Introduction

On 21 July 2020, after 5 days of continuous negotiations, the Heads of States or Governments of the European Union (EU), reached an agreement on a new Multiannual Financial Framework (MFF) and a 750 billion Covid Recovery Fund (named ‘Recovery and Resilience Facility’, RRF). The deal was hailed a historic breakthrough – a Hamiltonian moment even – as it was the first time the member states agreed to a form of debt mutualization through the joint issuance of European debt (Financial Times, 2020; Politico, 2020f; Ludlow, 2020dFootnote ¹). In terms of the size and set‐up this was an unprecedented step in European integration. What was most revolutionary was the principle of borrowing for spending, which allowed the European Commission to go out on the financial markets to acquire debt, guaranteed to be paid back by the member states, and hand it out in the shape of grants instead of only loans (within the context of Community programs). While there are some caveats to calling this the first ‘EU bond’ – the RRF is intended as a temporary, one‐off measure and there is no mutualization of existing national debt – this is nevertheless a ‘quantum leap’ in European integration, both financially and institutionally.Footnote ² In an EU that is supposedly characterized by reluctant member states coming up with ‘too little, too late’ type of solutions (Börzel, Reference Börzel2016; Jones et al., Reference Jones, Kelemen and Meunier2016; Mody, Reference Mody2018), how do we explain such a major step?

The deal was also remarkable in terms of the remarkably short negotiation process. The idea had its first appearance under the label of ‘Corona bonds’, which landed on the table of the Heads at their second video‐tele‐conference of 17 March 2020. However, the idea was dismissed almost immediately by the Northern member states, including Germany, and by Commission President von der Leyen (Politico, 2020a, 2020b). The stage appeared to be set for a lengthy process of intergovernmental bargaining resulting in a lowest common denominator type of solution. Nevertheless, only 4 months later, on 21 July 2020, the EU member states agreed to what was in essence a similar solution: a massive financial support package, to be financed through an unprecedented sharing of financial debts and obligations.

Moreover, while there had been intense negotiation about the overall size of the package, the division between grants and loans, the conditions (if any) that should be attached to financial assistance, the principle as such of the EU entering the financial markets and issuing debt on behalf of the member states sailed through with remarkable ease. As will become clear from our reconstruction, the speed and scope of the developments from March to July also took many insiders by surprise. The Covid crisis had somehow provided the momentum for agreeing on a 1.82 trillion euro financial package that, all things considered, had less to do with addressing the immediate consequences of the Covid crisis, and was more about building what the Commission likes to refer to as, the ‘Next Generation European Union’ (NGEU).

This article goes beyond the headlines to explore the forces at work behind the scenes that produced this unprecedented outcome. We contend that the existing literature has difficulty accounting for the speed and scope of the reform. In the literature on crisis management and policy change, crisis and reforms are not considered a happy couple (Boin et al., Reference Boin, ’t Hart, Stern and Sundelius2017, p. 134). In the classic Kingdon model, three independent streams: the problem stream, policy stream and political stream, need to come together to open a policy window for change (Kingdon, Reference Kingdon1984, p. 21, see also Ackrill et al., Reference Ackrill, Kay and Zahariadis2013). It requires ‘enlightened’ leadership and a very deep crisis to overcome bureaucratically anchored resistance to reform (Kamkhaji & Radaelli, Reference Kamkhaji and Radaelli2017, p. 717). Moreover, this leadership needs to have a sufficiently broad and strong political mandate that authorizes them to drive forward policy reforms (Keeler, Reference Keeler1993, pp. 437–438). Sometimes a crisis can provide such a mandate, but in more occasions, a crisis constraints leaders, forcing them to focus on fire‐fighting (Keeler, Reference Keeler1993, p. 440). In general, crises only provide a very limited window of opportunity to ‘punctuate the equilibrium’ and push through substantive reforms, before pre‐crisis routines kick back in (Jones & Baumgartner, Reference Jones and Baumgartner2005, p. 326).

Also, within the field of EU studies, there are many theories and explanations for why the EU fails to meet crises with reforms, with few approaches able to explain successful reforms. Evaluations of the EU's performance in dealing with these major challenges have been predominantly negative (see e.g., Bastasin, Reference Bastasin2012; Dinan et al., Reference Dinan, Nugent and Paterson2017). Different integration theories – liberal intergovernmentalism, new intergovernmentalism, post functionalism and the failing forward thesis – all provide their own explanation for why the EU fails to deliver on major reforms (Bickerton et al., Reference Bickerton, Hodson and Puetter2015; Hooghe & Marks, Reference Hooghe and Marks2019; Jones et al., Reference Jones, Kelemen and Meunier2016; Moravcsik, Reference Moravcsik2018). Post functionalism highlights domestic contestation as the constraining factor, liberal intergovernmentalism highlights incompatible national preferences. Liberal intergovernmentalism and failing forward both point to a process of distributive intergovernmental bargaining resulting in lowest common denominator (LCD) outcomes. New intergovernmentalism stresses the dominance of intergovernmental actors and action channels, which limits the potential for major policy renewal by supranational policy entrepreneurs.

Leadership is considered a crucial, but at the same time a rather elusive, factor in the literature. It is generally assumed that ‘entrepreneurial’ (Young, Reference Young1991, p. 285) and/or ‘transforming’ (Burns, Reference Burns1978, p. 20) leadership is necessary to drive major policy reforms. Through negotiation skills and process management, such leaders are able to ‘raise’ the motives and ambitions of the led. This is also why such leadership tends to create its own opposition. Major policy reforms target sensitive and salient policy‐domains, for example in the area of fiscal and socioeconomic policymaking, of which established, political and bureaucratic actors are reluctant to diminish their control (Genschel & Jachtenfuchs, Reference Genschel and Jachtenfuchs2014). While some leaders might be pressured, or even personally inclined, to strive for a fundamental overhaul of existing policies and institutional arrangements, in the post‐crisis phase such pressures and inclinations tend to evaporate quickly as other priorities take over, leading to stalled reforms. Some authors even argue ‘that crisis management and reform leadership cannot be the province of the same executive’ (Boin et al., Reference Boin, ’t Hart, Stern and Sundelius2017, p. 139).

Within the field of EU studies, this transforming (‘raising’) leadership effect is generally operationalized as leading to the transfer of competences to the EU level (see e.g., Müller & Van Esch, Reference Müller and Van Esch2020; Tömmel & Verdun, Reference Tömmel and Verdun2017). Yet when crisis reforms involve the transfer of competences in politically sensitive issues, they have to be dealt with by the highest political authorities. In the EU the highest political level is the European Council (EUCO), consisting of the Heads of State or Government of the member states). As a result, major crisis reforms are considered to be ‘Chefsache’, that is, ‘matters for the bosses’. Many EU scholars have argued that, because of a series of major crises, the centre of gravity has shifted systematically towards the Heads, with a primary focus on the German Chancellor (Blyth & Matthijs, Reference Blyth and Matthijs2011; Bocquillon & Dobbels, Reference Bocquillon and Dobbels2014; Bulmer & Paterson, Reference Bulmer and Paterson2013; Puetter, Reference Puetter2014; Van Middelaar, Reference Van Middelaar2019). Initial accounts of the Covid Recovery Fund also gave primacy to German Chancellor, Merkel, and to a somewhat lesser extent French President, Macron, for championing the initiative, primarily through the Franco‐German proposal of 18 May 2020 (see e.g., Genschel & Jachtenfuchs, Reference Genschel and Jachtenfuchs2021, p. 364).

However, as will become clear from our empirical reconstruction, this Franco‐German ‘initiative’ was shaped by Commission proposals that had been developed and discussed within EU institutional cycles and shared with these same member states beforehand. In general, what existing research misses is the increased autonomy and discretion that crises grant to the unseen bureaucratic hands behind the scenes that structure and manage the process on behalf of the leaders, helping them to translate vague ideas into workable solutions. The European Council might be the locus of decision making on major crises reforms, but it is ill‐equipped to come up with substantively meaningful and legally sound solutions fitting within the EU framework (Smeets & Beach, Reference Smeets and Beach2020a, p. 1143). For this it is highly dependent on input from the EU institutions, first and foremost the European Commission, but also the European Council President's Cabinet, the Council Secretariat and the Commission and Council's Legal Services. The relevant question then becomes, who is providing input to (‘feeding’) the leaders?

We contend that major crisis reforms require the involvement of multiple actors (political, institutional) at multiple levels (political, technical) and multiple stages (agenda setting, negotiation, endgame) of the decision making. This article traces the role and influence of the EU institutions in managing major crisis reforms. In the Analytical framework section we present our analytical framework, delineating the main actors, their activities, the levels and stages in a crisis reform process. The third section 3 is the central part of the article, dedicated to the empirical reconstruction of the RRF negotiations. We will use the method of embedded process tracing (EPT) to disentangle the dynamics at different levels of the negotiations about the Covid Recovery Fund. In fourth section we relate our findings to the general literature on crisis reforms and policy change, while also drawing comparisons to past cases of major EU crisis reform. While the literature has been able to identify the main conditions for effective reform, we explain how these conditions play out at the process level, which to a large extent determines their effective contribution to producing major policy reform. In the case of the Covid Recovery Fund, the activities of EU institutional officials significantly shaped the deal and steered process and thus helped to produce a more ambitious outcome than governments left to themselves would have been able to agree on.

Analytical framework: Actors, levels and stages

How to get from crisis to reforms? In this section, we flesh out the main actors, their activities, the levels and stages of the decision making in which these actors are operating, and how these activities jointly affect the outcome. This article uses process tracing methods to reconstruct the causal pathway between the initial cause (major crisis) and the final outcome (ambitious reform) (Gerring, Reference Gerring2017, p. 105).Footnote ³ Process‐tracing is an in‐depth case study method that can empirically assess processual causal claims (Beach & Pedersen, Reference Beach and Pedersen2019). Process tracing typically utilizes an abductive research design, drawing on existing theorizing to identify central concepts that are expected to be causally relevant (actors, levels and stages). The process tracing analyses then seeks to unpack empirically a causal process by tracing the activities of these actors (at various levels and stages) and how these activities are related. In our EPT design, the empirical analysis is embedded in the sense that we go beyond the usual, publicly available sources: official documents and press reports, typically complemented by some on‐off, ex post interviews. Instead, we had regular, in‐depth conversations with insiders from the different levels: European Council, Council Secretariat, Commission and member states both during the negotiations and afterwards. We joined these insiders as they were trying to keep track and make sense of ongoing events and initiatives.

The central actors were already identified in the introduction. Effective reforms require political leadership, to provide impetus and overcome resistance to change, as well as technical or instrumental leadership, to design and broker reform proposals (see e.g., Tallberg, Reference Tallberg2006, p. 37–39). On major EU crisis reforms, political leadership is primarily expected to come from the European Council (Van Middelaar, Reference Van Middelaar2019). Instrumental leadership is generally attributed to institutional actors, the presidents and civil servants in the different EU institutions (see e.g., Becker et al., Reference Becker, Bauer, Connolly and Kassim2016). Through their ‘entrepreneurial’ activities, institutional actors try to shape and steer policy reform processes.

With regard to levels, we use the distinction between the control room and the machine room (Smeets et al., Reference Smeets, Jaschke and Beach2019, p. 677; Van Middelaar, Reference Van Middelaar2019, pp. 197–199). The control room refers to the level that is formally/nominally in charge of the process, which in the case of major crisis reforms is typically the highest political decision‐making level. It is the European Council that needs to acknowledge ownership of the problem and the need for an EU level policy response. Actual European Council involvement might be limited and erratic, as it lacks the institutional infrastructure to manage and oversee policy development and debate. The Heads therefore need to delegate policy design and process management tasks to the institutional machinery. Also, if policy reform primarily takes the shape of legislative change, which can be processed through the ordinary legislative procedures, the appropriate level of political authority (i.e., control room) can also be the respective ministers, united in the Council of Ministers. The machine room refers to the technical and preparatory levels of the negotiations, which prepare, pre‐discuss and potentially reshape the reform proposals, but which lack the authority to provide the formal approval.

Third, with regard to the stages, we build on the distinction between agenda setting, negotiation, endgame stage, known from the policy studies and negotiation literature (see Pollack, Reference Pollack1997; Tallberg, Reference Tallberg2006). The stages are used to structure the analysis, connecting the initial cause – a crisis with a recognized demand for collective action – to the final outcome, that is, major policy reform (see Figure 1).

Figure 1.

Stages in a crisis reform process.

Crises are defined as a situation in which: ‘a social system – a community or organization – experiences an urgent threat to its basic values, which appears to require a far‐reaching response’ (Rosenthal et al., Reference Rosenthal, Charles and ’t Hart1989, p. 10). As we mentioned in the Introduction, crises do not by themselves launch a reform process. Instead, the process is triggered by a shared understanding that the collective level has a responsibility to (‘should’) act. There does not necessarily have to be consensus about the need for collective action. In fact, a reform process can be proceeded by ‘framing contests’ about whether the collective level indeed has a responsibility to act (Boin et al., Reference Boin, ’t Hart and McConnell2009). However, a reform process can only start when there is critical mass to support collective ownership of the problem.Footnote ⁴

We define the outcome major reform as: ‘a policy innovation manifesting an unusually substantial redirection or reinforcement of previous public policy’ (Keeler, Reference Keeler1993, p. 434). Like Keeler, we assess the magnitude of a reform by the degree of change it encompasses, not its impact or success (i.e., whether the reform serves to solve the problem it was intended to solve).Footnote ⁵ To be able to attribute causal leverage to institutional activities, this outcome should be more ambitious/go further than what governments left to themselves would have agreed on. This interpretation of causal influence fits with the classic understanding of ‘entrepreneurial’ or ‘transforming’ (supranational) leadership, which in the EU induces further steps in European integration that transcend LCD outcomes typically produced by intergovernmental decision making (Moravcsik, Reference Moravcsik1999, p. 271). In the empirical reconstruction we will show that other, less radical, reforms were available to governments, such as support through the European Stability Mechanism (ESM) or setting‐up a targeted and limited intergovernmental fund.

Between cause and effect, we identify three stages. In the agenda‐setting stage, the political (control room) level needs to move from acknowledging ownership to providing a mandate for action. Within the context of the EU, what is required is a decision by the European Council, and some form of delegation of responsibility to the machine room to develop solutions and deliver them to the Heads (Van Middelaar, Reference Van Middelaar2019, p. 180). At the technical (machine room) level, institutional actors seek to identify how the EU as a policy instrument can be used to tackle the crisis. Agenda setting primarily revolves around the drafting of texts, some of which may never reach the negotiation tables. The ones that do, however, lay out the tracks for subsequent negotiations.

The negotiating stage starts after the control room level has given the go‐ahead. This is the central stage of the reform process, in which institutional actors and member states representatives discuss and further develop the elements of solutions in the machine room. The main activity here is process management. Major crisis reforms require the involvement of multiple actors at multiple levels of decision making, from technical experts in national line ministries, to EU and national civil servants in working parties, to EU ambassadors (Coreper), to ministers, before going back to the Heads. Consequently, this is a process that cannot be managed by one single actor or institution (Smeets & Beach, Reference Smeets and Beach2020a). The European Council President can provide process management in the control room, but lacks the expertise to formulate solutions and has no presence in the machine room. The Commission is present in both the control and the machine rooms and might have the expertise, but is generally not trusted as the sole independent expert because it has its own substantive agenda. The Council Secretariat has the trust of the member states and the deep knowledge of the functioning of the EU system, but it lacks the capacity that the Commission has for developing substantive proposals. Moreover, while actors in the machine room can develop crisis reforms, they cannot provide the political endorsement. This can only come from the Heads. As said, even though the Heads’ involvement might be limited and superficial, providing their blessing to existing proposals and ongoing processes in the machine room is crucial. It is what keeps the normal bureaucratic actors and action channels – inside the Council and Commission – at bay.

The endgame stage is focused primarily on the control room level, where the Heads deal with the final, unresolved, contentious issues (in EU terminology ‘the brackets’) through classic, intergovernmental bargaining. Institutional actors support the Heads in finding ways to deal with member‐state specific concerns by suggesting legal or substantive solutions, drafting compromise proposals and chairing effective negotiations, either bilateral or multilateral. This final part is crucial for delivering the more ambitious outcome that has been developed in stages one and two.

In the next section, we use these distinctions between actors, levels and stages to reconstruct the process that resulted in the coming about of the Covid Recovery Fund.

Empirical analysis: The covid recovery fund

This section provides an empirical reconstruction of the coming about of the Covid Recovery Fund (RRF), on the basis of the analytical framework presented in the second section. Figure 2 provides an overview of the central actors that were involved in the RRF process. The network also shows the main formal and informal intra‐ and inter‐institutional links that were responsible for producing the deal and shepherding it through the appropriate decision‐making channels.Footnote ⁶ In the analysis, we trace the activities of these actors at the different levels and stages of the decision making. A more detailed explanation of the evidence for each of these activities and their impact on the process and outcome is provided in the methodological appendix.

Figure 2.

Map of the inter‐institutional network for the RRF.

Before we start, we point out two case‐specific particularities of the Covid crisis that are unrelated to the analytical framework. First was the lockdown. What usually happens in crisis situations is that the EU machinery goes into hyper drive. In contrast, in the weekend of 14 March 2020 Brussels came to an abrupt standstill, with much of EU politics transferred back to the capitals. There was the inability to organize physical meetings and informal get‐togethers, which left Coreper as the only body standing. Additionally, both the European Council President and rotating Council Presidency were severely limited in their administrative and coordinating abilities. Both presidencies were highly dependent on input from specific member states and from the European Commission. The Commission, for its part, jumped into the void, coming up with a plethora of initiatives in the first weeks in March, and even setting up a parallel system of ministerial videoconference meetings from the Berlaymont (the Commission's main premises).

The second case‐specific element concerns the role of the European Central Bank (ECB) and the German Constitutional Court (Bundesverfassungsgericht). In earlier crises, the Eurozone crisis in particular, when the Heads had failed to do what (according to some) was required, the ECB would step in to do ‘what‐ever it takes’ to save the Eurozone (see e.g., Bastasin, Reference Bastasin2012). Meanwhile, the Bundesverfassungsgericht would carefully monitor whether the ECB was not overstepping its mandate. In the early stages of the Covid crisis, the ECB intervened heavily, amongst others through its Pandemic Emergency Purchase Program (PEPP) and Pandemic Emergency Long Term Refinancing Operations (PELTRO). However, in a ruling of 5 May 2020, the Bundesverfassungsgericht stated that the 2015 ECB bond buying (Public Sector Purchase) Program would be illegal under German law unless the ECB provided adequate justification within 3 months (Bundesverfassungsgericht, 2020). This ruling raised alarm bells in Germany, and curtailed the role that the ECB itself could play in subsequent stages. One role that the new ECB President, Lagarde, would continue to play, though, was to spur on national leaders and finance ministers to consider radical solutions, such as Corona bonds (e.g., Reuters, 2020).

Conclusion and discussion

In this section we use our findings from the RRF case to (re)engage with the literature on crisis management and major policy change, within and beyond the EU. Our goal is to provide more clarity on the causal pathway that links crises to reforms. As said, most of the individual conditions for effective reform have been identified elsewhere and by others. We seek to identify how they are connected at the process level (when, by whom and at what level) to be able to produce reform. For this, we will also build on comparisons to previous cases of major crisis reforms within the EU.

First, reforms require leadership, political as well as technical/instrumental. In times of crisis, the general focus – of the media but also academia – tends to be on political leaders (Cross, Reference Cross2017; Kreuder‐Sonnen, Reference Kreuder‐Sonnen2018; Lipscy, Reference Lipscy2020). However, the same literature has also shown that these political leaders are limited in their ability to link crises to reforms (Boin et al., Reference Boin, ’t Hart, Stern and Sundelius2017; Keeler, Reference Keeler1993). Our findings echo and extend this line of reasoning. The RRF case revealed some of the difficulties that the Heads faced when trying to provide collective leadership, albeit in a very unusual environment, in which physical meetings were impossible until the very end. The crisis management literature would then expect that the initiative would lie with individual leaders who have the political clout to take charge of crisis management and are able to bring together key actors. This is indeed part of the story behind the RRF, where political level agreement was reached in no small part through bilateral or multilateral meetings between leaders like Merkel, Macron, Sanchez, Conte and Rutte. However, what was lacking at this level was the ability to engineer policy solutions that could provide the basis for leaders’ discussions. Without viable solutions, political level discussions can quickly turn into a shouting matches, as we saw in the early stages of the Covid crisis, where alarmism and insults echoed through the VTC channels. In an attempt to show domestic audiences that they took charge, leaders championed radical ideas (Corona bonds, perpetual debt, Treaty reform) and opted for ‘heroism’ at a time when they needed allies more than ever (see also Goldfinch & ‘t Hart, Reference Goldfinch and ’t Hart2003, p. 240).

To be sure, political leadership is required, but this energy needs to be channelled through appropriate policy ‘streams’ or channels (Kingdon, Reference Kingdon1984). Such channels are not created by the leaders themselves, nor by their domestic policy‐making machinery. The Spanish, Italian, French, Dutch and Franco‐German proposals for dealing with the financial implications of the Covid crisis were political‐level ideas only, and were not anchored to the policy‐making machinery. A second type of (instrumental) leadership is required to provide the linkage between political (control room) and technical (machine room) proceedings. Successful crisis reforms depend therefore on the effective interplay between these levels. From our reconstruction of the RRF case, we extrapolate what this interplay could look like more generally. Figure 3 serves as a heuristic tool that can be used to understand such processes. It does not intend to serve as an exact replica for every reform process, but instead seeks to capture the core dynamics. The remainder of this section discusses the main elements in this model, in the context of the broader literature.

Figure 3.

Process model of the interplay between political and institutional actors in managing major crisis reforms.

Agenda setting is a key condition for major policy change (Jones & Baumgartner, Reference Jones and Baumgartner2005; Pollack, Reference Pollack1997; Princen, Reference Princen2011). Yet, in the literature, agenda setting is generally portrayed as rather long‐term and gradual, leading to slow, but significant, reorientations of a policy course. Agenda‐setting activities tend to take place away from the spotlights, at the lower, technical levels of the decision making. However, agenda setting in a crisis is a much more dynamic and interactive process, and, moreover, in a crisis, political authorities need be actively involved in this. The leaders need to be ‘fed’ with ideas that they can subsequently sell as their own. Within and beyond the EU, there is the persistent misconception that major policy reforms need to be carried by strong, ‘entrepreneurial’, institutional leaders, who present and defend their own reform agenda's to their political counterparts. The image of former Commission President, Jacques Delors, hangs heavily over EU decision making as well as its theorizing (see e.g., Grant, Reference Grant1994; Ross, Reference Ross1995).

The problem with this line of reasoning is that these institutional entrepreneurs, the presidents of the European Commission and European Council, face the same, if not more, limitations as national political leaders when they try to take the reins in crisis reforms. Some recent Commission Presidents, for example, Barroso during the Eurozone crisis and Juncker during the migration crisis, sought to leave their mark on the European project through a fundamental overhaul of the existing EMU and asylum regimes. But their attempt to ‘do a Delors’ quickly ran into member states’ opposition. Agenda setting in a crisis environment is more effective if it does not take the shape of bold initiatives, but instead involves what we term ‘laying out of the tracks’ that national, political leaders are then willing to endorse. Agenda setting also requires that these leaders provide a mandate to further develop some of these tracks.

The sequence and interplay between these agenda setting activities at the two levels can of course vary. It can be that the internal machinery will start to lay out the tracks – that is, diagnose the problem and formulate solutions, fitting within the legal and operational framework of that institution – ahead of any political mandate, like we saw in the RRF case. In other major EU crisis reforms, like the banking union, the control room provided the first spark, after which the machinery sprang into action (see e.g., Nielsen & Smeets Reference Nielsen and Smeets2018). The key point is that agenda setting on major crisis reforms is a shared competence, in which the political level decides the broader direction and the technical level is able to develop and deliver on these ideas.

Process‐management of negotiations is another key condition for major policy change, as is recognized in the policy and negotiation literature (Keeler, Reference Keeler1993, p. 440; Lipscy, Reference Lipscy2020, p. 116). We sought to identify how these activities are related at the process level in producing successful reforms. We know from the literature that crisis is an enabling, rather than active, cause. Crises discredit prevalent policies and ideas, expand executive discretion and loosen the political, legal and bureaucratic constraints that hinder policy change in normal times.

However, this loosening of constraints does not happen automatically. Reform processes need to be actively shielded from national and bureaucratic policy actors. In a crisis situation, everybody typically wants to be part of the process, with the result that many reforms processes run aground (Boin et al., Reference Boin, ’t Hart, Stern and Sundelius2017, p. 133). In order to be successful, networks need to be small, lines need to be kept short and key actors need to be able to find each other across formal institutional divides. This also means that the process needs to be kept rather informal, arguably even ‘unrepresentative’ or ‘undemocratic’. Away from the spotlights, leaders need to know whom to call, or whom to invite to be part of their search for solutions.

Institutional actors are, of course, not the only ones that can come up with solutions and ‘feed’ them to the leaders. In other major crisis reforms, such as the EFSF and the EU Turkey deal, national civil servants played a key, engineering role (Smeets et al., Reference Smeets, Jaschke and Beach2019; Smeets & Beach, Reference Smeets and Beach2020b). However, institutional actors do have a competitive advantage in coming up with legally and operational sound solutions at the collective (EU) level and shepherding these to the policy making machinery. However, as we know from the literature, the effective management of major crisis reforms is not the prerogative of a single actor or institution (Boin et al., Reference Boin, ’t Hart, Stern and Sundelius2017). Instead, it requires collaboration across and a division of labour between various actors and institutions.

In earlier crises (Eurozone and migration crisis), there was mistrust of an overly ambitious Commission that tried to take charge of reform negotiations both at the control and machine room level. This resulted in political leaders hammering out solutions amongst themselves, outside of the Community framework, like the EFSF and EU Turkey deal. In June 2018, the Heads again tried, but failed, to reshape EU migration policy, drafting proposals and forging compromises, resulting in a plethora of measures, ranging from controlled centres to disembarkation platforms (Smeets & Beach, Reference Smeets and Beach2021). In the RRF case, the opposite happened. The Community framework made a comeback. The Covid Recovery Fund was, in fact, a Community product dressed up as an intergovernmental Franco‐German initiative. But this approach is not new. Also, during the Eurozone crisis, we saw this interplay between Franco‐German letters and institutional (Commission) proposals, specifically in the setting up of the ESM and the Fiscal Compact. At the time, the ideas were perceived to be in competition, but in reality they served to propel developments at different levels. Generally speaking, the Franco‐German letters would serve to provide the political backing at the control room level that would allow the institutions to move forward at the machine room level.

In the RRF case, the Commission provided forceful leadership in the machine room, but it did not feature prominently in the political negotiations around the EUCO summits or bilateral or multilateral meetings between leaders. Rather than opting for political activism, President von der Leyen showed strong management of ‘the house’, getting her Commissioners to fall in line, enabling the internal machinery to deliver on an ambitious package (Kassim et al Reference Kassim, Connolly, Dehousse, Rozenberg and Bendjaballah2017). While her management style was perceived as ‘slightly Olympian’, von der Leyen managed to keep many bureaucratic actors within and beyond the Commission at bay. For this approach to work, inter‐institutional cooperation was key. In the RRF case, the Council side – the European Council President, the Council Secretariat and its legal service – came in rather late in the process. However, it played a crucial role in getting the initial endorsement in the days between 20 and 23 April and shepherding the RRF through the final Summit in July. In the endgame process the Commission only played a marginal role. Creative solutions had to be found, primarily by the Council Legal Services and Council Secretariat in direct relations with national representatives, ahead of and at the July Summit.

We end with two broader observations. First, our process‐level analysis offers a more optimistic perspective on how to get from crisis to reforms. If we look beyond the media alarmism and political grandstanding – which seems to become more dramatic with each crisis – we note that the EU policy‐making system is remarkably resilient, and able to deliver in spite of, in fact even helped by, political clashes at the level of the leaders or ministers. Second, but related, we believe that there is reason to qualify the amount of political leadership that is required to get from crisis to reforms (cf. Ludlow, 2020b, p. 77; Van Middelaar Reference Van Middelaar2019, p. 1010). We contend that political leaders need to be brought into a position in which they can exercise guidance and control. High level interventions need to be anchored in the policy‐making machinery to be able to have an impact. If that happens, community‐based solutions can appear remarkably attractive, and agreeing on them remarkably easy, when compared to the vicious intergovernmental bargaining on ESM involvement or Corona bonds. In the Covid case, it resulted in a massive Recovery Fund, which introduced a form of debt mutualization that had been unthinkable as late as early April.