership, through a process of enhanced intergovernmental coordination, thereby increasing the scope of EU policy making, but limiting the transfer of sovereignty (Puetter 2014: 29). In this process of 'integration without supranationalisation, policies are Europeanised, without being Communitarised (Bickerton et al. 2015: 704). NI expects there to be less room for major legislative packagesa claim that is already somewhat disputed by the major packages (six-pack, two-pack, banking union, CEAS reform) launched during the two crisesand less scope for formal delegation of policy making initiatives to the European Commission in particular (Puetter 2014: 57-8).
This leaves us with the question: howmeaning by which means and through which mechanismsthe European Council steers the rest of the machinery? The European Council has been appropriately characterised as 'a political power station unconnected to the grid' (Van Middelaar 2019: 197-8), as is exemplified by many media-covered crisis summits with little impact or follow up. Agreements reached by the Heads still need to be transformed into feasible and meaningful solutions within the legal and institutional framework of the EU. The European Council does not have its own 'machine room' where policy proposals are further developed, nor is it sufficiently anchored to the existing machinery except through the office of the PEC, who has too small a cabinet and too little administrative support for it to effectively maintain overall guidance of machine room processes. The Sherpa-network of personal advisors to the Heads was very influential during the early stages of the Eurozone crisis, but it lacks the institutional infrastructure to become a permanent rival to the Coreper (II). 1 Moreover, we witnessed only very few occasions of a genuine 'deliberative intergovernmentalist' process in which the Heads themselves were actively crafting reform solutions. These include, most prominently an eventful weekend in May 2010, when the EFSF was set up and in March 2016 with the coming about of the EU Turkey Statement. For determining the role and influence of the European Council in the remaining reform process, we need to look beyond the level of the Heads and consider the interplay with the other EU institutions.
In this new system of intergovernmental policy coordination, there is one clear loser: the European Parliament is expected to be hardly involved (Puetter 2014: 233). NI's conjectures about the interplay between the European Council, the Council, and the Commission are more subtle than the labels 'competition' and 'decline' imply. The Council is generally expected to act in tandem with the European Council, enacting the decision of the Heads. The European Council then acts as the initiator, coordinator and overseer of the work of the Council. However, in dealing with the major crisis reforms, the height of the system (European Council) often operated rather autonomously fromand therefore out of sync withthe rest of the machinery. As will become clear from the subsequent comparison of the Eurozone and CEAS reforms, there are notable differences regarding the type and timing of European Council involvement that need to be considered, which explain why and how the European Council was more effective in the Eurozone than in the asylum reforms.
According to NI, the Commission is no longer the sole, or even the prime, initiator of new policy initiatives. However, by actively seeking coverage and endorsements of the European Council, the Commission has been able to codetermine the subsequent course of the decision making (Puetter and Fabbrini 2016: 636-7;Smeets et al. 2019: 679). All in all, NI thus constitutes a 'milder' version of intergovernmentalism, which leaves ample room for institutional initiative and collaboration across the intergovernmentalist-supranationalist divide. Again, there are notable differences in the interplay between the European Council and the Commission, during the Eurozone and asylum crises, specifically in the willingness of the Commission to operate within, instead of rallying against, these nominally intergovernmental actors and action channels.

The European Council in the Eurozone reform
Both insiders and close observers still disagree about the degree to which the European Council has provided leadership in the EMU reform. There is a 'benevolent' reading that sees the European Council providing active guidance on key elements of the EFSF, ESM, and banking union, resolving deadlock when the finance ministers failed to deliver. There is also a more 'sceptical' reading considering the European Council as a somewhat detached, and slightly erratic overseer, whose main job was to set impossible deadlines which were subsequently ignored (Interview PEC 1/2015; Interview Council 1/2015). A similar dichotomy applies to the PEC, which according to some merely acted as the go-between the German Chancellor and French President(s), while according to others operated as a crucial, independent node in the inter-institutional network (Interviews PEC 2/2015, 3/2015).
Both views are, in fact, correct and applicable at the same time. Both in the ESM and banking union negotiations, the European Council and its President provided guidance and oversight through a limited number of well-timed interventions. In NI, European Council involvement on the EFSF is exemplary of a process aimed at reaching consensus by means of deliberative coordination at the highest political level (Puetter 2014: 78). However, we need to bear in mind that the EFSF was a real crisis measure, taken under extreme (political and time) pressure, when the Eurozone was under attack from the financial markets and there was a dire need to come up with a response before the markets opened on Monday. This intensive type of European Council involvement was only exemplary for immediate crisis measures. In the medium to long term processes of setting up the ESM and banking union, the European Council's role was quite different.
For the ESM, European Council involvement was limited to a few, albeit crucial, interventions (Smeets et al. 2019: 684, 686). In December 2010, it was the European Council that needed to give the go-ahead to the minor, but for Germany 'highly significant', revision of (or rather addition to) Article 136 TFEU. The March 2011 Eurozone summit was required to give the finance ministers the go-ahead to look for ways to increase the lending capacity of the EFSF/ESM. The May 2011 Summit was required to endorse (without discussion) the set-up of the ESM, and to allow for further explorations of potential enhancements of the funds at technical levels. The July 2011 Summit was crucial for endorsing the work that was already being done to enhance the size and scope of the funds. In October 2011, the European Council endorsed the idea of EFSF/ ESM leveraging. Finally, in December 2011 the European Council set the stage for reassessing the combined lending capacity of the funds. At none of these European Council meetings was the ESM the central topic of debate between the Heads. The ESM was discussed briefly, while other, more immediate concerns such as private sector involvement in July 2011, or the Fiscal Compact in December 2011, held centre stage in the debate (European Council 2011a, 2011b). The main role of the European Council was to spur on the finance ministers, by providing them with instructions and making sure that they delivered.
For the banking union we see the same pattern. The June 2012 Eurozone Summit and subsequent European Council are widely portrayed as the moment at which the Heads launched the banking union (European Council 2012a). In reality, the debate had been about recapitalising the Spanish banking sector, and common supervision was considered by some, most notably the German Chancellor Merkel, a precondition. It is highly doubtful that the Heads were aware that they had taken a major step in European integration. The October 2012 European Council removed the highly contentious issue of a common deposit guarantee scheme from the equation, again without substantive debate. This was rather decided in the run-up to the Summit, by actors close to PEC, Van Rompuy (Nielsen andSmeets 2018: 1241). The December 2012 European Council endorsed the banking union package and set new deadlines for the finance ministers on the single rule book (SRB) and the single resolution mechanism (SRM).
The June and December 2013 European Councils subsequently welcomed the deliverables, the bank recovery and resolution directive (BRRD), and the Single Resolution Mechanism (SRM) (European Council 2013a, 2013b). Work on the final element, a single resolution fund (SRF) that was to accompany the SRM, was again left to the finance ministers, and the Chair of the Eurogroup in particular. Generally speaking, the Heads did not feel inclined to intervene in the legislative dossiers, nor in the nominally intergovernmental negotiations taking place within the confines of the Eurogroup, Eurogroup Working Group (EWG), or in the IGC that was set up to deal with the SRF (Ludlow 2013: 18). Its role was again predominantly procedural, providing instructions, setting deadlines, monitoring progress, and welcoming agreements. This type of European Council involvement was quite effective in steering, rather than substantively guiding, the rest of the machinery.

The European Council in the reform of asylum policy
For the asylum reform, European Council involvement would look quite different. The Heads provided less procedural guidance but engaged more with the substance where it largely played the role of a blocker.
The immediate crisis response, including discussions on quota-based asylum-seeker relocation from Italy and Greece, was characterised by heated debates between on the one hand German Chancellor Merkel, backed by Dutch Prime Minister Rutte, the Swedish and initially the Austrian government and on the other hand the Visegrad states (V4), supported by the PEC, Tusk (Ludlow 2015(Ludlow : 10-2, 2016. Rather than acting in tandem with the European Council, the JHA Council, more specifically the Luxembourg Presidency, joint forces with the Commission, forcing through a decision on relocation by means of a Qualified Majority Voting, the day before the European Council was to meet and discuss the matter (Ludlow 2015: 20; Interview Council 1/2020). While Germany and its allies had a vested interest in supporting border countries, not least because they were negatively affected by onward movements of refugees, the V4 were not ready to take in additional refugees to support border countries andby extensionfinal destinations in North-Western Europe (Zaun 2018; Interview PermRep 1/2016).
The September 2015 QMV decision on relocation poisoned the atmosphere, and ensured that at the level of the Heads, the V4 were on their guard (Interviews Council 4/2019, 5/2019). Given the divisions in the European Council on temporary quotas, debates on permanent quotas never took off. As a compromise, some Member states suggested following the Commission communication from 23 September 2015 and addressing the issue under a comprehensive CEAS reform (Council 2015: 3).
In the meantime, some members of the European Council took it upon themselves to flesh out a deal with Turkey on stemming the flow of refugees/migrants crossing the Aegean Sea. The initiative came from Germany where chancellor Merkel was under pressure domestically from right-wing populists for opening the country's borders to refugees (Ludlow 2016: 46). Rather than working with the European Council as a whole, or mandating the Council, the German Chancellery co-opted the European Commission. This is an example of a supranational institution being very eager to contribute to an intergovernmental process, on a legally speaking purely intergovernmental deal. Within this intergovernmental framework, however, the Commission, personified by Vice President Timmermans, was able to put a strong steer on the process leading up to the EU Turkey statement of 18 March 2016. Germany and the Commission were supported by a 'coalition of the willing', yet, most member states were hardly involved in the process but instead were brought in at the very last minute to endorse a deal that had been made for them. This resulted in a deal that cannot be attributed to the EU, nor challenged before the CJEU, because it has no supranational legal quality.
In spring 2016, the Commission submitted its package of proposals for the reform of the CEAS. Unlike previous reforms, Council negotiations were highly controversial, and decisions therefore usually taken at the Coreper rather than Working Group level (Interviews PermReps 4/2019, 5/2019, 6/2019, 10/2019). The question of mandatory solidarity remained central to the reform package, particularly the Dublin Regulation, which included a corrective allocation mechanism comparable to refugee quotas (Interviews COM 1-3/2019; PermRep 1-11/2019; EP 1/2019; Council 1-3/2019). Several Council Presidencies worked hard on a compromise. The Bulgarian Council Presidency came closest to finding a compromise on Dublin IV, watering down significantly all commitments on solidarity (Council 2018). Still, divisions at the level of the Heads made an agreement impossible. In December 2017, Tusk had made a brave or clumsydepending on whom you askattempt to get the, according to him, 'highly divisive' and 'ineffective' mandatory quotes removed altogether, thereby allowing the Council to proceed with legislative business (Ludlow 2017: 15; Interviews Council 1/2020, 2/2020). But this was not acceptable to the German Chancellor and the Dutch Prime Minister at the time.
The deadlock perpetuated through the June 2018 European Council Conclusions, which reiterated that decisions on Dublin should be taken under consensus (European Council 2018). Many observers interpret this as a requirement for unanimity (Interviews PermReps 1, 5, 6; COM 2, 3; EP 1). The V4 effectively used the venue of the European Council to circumvent the qualified majority rule applied in the Council. On their own, the V4 would have been unable to build a blocking minority in the Council. Towards the end of the Bulgarian Council Presidency, however, the V4 were joined by the Italian Interior Minister Salvini who, for strategic reasons, suggested that the Council should follow the proposal of EP Dublin IV rapporteur Wikstr€ om (Interview PermRep 4/2019). She had suggested abolishing the first country of entry principle altogether and replace it by a 'free choice' or comprehensive relocation principle (European Parliament 2017: 9). The Dublin IV reform thus exemplifies a rather ineffective interplay between the European Council and the Council. After being out of sync on the relocation decision of September 2015, 'normal' legislative decision making was curbed by repeated interventions by the Heads.
With the negotiations on the Dublin Regulation deadlocked, CEAS reform as a whole came to a standstill in spring 2018 (Interviews COM 1-3; PermRep 1-11; EP 2; Council 1-3). While the majority of North-Western Member states, the Baltic states and even some of the Central Eastern European States were ready to adopt a mini-package (consisting of the Asylum Agency, the Resettlement Regulation, and the three directives/regulations aiming at harmonisation), the Southern and V4 Member states stuck to a 'package approach' (Interviews PermReps 4/2019, 5/2019, 2/2019). The Southerners were afraid that the Dublin reform would be easily dropped once they lost the leverage of the Agency, while the V4 were afraid of making any commitments that could backfire, depending on the outcome of the Dublin IV negotiations (Interviews PermReps 6/ 2019, 8/2019). The shadow of the European Council hung heavily over the Council. Protracted European Council involvement had politicised attempts to harmonise asylum systems, thus resulting in complete deadlock (Interviews, PEC Cabinet 1/2020; Council 2/2020; COM 4/2019).

The role of institutional expertise
This section discusses the role of regulatory expertise as a power differential in EU level bargaining, as theorized by LI. After this, we compare the important role of institutional expertise in the Eurozone reform with the limited role in the asylum policy reform.

Institutional expertise in theory: engaging with liberal intergovernmentalism
Moravcsik engages most extensively with the role of institutional expertise in his article on supranational entrepreneurship (Moravcsik 1999), delineating the conditions under which supranational institutional actors can provide leadership in EU negotiations. Their ability to provide leadership stems from 'the persuasive manipulation of information and ideas' (Moravcsik 1999: 268). This implies that there are bottlenecks or information asymmetries which favour certain actors over others. Those with privileged access to information or ideas, be it national representatives or institutional actors, have an advantage in steering subsequent negotiations. Expertise is hence a 'power differential' in EU negotiations. Moravcsik (1998) was able to show that institutional actors did not have such an informational advantage over the member states in previous major (treaty) reform negotiations, except for the Single European Act.
It is important to note that the ideas or information that LI refers to are ideas about the configuration of member states' interests and problems, which would help the Commission to identify potential zones of agreement, for instance on agricultural or competition policy (Moravcsik 1999: 281). The recent major crisis reforms were, instead, about European level problems forcing their way onto national political agendas, for example problems caused by cross-border banking or large inflows of migrants. There is of course still a distributive dimension to these negotiations, for instance whether a particular solution favoured German banks over Greek pensioners, or 'frontline' states versus those less affected by migrant flows. However, these were not negotiations in which member states came together to decide which competences they would want to pool. Moravcsik (1999: 276) posits five models of what institutional leadership could look like, of which two are relevant for assessing the current major crisis reforms. There is little reason and little evidence to support the idea that the EU institutions were able to act as 'honest brokers', sources of legitimisation ('champion for Europe'), or as mobilisers of domestic support ('two-level network manager') during these crises. With the possible exception of the ECB in the early stages of Eurozone crisis, EU institutions were perceived as stakeholders rather than neutral intermediaries serving the common European good. The two models that are potentially applicable in major crisis reforms are 1) the 'great man' type of leadership, which focuses on the Commission advancing bold, imaginative proposals at the political level; and 2) the 'technocratic' type of leadership, which focuses on the ability of lower level institutional experts to formulate substantively and legally competent proposals, that allow the decision making to move forward. We will compare the role of institutional expertise in the Eurozone and asylum crisis, to determine whether the conditions (bottleneck) for institutional leadership role were present and what kind of role supranational institutions were able to play.

Institutional expertise in the Eurozone reform
The crucial precondition for institutional expertise to have an influence is an informational bottleneck. Scarce information amongst principals (the member states) opens a window of opportunity for the agents (EU institutions) to put forward their own ideas and proposals. This precondition is validated for the Eurozone crisis reform process. With the launch of the EFSF, ESM, and banking union, the member states, and the supranational institutions, were entering uncharted territory in which they lacked a clear vision on where EMU deepening was going as well as technical and legal expertise to implement it (Interview Eurogroup 2016/1). The first forum created for dealing with this complexity was the 'Van Rompuy' taskforce on economic governance in 2010. Within this taskforce, the ECB and to a lesser extent the Commission had the clearest ideas about what EMU deepening could look like (Interview PEC 2015/1). Both were keen on setting up a permanent crisis resolution mechanism or stability fund, whereas, the debate between the member states was limited to a strengthening of the existing Stability and Growth Pact (SGP). However, the ideational entrepreneurs, ECB and Commission, were unable to put their mark on the first temporary fund, the European Financial Stability Facility (EFSF). The EFSF resulted from summit diplomacy, at which national representatives provided the ideas and expertise (Gocaj and Meunie 2013;Smeets et al. 2019: 684). Therefore, the EFSF was not a radical innovation, but an extension of the Greek Loan Facility, offering a framework for providing further loans to member states facing problems financing their debt.
It is misleading to portray the European Stability Mechanism (ESM) merely as the permanent successor of the EFSF. Both in terms of the size and the scope, the ESM constitutes the first step towards a European Monetary Fund (EMF) rather than just loan facility. There were some instances of 'great man' type of leadership, specifically by ECB Presidents, Trichet and Draghi, who played an entrepreneurial role at the level of the Heads (see De Rynck 2016;Nielsen and Smeets 2018). ECB President Draghi was also one of the early propagators of the idea of a banking union, thereby filling the ideational void on how to break the 'vicious circle' or 'doom loop' between banks and sovereigns.
Still, it was the technocratic type that turned out to be more important. The impetus and ideas for reforming the Eurozone came from the ECB and the Commission. They provided their input via the nominally intergovernmental fora of the EFC/EWG and the taskforce on coordination action (TFCA), which had replaced the Van Rompuy taskforce. At these technical levels the ECB acted forcefully, pushing for a permanent stability fund to replace the Security Market Programme (SMP) already in the first half of 2010 (ECB 2010; Interviews COM 1/2017, Council 1/2017). The ECB also had the most developed ideas about the concept of a banking union in spring 2012. In both cases, the Commission, primarily at services level, picked up the idea and started running with it. Commission officials at TFCA and EWG level developed the proposals on the size and scope of the ESM, to allow the fund to engage in primary and secondary market interventions, to recapitalise financial institutions instead of just governments, and to provide precautionary loans (Commission 2012). The services level of the Commission also turned the preliminary idea of a supervisory mechanism into a full-fledged roadmap towards a banking union The dominant role of supranational institutions, the Commission and the ECB, in these fora points to the limits of intergovernmental coordination as a mechanism for driving European integration. Member states were not kept out of the loop, but they were not in the driving-seat either. Yet, this 'triumph of technocracy' (Moravcsik 1999: 281) was not the result of smarter of more skilful policymakers at supranational than at the national level, nor the number of experts involved. Technocratic experts mattered, due to their central position in the Eurozone system, which provided them with better opportunities for designing substantively sensible EU-level solutions, fitting within the EMU's legal framework. It was thus the kind of expertise required that favoured the input of supranational actors that were constantly operating at that level. The situation was quite different for the asylum crisis.

Institutional expertise in the reform of asylum policy
Unlike in the Eurozone area, the Commission did not have any informational advantage resulting from a central position in the CEAS reform.
The CEAS is still largely decentralised, with member states' asylum administrations holding most of the legal and practical expertise on the processing of asylum applications, including legal precedence. According to one observer, policymaking in this area therefore is 'a tango led by the Council' (Interview COM 1/2012). Indeed, in the immediate crisis response and the debate around refugee quotas, the Commission drew on policy ideas that were inspired by distribution keys in European federal states, such as the German K€ onigsteiner Schl€ ussel, which also considers population size of a state and its wealth when distributing asylum-seekers.
The European Parliament is usually considered even less of an expert than the Commission, as rapporteurs usually work on several dossiers and tend to be generalists (Interview EP 1/2012; Interview COM 1/2012). The Commission has more manpower working in the area. Still, the Asylum Unit in DG Home is small compared to national asylum administrations, especially in big member states (Interview COM 1/2012). In the past, both the Commission and EP have used the expertise of UNHCR and partly NGOs to fill their gaps in expertise. However, this has often led them to suggest very liberal policies, which were strongly opposed by a majority in the Council.
While the EP and the Commission are usually in favour of Europeanising any policy area, as it strengthens their own role and mandate, member states are generally more hesitant to introduce new policies in the area of asylum, especially if they mean important changes of current policies or increase costs related, for example resulting from higher asylum-seeker recognition rates or better reception conditions (Interview PermRep 1/2013). Consequentially, the Commission and the EP have more to lose vis-a-vis the Council (Interview PermRep 1/2012). As member states are the ones implementing asylum policies, they have greater leverage in EU asylum policy making more generally.
The strong divisions on refugee quotas in September 2015 further weakened the role of the Commission. Normally, if there are two opposing camps, the Commission can try to craft a package deal accommodating the two and 'upgrading the common interest'. However, the situation in the CEAS reform was more complex: Southern border countries were keen to get stronger commitments on solidarity and relocation. North-Western traditional destination countries wanted to get more commitments from border countries that these complied with CEAS law and especially with the Dublin Regulation. The presence of a third camp, the V4, made a deal difficult, as they neither wanted more solidarity nor were they particularly keen on stronger enforcement of the CEAS, for instance, through a European Asylum Agency (EUAA) (Interview PermReps 1-11/ 2019; Interview Council 3/2019). As the Commission had nothing to offer to this third camp, it could not assume the role of honest broker (Interview COM 3/2019). Instead, Commission President Juncker tried to act as a 'strong man' leader, openly siding with North-Western member states that initially wanted refugee quotas and later pushed for a stronger role of the EUAA. This only enhanced the perception of the Commission being a political rather than a neutral actor (Ludlow 2015: 9; Interview PermRep 9/2019).
On the few dossiers that made it into trilogues, the EP took a relatively pragmatic approach, knowing that if they did not largely agree with the Council, no policy would be adopted (Interview EP 2/2019). The positions of NGOs were much less considered than in earlier phases, which had to do with the fact that the reform proposals were drafted under time pressure and with no time for longer consultations or impact assessments (Interview NGO 1/2019; Interview COM 1/2019). The Commission's initial political role undermined its impact later on. As we have seen, only in the process leading up to the EU-Turkey deal could the Commission act as a technocratic leader, working under an intergovernmental umbrella, crafting a package of measures, for financial support (for hosting refugees), visa liberalisation, re-energising accession and, resettlement of refugees from Turkey, and working together with the Council Secretariat and the PEC Cabinet to shepherd the proposal past the member states (Smeets and Beach 2020: 141).

Conclusion
This article makes an empirical and a theoretical contribution. Empirically, we explain the different role and influence of the intergovernmental and supranational EU institutions on the major EU crisis reforms of the past decade, thereby accounting for their diverse policy outputs. Theoretically, we contribute to research suggesting that EU policy making has become more 'intergovernmental' in recent years, by nuancing two central claims of intergovernmentalist theorising: the perceived dominance of intergovernmental actors, especially the European Council, in these reform processes and the supposedly limited role of supranational actors and expertise.
We have provided condensed comparative analyses of six major crisisinduced reforms, the EFSF, ESM, the banking union, the relocation mechanisms, the EU Turkey deal, and the (ongoing) CEAS reform. These reveal important variations in the interplay between the intergovernmental (European Council and Council) and supranational institutions (Commission and EP) and in the balance of power between member states and institutions. The first big difference between the two crises concerns the timing and the type of European Council involvement. During the Eurozone crisis, calling on the European Council was a clever option used at strategic moments in the ESM and banking union negotiations, to push forward the technical work in the Eurogroup/Ecofin Council and in the preparatory bodies. In terms of the type, the involvement of the Heads was predominantly general and procedural: providing general guidelines, reflecting on progress, setting new deadlines. Broadly speaking, the job of the Heads was to show that there was political commitment, thereby providing ministers and civil servants with enough leeway to break technical deadlocks, for instance on the bail-in conditions or the supervisory and resolution mechanisms. During the asylum crisis, on the other hand, European Council involvement was anticipated long beforehand, and it therefore worked more as a brake on the work of the JHA Council. In terms of the type, the Heads were directly involved in the substance of the matter, thereby politicising discussions that would have probably fared better if they had been kept at the technical level.
The second main difference we identified between the two crisis lies in the different kinds of expertise that were required: highly conceptual and technical EU level expertise in the Eurozone crisis, versus legal and operational expertise in the asylum crisis. These types of expertise favoured different holders of this expertise: EU level actors, like the Commission and the ECB, in the Eurozone crisis, versus national actors from line ministries and implementing agencies in the asylum crisis.
These two differences are related. The effective interplay between European Council and Council during the Eurozone crisis was, in no small part, due to an effective division of labour. Eurozone policies required high levels of financial and economic expertise, and there was less of a risk that the Heads would tread on the territory of the finance ministers, when dealing with highly specialised dossiers like capital requirements or steps in a resolution mechanism. Migration policies are complex in a legal sense, as they revolve around legal obligations and (im)possibilities, but they are less complex in terms of the practical implication of certain policies. Hence, the risk of the Heads taking the seat of the JHA ministers and replicating Council level debates was much greater.
Finally, how do we explain these differences and thereby contribute to intergovernmentalist theorising? First, contrary to (some strands of) NI, we note that a more prominent role of the European Council does not necessarily make a decision-making process more intergovernmental. Quite to the contrary, European Council involvement can have a liberating effect, helping line ministers to transcend entrenched positions. It is in fact the Council level negotiations about the CEAS reform that most closely resemble the kind of domestically-determined, interest-based interstate bargaining that we know from classic LI theorising. European Council involvement in the Eurozone reforms did not take the shape of consensus-oriented deliberations, as hypothesized by NI, but the European Council was able to provide political impetus and procedural guidance, thereby spurring on work at the technical level.
Second, contrary to LI theorising, we have shown that there is no theoretical reason to presume that information asymmetries (primarily) work to the advantage of domestic actors. Expertise is a power differential that can work both in the favour of supranational and national actors. Even in a nominally intergovernmental framework, institutional expertise can play an important role, depending on the kind of expertise that is required. LI's theoretical indicators for technocratic leadership, the number and quality (or 'skill') of institutional experts involved, are ill-suited for capturing this role. Due to high levels of centralisation of monetary policies, ECB and Commission experts were better suited to provide technocratic leadership in the Eurozone reforms, while low levels of integration of asylum policies, combined with high levels of political ambition, turned the Commission from a technocratic body into a stakeholder. A comparison of the EU Turkey deal with the faltering CEAS reform, reveals that the legal shape matters less than the substance of the deal. Moravcsik observation that supranational entrepreneurs 'enjoy brief successes and long periods of failure' (Moravcsik 1999: 285) still holds. However, our analyses have shown that ideational entrepreneurship can flourish even in a seemingly inhospitable intergovernmental environment. Note 1. The Sherpa network primarily operates through informal, bilateral or trilateral e-mail and telephone contacts between key Sherpas. Plenary meetings have become less important and less frequent.

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

Interviews and documents
The interviews and documents that were used for the article were part of two large scale research projects by the authors. More information on these research projects and a complete list of the interviews can be found in the online methodological appendix.