Models of EU Constitutional Reform: What do we learn from the Conference on the Future of Europe?

Abstract

Constitutional reform in the European Union suffers from a post-functionalist dilemma: the options that are politically viable are not democratically legitimate and the options that are democratically legitimate are not politically viable. Against the background of the recent Conference on the Future of Europe and the involvement of transnational European Citizens’ Panels, this article asks whether there is any prospect of overcoming this dilemma and organizing fundamental reform of EU institutions that is both normatively legitimate and politically viable. For this, it examines four models of EU treaty reform and the way these have figured in actual EU reform processes: Intergovernmental Conference, European Convention, informal intergovernmentalism and a Citizens Convention. The article concludes that, as long as the European Union is best characterized as a ‘demoi-cracy’ in which political deliberation takes place primarily in national public spheres, the Intergovernmental Conference remains its main and inevitable forum for constitutional reform. Hence, alternative models of EU constitutional reform should be evaluated not so much on the basis of their potential to substitute the IGC but rather on their ability to catalyse the process and to pre-commit the member state governments.

I. Introduction: The post-functionalist dilemma of EU constitutional reform

Constitutional reform in the European Union suffers from a ‘post-functionalist’ dilemma: the options that are politically viable are not democratically legitimate and the options that would be democratically legitimate are not politically viable. Liesbet Hooghe and Gary Marks (2009) coined the term ‘post-functionalism’ in the mid-2000s, in the aftermath of the failure of the EU Constitutional Treaty. Post-functionalism shares the premise of (neo-)functionalism and most other classical theories of European integration that there is ‘a mismatch between efficiency and the existing structure of authority’ (Hooghe and Marks 2009: 2). In other words, there are pervasive reasons of functionality and efficiency to exercise ever more collective competences at the European level. Where post-functionalism distinguishes itself from previous theories of European integration is in its recognition that, in the era since the 1991 Treaty of Maastricht, ‘[p]olitical conflict makes all the difference’ to whether these functional imperatives are followed up. Moreover, Hooghe and Marks (2009: 2) observe that political entrepreneurs in Europe have successfully mobilized ‘exclusive national identity among mass publics’. This mobilization of nationalist sentiments ‘is likely to raise the heat of debate, narrow the substantive ground of possible agreement and make key actors, including particularly national governments, less willing to compromise’.

Essentially, then, the process of European integration is caught between, on the one hand, functional imperatives that call for greater integration and, on the other, substantial parts of the population that resist it on ideological and identitarian grounds. This condition implies that the functional benefits that further European integration can offer may remain unfulfilled for political reasons. However, it also means that political elites may still try to reap those benefits by downplaying the politicization of European integration and by avoiding popular engagement through, for instance, referenda (Hooghe and Marks 2009).

The post-functionalist dilemma has had no bigger impact than on EU treaty reform. There remains a pervasive sense that the political project of European integration is still fundamentally incomplete and undemocratic (Walker 2023). This is evidenced by the recurrent theme of ‘future of Europe’ debates that are ignited time and again. Examples are the installation by the European Council of the Reflection Group on the future of Europe chaired by former Spanish Prime Minister Felipe Gonzalez in 2007; the 2017 White Paper on the Future of Europe by the European Commission; the European Council’s Sibiu declaration on the Future of Europe in 2019; and, most recently, the inauguration of the Conference on the Future of Europe under the auspices of all three main political institutions of the European Union in 2021. These initiatives do not just reflect the fact that the European Union remains a work in progress; they also imply the recognition that the democratic foundations of EU decision-making remain wanting – that many citizens feel detached from these decisions, and experience them as being taken over them through some kind of obscure international process rather than that they co-own them through a process of carefully calibrated democratic mechanisms.

At the same time, as the European Union has encountered a poly-crisis since 2009 (Zeitlin, Nicoli and Laffan 2019), it has often been apparent that its established competences and mode of operation were not up to the challenges it faced. Adequate responses to the euro crisis, the migration management crisis, the coronavirus crisis and the Russian invasion of Ukraine all required institutional improvisation that regularly stretched the boundaries of what the treaties provided for (de Witte 2021). Notably, critical decisions – such as the European Stability Mechanism, the Fiscal Compact and the Turkey deal – were established outside of the Treaty framework and the constitutional guarantees it provides. Crucially, ever since the agreement on the Treaty of Lisbon in 2007, member states have refused to revisit the EU treaties that essentially embody the constitution of EU decision-making (Fossum 2008; Seubert 2022). Much of this refusal can be explained by the fear that treaty revision would trigger politicization and calls for referenda, which government leaders do not expect to be able to win – at least not at the level of the cross-national majority that would be required for any substantial treaty changes to be ratified.

In this article, I examine whether there is any prospect of overcoming the post-functionalist dilemma and organizing fundamental reform of EU institutions that is both normatively legitimate and politically viable. This examination is inspired in particular by the involvement of randomly selected citizens’ panels in the most recent EU reflection initiative, the Conference on the Future of Europe as it took place in 2021 and 2022 (Conference on the Future of Europe 2022). Various actors – not least the European Parliament (2022a) – have looked at the Conference as a prelude to a formal treaty revision process. At the same time, the involvement of citizens’ panels also opens new ways of thinking about how a treaty revision process may be organized and legitimated. Possibly the use of citizens’ panels can produce both the kind of reform proposals as well as the kind of process that would allow to overcome the public scepticism that has been preventing established political institutions from undertaking any formal constitutional reform of the European Union (cf. Landemore 2020).

In Part II, I outline some premises regarding the specific nature of the EU polity, the challenges this raises for constitutional reform and the specific standard of legitimacy that can be derived from that. The remainder of the article then examines four models of constitutional mediation that are suggested in the context of the European Union: Intergovernmental Conference, European Convention, informal intergovernmentalism and a Citizens Convention based on the citizens’ panels experience. Each of these models strikes its own balance between the two horns of the post-functionalist dilemma: political viability and democratic legitimacy. This balance is partly implied by the internal logic of each model, but it can also be substantiated by actual experiences to the extent that models have been tried in practice. Obviously, these models are ideal-types. In practice, one can think of combinations in which different bodies are merged or different models are used in sequence (as in the current provisions of article 48 TEU in which the conclusions of a European Convention still need to be adopted by an Intergovernmental Conference). Possibly, such combinations can help to secure the merits of different models, but they always come with the risk of complicating the process and its readability by citizens.

II. Challenges for constitutional reform in a demoi-cracy

Constitutional reform faces different legitimacy demands than everyday political decision-making (Ackerman 1991; Siéyès 1982). When it is about the definition of the rules of the ‘political game’, they need to rest on a widespread consensus of the people for a political system to be considered democratic. At the same time, more than any other political issues, constitutional questions ask for a specific form of deliberation. They cannot be settled if everyone simply seeks to maximize their own interests. They can only be settled if people are willing and able to abstract from their specific social positions and world-views and evaluate the constitutional options in terms of the fairness with which they mediate between the diversity of positions present in society (Rawls 1993). Constitutional deliberation is thus inherently premised on the building of an ‘overlapping consensus’ between people(s) with deeply diverse backgrounds.

Such constitutional deliberation faces particular challenges in the European Union because in its present constellation it is best characterized as a demoi-cracy – that is, ‘a Union of peoples, understood both as states and as citizens, who govern together but not as one’ (Nicolaïdis 2013). For constitutional deliberation, this means that, for better or worse, political deliberation remains concentrated primarily at the national level. While there may be an increasing parallelism in the political debates across all member states, certainly when it concerns matters that affect Europe as whole (Koopmans and Erbe 2004; Koopmans and Statham 2010), there is no effective overarching supranational public sphere in which EU citizens directly engage with each other across borders (Habermas 1995).

Under these conditions, the forging of the overlapping consensus required for constitutional deliberation in the European Union is best understood as a ‘two-level game’ (Savage and Weale 2009, drawing on Putnam 1988): it involves national debates, the outcomes of which then serve as input into the negotiations between political representatives at the EU level. In turn, at the EU level, the political representatives receive the views from their counterparts from other member states and these views can then be reinserted into their national debate. Ideally, these two-level deliberations converge to an equilibrium in which the debates in the different member states come to respect and incorporate the concerns of the others and are willing to sign up to common solutions that abstract from their own specific interests and world-views (Crum 2012: Ch. 2). Obviously, however, with a great many different positions and debates, it can be hard to find solutions that are acceptable to all. Given that the nation-state remains the primary object of allegiance for most European citizens, deliberative convergence cannot be taken for granted and it may be that the dynamics of national debates turn against a common European solution.

Ultimately, in evaluating the legitimacy of the different models of EU constitutional reform, the question is about the most appropriate form of mediation. Inevitably, any kind of constitutional reform in the European Union must rely on some form of mediation. The issue is which actors can act as credible mediators and what their ability is to communicate with the people at large. The process of constitutional reform can only be democratically legitimate if it is the subject of public deliberation. The critical additional constraint that the European Union, understood as a demoi-cracy, raises is that we cannot assume the constitutional deliberation will take place in one integrated European public sphere; rather, the process has to trigger and facilitate the debates across the different national public spheres. What is more, in the absence of effective transnational deliberation in which disagreements can be discussed and settled, the fundamental nature of constitutional decisions precludes the possibility that a decision can simply be imposed upon a member state and its people. As long as a people cannot have their voice fully heard in the other public spheres, this would be an instance of domination. In short, the standard of democratic legitimacy for the analysis of the four models below involves the combination of, on the one hand, their likelihood to trigger public deliberation across the European Union and, on the other, them precluding any demos being dominated by the others.

III. Intergovernmental Conference

The Intergovernmental Conference (IGC) is the model through which fundamental institutional reform in the European Union has traditionally been processed. The use of an IGC to amend the European treaties was already explicitly provided in Article 236 of the 1957 Treaty of Rome, remnants of which survive in the present-day Article 48 TEU. An IGC is organized around representatives of the governments of the member states. Initial negotiations may start at the level of civil servants. During most of the process, responsibility usually remains with the Ministers of Foreign Affairs. Ultimately, the conference and its outcome are concluded by the Heads of State and Government. Throughout, IGC meetings take place behind closed doors, in line with common practice in international negotiations.

The Treaty of Rome already provided four critical flanking measures, namely that any member state or the Commission can ask for a revision of the treaty; that it is up to the Council to decide on such a request and to initiate an IGC; that the Intergovernmental Conference decides by consensus (‘common accord’); and that any amendments agreed by the IGC only enter into force after having been ratified by all member states according to their respective constitutional requirements. Over time, this procedure has essentially remained the same, even if the role of the Council in agreeing on the initiation of an IGC has been taken over by the European Council and the use of a European Convention has been added by the Treaty of Lisbon. Thus, for now, in some way or another, any treaty reform will have to pass through an IGC.

In principle, the chances for an IGC to reach the required unanimity on proposals for constitutional reform are actually quite small. A defining characteristic of an IGC is the bargaining logic in which the defence of national interests retains priority over the potential exploitation of collective gains (Moravcsik 1998). This bargaining logic is entrenched by the fact that the parties to an IGC are member state representatives and that any agreement will subsequently have to be approved in a national ratification process. Moreover, the stakes of IGCs are known to be high as they involve fundamental issues of power allocation. Furthermore, member state governments are also well aware that, once a concession has been approved on these issues, it will be very difficult to roll it back again (Scharpf 1988). Given that that decisions in the IGC need to be made by ‘common accord’, it only requires one member state to hold back for a decision to fail.

The chances of an IGC reaching an agreement increase when there is significant pressure by external events and when a wide range of issues can be negotiated in parallel so issue-linkages can be exploited by compensating member states’ concessions on one issue with gains on another. In this light, we can understand the series of IGCs that took place in the late 1980s and 1990s. In these decades, the pressure on Europe to integrate markets as a means to retain global competitiveness coincided with fundamental geopolitical changes in its environment and, following the demise of the communist regimes in Central and Eastern Europe, requests for accession to the European Communities. Ultimately, this led to comprehensive treaty reform packages, with the 1991 Treaty of Maastricht the main example, combining the development of new major EU policy competences (such as the single market, the Euro and common migration initiatives) with fundamental institutional and procedural reform (most notably the extension of qualified majority voting in the Council and increasing powers for the European Parliament) and preparations for enlargement of membership. Notably, even these packages almost floundered, as when the Treaty of Maastricht was initially rejected in a referendum in Denmark.

Indeed, if institutional reform has to rely on the internal logic of the IGC itself, probably very little if anything would come from it. That is why most IGCs build on preparatory work that is done with a clear collective objective in mind and with much less direct bargaining logic. For example, most of the reforms incorporated in the Treaty of Amsterdam were precooked by a ‘Reflection Group’ chaired by Spanish Secretary of State for European affairs Carlos Westendorp.

In principle, the legitimacy of IGCs is high. Member state governments usually enjoy a strong popular mandate (stronger than any EU institution can claim) that has been won in widely publicized and hotly contested elections.¹ The rules of the IGC prevent inter-state domination as they prescribe that no reform can be adopted against the will of any member state, as each of them retains a veto. The legitimacy of IGCs is reinforced further by the additional requirement to have all reforms ratified according to the national procedures. Notably, this requirement allows member states to secure additional legitimacy for their consent to EU institutional reform by making ratification subject to a popular referendum and the public debate this is likely to unleash. Overall, then, IGCs have been set up to prevent any member state finding fundamental EU reforms imposed upon it against its declared will.

However, the flipside of warding off any risk of member state domination is that any reform that would be deemed desirable by a great majority of member states or from the perspective of the collective can be held hostage by a single reluctant member state. This is where the IGC method not only becomes rather ineffective but its legitimacy is fundamentally challenged: the way the procedure is set up essentially inhibits many European actors from pursuing collective interests that they may see but know will run into a minority of (or even only one) opposing member states (Closa 2011).

As it is, the IGC as a method for fundamental EU reform seems essentially stuck. This is partly a function of the growth of the number of EU member states; with 27 member states, it is almost unimaginable to secure a broad consensus and to maintain it through the ratification phase as well. This has even become harder now European integration has increasingly become an issue that divides national parties and citizens, with any treaty changes likely to become subject to the unpredictable verdict of a national referendum (Hooghe and Marks 2009). Treaty-ratification referenda are not only likely in member states where this has long been a constitutional practice (such as Denmark and Ireland) but also in member states in which governments face increasing criticism for an all-too-integration-minded EU policy. Notably, in the case of the 2004 EU Constitutional Treaty, ratification referenda had been foreseen in ten of the then 25 member states before the process was suspended because of negative referendum outcomes in the Netherlands and France (Crum 2007).

Since an IGC remains legally an inescapable phase in any formal process of EU treaty reform, any chance of success essentially relies on the ability to build momentum ahead of the IGC and in its aftermath. Momentum ahead of an IGC can be build up by preparatory gatherings that are able to agree to a common package of proposals that all member states can recognize as superior to the status quo. This is exactly what may be aimed at by the establishment of ‘wise men committees’, future-of-Europe debates and, most recently, the Conference on the Future of Europe. In this sense, we can also understand the insertion of having the IGC preceded by a European Convention in the most recent revision of the treaty reform procedure in the Treaty of Lisbon (see Part IV).

Momentum after an IGC is needed to carry any agreed reforms through the ratification procedures. This requires a strong and affirmative sense of popular engagement and urgency that is widely spread throughout all EU member states. This is probably the element that national governments and EU institutions can control least. Even in the series of crises that has hit Europe over the last fifteen years, experiences in member states, and the responses they elicited, have remained widely diverse and failed to provide a common platform for fundamental EU reform (Hutter and Kriesi 2019).

It is clear that, under the present conditions, the IGC model by itself is fundamentally stuck and can no longer be relied upon to yield any fundamental EU reform. Its limited chances to be effective disappeared altogether with the expansion of the number of EU member states and the way the politicization of European integration has been turned turned into a rallying issue for political discontent in many member states. Moreover, the legitimacy that initially derived from the protection that it offers to member states’ autonomy is fundamentally undermined by the fact that by now all reasonable desires to improve European cooperation are essentially held hostage by the most reluctant member states. The bottleneck that the IGC forms for effective and legitimate EU institutional reform can only be overcome if a procedure can be devised that can mobilize a strong and affirmative sense of popular engagement and urgency across all member states of the European Union and that is not only able to pave the way for a consensus among the member states about the reforms needed but can also retain this momentum throughout the national ratification processes.

IV. European Convention

A European Convention, as it is now foreseen in the EU treaties (Art. 48 TEU), is a broad-based body of political representatives of ‘the national Parliaments, of the Heads of State or Government of the Member States, of the European Parliament and of the Commission’. As it was, the 2002/03 European Convention that prepared the EU Constitutional Treaty was composed of 66 members from the then fifteen EU member states and another 39 members of countries that were involved in accession negotiations at the time (excluding later member Croatia but including later non-member Turkey). Specifically, the European Convention involved one representative from the government and two from the parliament of each state, complemented by sixteen members of the European Parliament, two members of the European Commission and the three-headed Convention presidency. As codified in the 2009 Treaty of Lisbon, a European Convention is always to be succeeded by an IGC that is to formally adopt its reform proposals. In that sense, a European Convention is a preparatory body of the IGC. It is to ease IGC decision-making and it may also be seen as contributing to its legitimacy. In fact, the treaty now prescribes the establishment of a Convention as the default procedure to initiate EU treaty reform, a procedure that is only to be deviated from when the envisaged reforms are considered so straightforward that a Convention is considered to be superfluous. In that sense, the EU treaties recognize the IGC(only) model to be exhausted in terms of both its effectiveness and legitimacy.

A Convention is expected to proceed in a markedly different way from an IGC when it comes to the consideration of possible reforms. This is due to a number of characteristics. For a start, a Convention simply involves a greater number of members. The consequence of this is that it really operates as a collective body and that, contrary to an IGC, it is also clear from the start that no individual actor can hold the deliberations hostage on their own preferences. Moreover, the members of a Convention hold qualitatively different mandates: some (like parliamentarians) have a direct electoral mandate, while others are political appointees (like most government representatives and the representatives from the Commission) with an indirect electoral mandate. Some also hold a national mandate (as parliamentarians or government representatives), while others explicitly represent the EU constituency as a whole (the representatives of the European Commission). This encounter between a variety of mandates also has the virtue that, to some extent, detaches the proceedings in the Convention and the focus on constitutional reform from the day-to-day policy discussions in the EU institutions. Interestingly, these different mandates allow for the formation of different kinds of cross-cutting coalitions, sometimes nationality based, sometimes based on level of governance, sometimes institution-based, sometimes party based.

The great variation in terms of political status prevents the use of any kind of voting rule that could reflect in any adequate way the political weight of the Convention members relative to each other. Since it is obviously also impossible for a Convention to reach decisions by absolute unanimity, there is no alternative then to seek to adopt its position on the basis of some sense of a widespread consensus among its members. A further characteristic of a Convention is that, in terms of its proceedings, it tends to work like an assembly, or even a parliament, in that its meetings take place in public and that for parts of its work it may actually split up into committees or sub-sections. The critical upshot of these different features is that a Convention’s proceedings are much less characterized by hard-nosed bargaining than the IGC. Instead, the logic of the Convention setting compels its members to engage in deliberation – that is, to engage with each other on the basis of common principles and to use arguments to win each other over and to find positions that are acceptable to all sides involved (Closa 2003; Magnette and Nicolaïdis 2004).

The combination of multiple lines of representation with public reason-based proceedings gives the Convention-model a distinctive edge over the IGC in terms of legitimacy. It implies that its conclusions are to be assessed less in terms of gains and losses and more on the basis of the arguments exchanged and the reasonableness of the conclusions arrived at. Thus, citizens can track the reasoning of a European Convention as it takes place in public. Indeed, in principle, the public nature of a Convention’s proceedings can invite public debate beyond its confines. In turn, the upshot of these debates can again be incorporated in real time by the Convention members in their proceedings.

In fact, the experience of the 2002/03 European Convention suggests that the actual occurrence of such external media interactions is essential to the legitimacy of the proceedings. In their absence, a Convention may well come to adopt a logic of its own (as a kind of ‘bubble’) and become too detached from the interests of its citizen constituencies (Crum 2012). This explains how the hard-fought consensus on an EU Constitutional Treaty was eventually rejected by a majority of voters in France and the Netherlands. Obviously, there is a tradeoff in this respect because – assuming that different EU constituencies hold competing interests – the more Convention members are held to account by them, the more challenging it will be for them to reach a consensus. However, instead of approaching this as a zero-sum game, such interactions can also to be seen as a process of collective deliberation in which all actors – within and outside the Convention – are to attain greater appreciation of the diversity of interests and values involved (Savage and Weale 2009).

Ultimately, the European Convention model does rely on rather heroic assumptions about the legitimacy that it can achieve for the reform proposals that it proposes. The 2002/03 European Convention failed to redeem these assumptions (Crum 2012: Ch. 9), even if it operated under distinctively beneficial conditions: the number of member states was smaller than today and, as a first-off attempt, the ‘Constitutional Convention’ enjoyed considerable goodwill. The fact that this European Convention ultimately failed to command the support of the citizens and the fact that most of its recommendations only came to be adopted after being reconsidered by a second IGC several years later – which, most notably, focused on playing down all too great ambitions – seems to have left a considerable ‘trauma’ in the mind of many European politicians (Seubert 2022).

Thus, even if the European Convention has now been codified in the treaties as the default method for developing EU constitutional reform, EU governments have avoided to activate it over the last fifteen turbulent years. This reluctance presumably reflects their scepticism about a Convention’s ability to bridge the political divisions that the European Union experiences on key issues today and about its ability to sway the EU peoples behind a set of common reforms. Indeed, given its volume and its public nature, a European Convention may actually risk amplifying societal polarization and hardening opposing positions instead of paving the way for mutual understanding and consensus. Such considerations may account for the fact that member states have rather relied on less-obtrusive methods whenever fundamental institutional reform could not be avoided over the last fifteen years.

V. Informal intergovernmentalism

As the European Union has faced a series of crises since 2009 (the financial crisis, the migration management crisis, Brexit, COVID-19 and Russia’s invasion of Ukraine), it has regularly turned to unprecedented measures to deal with them. Some of these could be activated within the confines of the established treaty structure and the competences and procedures it provides. Others, however, required the European Union to exercise new kinds of powers or to resort to procedures that had not been foreseen in the treaties so far. Even if the EU leaders had the political will to revise the treaties to enable the necessary measures, the urgency of the crisis generally did not allow the time to resort to the formal treaty change procedures as such procedures usually take several years.

Starting with the euro crisis response, a pattern can be discerned by which the crises lead the EU institutions to extend the interpretation of the existing treaty competences and procedures to the maximum and, at the extreme, to resort to measures outside of the treaties. Typically, these measures are adopted under the close control of the EU member states; the European Parliament does not play any role in their adoption and the role of the Commission is mostly an administrative one. When measures are adopted outside of the EU treaty framework, their adoption and execution are also exempted from judicial review and scrutiny by the European Court of Justice. Such decision modes are very much in line with what Bickerton, Hodson and Puetter (2015) have labelled the ‘new intergovernmentalism’ by which member states have brought EU action under their direct control and moved it outside of the codified ‘community framework’ that involves the European Commission and the European Parliament in full. In contrast, the cases that operate on the edges of the treaty have been problematized as a manifestation of ‘emergency Europe’ in which executive actors appeal to crisis conditions to suspend established constitutional principles and procedures (Kreuder-Sonnen 2016; White 2015).

Nowhere have the limits of the EU treaties been more severely tested than in the case of the euro crisis. As it was, in response to the euro crisis in March 2011, the EU member states did use the official simplified treaty revision procedure to amend Article 136 of the Treaty on the Functioning of the European Union to allow for the establishment of a European Stability Mechanism (ESM). Notably, however, a year later the ESM was established as a separate treaty between the nineteen states in the Eurozone and outside the overarching EU framework.² Similarly, that same spring, the EU member states agreed on the EU Fiscal Compact, committing them to strict budget discipline, which was again adopted outside of the existing treaty framework.³ The euro crisis also led the European Union to build a previously unforeseen ‘European Banking Union’. Besides the development of a common regulatory framework, this involved the establishment of new regulatory authorities (the European Banking Authority and the European Securities and Markets Authority); the assignment of a supervisory role to the European Central Bank (ECB); and the establishment of a Single Resolution Fund.

Ultimately, much of the ability of the European Union to handle the euro crisis hinged on the European Central Bank resorting to new measures. The ECB adopted an unprecedented interpretation of its role in keeping member states solvent by introducing the Public Sector Purchase Programme (PSPP) to buy up member state debt. The conformity of these measures with the EU treaties and with national constitutions has been challenged, most notably in Germany. While the European Court of Justice has consistently upheld the compatibility of the ECB’s actions with the treaties, the German Bundesverfassungsgericht (2020) has found that the ECB’s decisions on the PSPP exceeded the EU’s competences (see also German Law Journal 2020).

The limits of the treaties have also been tested in dealing with other crises. Notable examples are the agreement reached by EU governments with the Turkish government in March 2026 to stop the flow of irregular migrants from Turkey to Greece (Carrera, den Hertog and Stefan 2019). A more recent example is the EURI-regulation, which enables the Commission to actively borrow on the financial markets to cover post-pandemic recovery and resilience measures (de Witte 2021).

This series of examples demonstrates that the European Union has been able to undertake quite wide-ranging reforms in response to the major challenges it has faced since the Treaty of Lisbon. New agencies have been established, whole new procedures have been introduced (most notably the possibility of bailing out member states), existing institutions have taken up roles and competences that were previously unforeseen and member states have taken on new commitments outside of the existing treaties. In terms of effectiveness, these measures are generally seen as having been essential for the European Union to contain the impact of the crises, to bring them to an end and to prevent similar crises from occurring in the future.

However, if these measures turned out to be politically viable, they were so only at the cost of a systematic neglect of democratic legitimacy. Given their crisis nature, public deliberation on them is bound to be limited and constrained. Moreover, their authorization relies exclusively on governments’ willingness to grant their consent. However, even if these governments usually carry considerable authority, in many of these cases their accountability has been severely compromised. In fact, in some cases – most notably the assent of the Greek Syriza government to its third bailout package in 2015 after a national referendum had come out against it – one may even question whether their consent was freely given.

Critically, the formal revision procedures in the treaties are there to protect European citizens from being dominated by the unilateral exercise of EU power, to ensure that any revision of EU powers and procedures is subject to checks and balances, and to ensure that it can rely on the broad consent of the people. In bypassing these procedures, member states essentially imposed their measures without allowing for the broad-based negotiations and ratification according to the respective national constitutional procedures. There are good reasons to question whether such bypassing was sufficiently justified by the nature of the crises faced by the European Union (Kreuder-Sonnen 2016; White 2015). Even if the crises required quick action, democracy would demand the formal procedures to be followed at a later stage still to legitimate the measures adopted – especially to the extent that they continue to stay in force after the crisis.

In fact, it is far from certain that many of the measures adopted would have been able to command the broad popular support that fundamental EU reforms are expected to meet. Even measures that can be seen as involving benefits to all EU member states and their citizens (like the Banking Union, the EU–Turkey deal and the EURI-regulation) might still be objectionable to certain (sub-)constituencies – for instance, on humanitarian grounds or because they ultimately do entail a substantial increase in European integration. In other cases, the willingness of certain European peoples to endorse the decisions taken may have been limited because they essentially involve their governments to vouch for (financial) risks that are considered to be more dependent on the choices of other member states than on their own.

While these cases are problematic because it is questionable that, counterfactually, they would have met the requirements of formal treaty reform, the more dramatic cases are those where reform measures have been imposed on member states under conditions that most certainly compromised their preferences. Most notably, this applies for the bail-out arrangements as they were developed step by step in the euro crisis, eventually culminating in the ESM. Ultimately, member states in acute financial trouble had little choice but to accept the procedures and conditions proposed to them. Alternative arrangements, which would have approached the financial problems of euro zone members more as a collective responsibility and less in individualized and contractual terms, were simply not considered (Warren 2018). Similarly, in the case of the Fiscal Compact, the potential risks of political reticence presumably obliged multiple member states to consent even if it was unclear which of their interests it served.

Overall, the European Union certainly has been able to engage in far-reaching institutional reform over the last fifteen years. However, to do so the member states have invoked the crisis conditions to adopt measures that have stretched existing competences and procedures to their very limits (and arguably even beyond them) and bypassed the formal Treaty revision procedures – and, as a consequence, the European people(s).

VI. Citizens Convention

In recent times, transnational EU citizens’ panels have emerged as a new means to enrich and strengthen EU democracy. This interest emerges as part of a much wider trend in the use of deliberative mini-publics in politics around the world (OECD 2020). Citizen panels involve the convening of a randomly selected, socially diverse group of citizens who are invited to establish a common position on a matter of public interest. Importantly, they do so through a series of meetings, in which they are provided with information and, above all, exchange views and opinions among themselves. Usually, the panels ultimately arrive at a relatively broad consensus view that is then documented in a report and serves as a – more or less binding – advice to elected politicians. Classic examples of the use of citizens’ panels come from Canada, New Zealand and some states in the United States (e.g. Warren and Pearce 2008). Over the last decades, also European countries have used them, often at the sub-national or local levels. Prominent recent nationwide examples come from Iceland (Landemore 2020), France (the 2019/20 Citizens Convention on Climate: Giraudet et al. 2021) and Ireland (the 2012-14 Constitutional Convention and the 2016–18 Citizens Assembly: Farrell and Suiter 2019). Notably, in the cases of Iceland and Ireland, the Citizens’ panels were tasked with a mandate to revisit constitutional provisions, while the French convention was also to address an issue of fundamental importance that transcended a simple individual policy proposal. By most indicators, these latter examples have been quite successful, as they have paved the way for genuine policy change in contested domains (same-sex marriage, abortion, climate policy) and provided these measures with considerable legitimacy.

At the EU level, four Citizens’ panels have been held as part of the Conference on the Future of Europe (CoFoE) that took place in 2021–22. The CoFoE is the latest attempt of EU politicians to respond to the lingering sense that the organization of EU decision-making remains deficient democratically (Oleart 2023). It originated in a call to organize a fundamental EU-wide debate by French President Emmanuel Macron (2017), a call that was subsequently adopted by Ursula von der Leyen as she became President of the European Commission in 2019. Despite many other politicians being much less enthusiastic about the idea, the initiative gained a momentum of its own. In May 2021, somewhat delayed by the outbreak of the COVID-19 pandemic, the three main EU political institutions set up a layered structure, in which decentralized debates and four Citizens’ panels were to provide input into a convention-like Conference Assembly. Ultimately, the Assembly was to draft a report with proposals on the future of Europe for the three institutions that had authorized it: the European Parliament, the Council and the European Commission. The use of EU Citizen panels built upon the experience that the European Commission had gained with the Citizens Dialogues it has been holding since 2013 (Costello 2021). In an extension of this, the European Commission has been experimenting with various pilot forms of transnational citizens’ dialogues (Bertelsmann Stiftung 2021; Boucher 2009; Kies and Nanz 2013).

The four CoFoE Citizens’ panels were given mandates on the broad topics of, respectively, ‘Stronger economy, social justice, jobs, education, culture, sport, digital transformation’; ‘EU democracy, values, rights, rule of law, security’; ‘Climate change, environment, health’; and ‘EU in the world, migration’. Each of them proceeded through three sessions of three days. A first, in-person, session took place in Strasbourg and was used to audit selected experts and to identify key issues. A second, online session, was dedicated to translating these topics into concrete proposals. A third, in-person (and partly hybrid) session involved the fine-tuning of the proposals and the eventual adoption (or rejection) of them (Alemanno and Nicolaïdis. 2021). The EU Citizens’ panels were successful in that many of their proposals made it into the report that the Conference Assembly eventually adopted. Indeed, in many respects, the conclusions of EU Citizens’ panels very much came to set the agenda for the CoFoE report (which included 326 measures organized around 49 proposals).

All three EU political institutions have committed to follow up on the CoFoE conclusions (European Parliament 2022a, 2022b; European Commission 2022; European Council 2022). Various of the measures suggested fit with already ongoing policy actions, while on others the EU institutions are committed to taking new initiatives. Ultimately, for all three institutions, there will also be measures that they resist, either because they run counter to their prevalent preferences or because they are considered to be unfeasible or too demanding politically. In the latter regard, measures that require changes to the EU treaties pose a particular challenge. Notably, the European Parliament (2022b) has used the momentum of the CoFoE to ask the European Council to start a treaty revision process by establishing a European Convention. In turn, Commission President von der Leyen agreed in her 2022 State of the Union speech that ‘the moment has arrived for a European Convention’. However, in line with their resistance of the last fifteen years, the response by the member states has been delayed.

Although at this stage it cannot be claimed that the CoFoE and its Citizens’ panels have led to fundamental EU reform (Patberg 2022), the experience with them and the kinds of proposals they have produced suggest that Citizens’ panels may offer a further alternative trajectory to EU constitutional reform. Given the present procedures in the EU treaties, the Citizens’ panels will not be able to replace an IGC, or even a European Convention, in adopting treaty reforms. However, one could envisage a situation in which the conclusions of a European Citizens’ panel, supported by a wave of public engagement and support, would carry so much authority and legitimacy that they would turn the IGC and a European Convention into mere formalities (or even allow the European Convention to be skipped).

So can Citizens’ panels succeed in overcoming divisions and building EU-wide popular momentum for effective EU constitutional reform? In considering this question, the key issue is why a citizens-based constitutional convention might succeed where a politicians-based convention falls short. Essentially, a Citizens Convention would radicalize the democratic advantages that a European Convention has over an IGC: it would be expected to take the openness and the deliberative character to the next level. First, whereas a European Convention involves a wide range of different kinds of politicians, participation in a citizens’ panel is not subject to any substantial selection criterion. All citizens are genuinely treated as full equals in that they have a strictly equal chance of being selected. Further, assisted by stratified sampling, the composition of a citizen panel can be guaranteed to involve a balanced reflection of the full diversity of the polity at large in terms of, for instance, nationality, gender, age, level of education, and wealth.

At the same time, each of the selected members is bound by no other interest and judgment than their own. This feature is expected to facilitate the likelihood that a citizens’ panel is to proceed by pure deliberation. Since the members of citizens’ panels are not tied by parties or any other affiliations, they are unlikely to enter the process with deeply entrenched positions and there are no significant costs attached to them changing their views. Thus, they can openly engage with the questions they are asked and the information that they encounter, and they are free to raise any new idea that may come to mind. For that reason, deliberation in citizens’ panels can be much more open, creative and fruitful than the often predictable and chequered interactions that are typical of deeply entrenched party politics (Dryzek et al. 2019).

Overall, the primary contribution that citizens’ panels can make to the democratic system at large is as ‘generators of ideas’. Citizens’ panels may (well) come up with ideas that are unlikely to gain traction in the party system as it stands. One reason is that they do not suffer from any (party) prejudices. Hence, they may be more open to new ideas and be more restrained than representative institutions in disqualifying certain ideas already before they are even properly considered. On top of inviting new ideas, citizens’ panels may well be able to test the bridging potential of certain ideas and proposals – that is, their ability to offer something to people with different points of view and interests even if they do so in different ways or on different grounds. This potential has particular added value in the context of the European Union, given its lack of an overarching supranational public sphere in which citizens might otherwise engage with each other. Thus, transnational citizens’ panels may offer a simulation of the EU’s missing transnational public sphere.

However, citizens’ panels also come with some structural constraints to their ability to offer new and creative solutions where elected politicians fail. The most fundamental challenge faced by citizens’ panels concerns their external legitimacy: on what basis can we justify that what is agreed among the members of the citizens’ panel should apply to society at large? Critically, elected assemblies have been authorized by the people for whom they decide through elections, and they are accountable for their decisions because they can be voted out of office in the next round. By definition, none of this applies to citizens’ panels. In contrast, the main sources of legitimacy to which citizens’ panels can appeal are procedural and substantive in character: their statistical representativeness, the openness of their establishment, the quality of their deliberations and the ability of their conclusions to command the support of a broad and diverse range of people (Landemore 2020: 106ff). However, it is not self-evident that these considerations are in and of themselves compelling enough for society to adopt the conclusions of a citizens’ panel as collectively binding rules.

In the absence of some act of societal authorization, it is only natural that the conclusions of citizens’ panels tend to be non-binding, and that their eventual adoption remains subject to elected institutions – as, in the case of EU constitutional reform, an IGC and/or a European Convention. That means the adoption of the results of citizens’ panels remains hostage to the very political (electoral) logic that it sought to overcome (Setälä 2017). In the worst-case scenario, when electoral institutions decide to reject (part of) the conclusions of the citizens’ panels for (valid or not) reasons of their own, they create a ‘legitimacy backlash’ as expectations that were projected on the citizens’ panels are undermined (compare situations in which elected institutions reject the outcome of an advisory referendum: Setälä 2011).

Overall, a Citizens Convention risks becoming only one additional step on an ever-longer chain towards EU constitutional reform, a preparatory body to a European Convention that, in turn, is to prepare an IGC. Such an extra step can only be justified if it works as a catalyst and helps to increase the visibility and momentum of the process overall. This all hinges on the external visibility, engagement and support that a Citizens Convention would be able to mobilize in the course of its work. If a Citizens Convention works mostly in isolation and proceeds as good as unnoticed by the media and the wider public, then this will not only limit its external legitimacy but also its leverage over the electoral institutions. To be successful and politically compelling, it is not only the quality of the internal deliberation of the citizens’ panel that is critical but as much the quantity and quality of its interaction with the wider societal debate. It are exactly these conditions that speak to the success of the celebrated cases of citizens’ panels in Iceland and Ireland (Landemore 2020; Farrell and Suiter 2019).

Like those cases, the legitimacy of an EU Citizens Convention would benefit from it being widely publicized affair, if its deliberations would receive widespread media attention and if the reception of its deliberations in the wider society would already have been fed back into its work. Ideally, one would want a Citizens Convention to inspire a wider societal debate and to be able to incorporate the dynamics of that debate directly into its own deliberations and conclusions. Moreover, such a debate would have to engage all EU member states equally and stimulate transnational interactions between them. If the wider public felt it had been actively engaged by the process of the Citizens Convention, then this would be a compelling force for the subsequent European Convention and IGC to adopt its conclusions.

Unfortunately, if there has been one respect in which the CoFoE Citizens’ panels have fallen short, it has been in their ability to elicit media attention, let alone to trigger a genuine transnational debate in the European Union.⁴ We do not know yet why exactly the CoFoE Citizens’ panels attracted so little public interest, but one relevant reason would seem to be that they remained far removed from actual decisions. Some improvements are conceivable in this respect, such as a more precise mandate and greater clarity about the follow-up process from the start. Still, given the advisory character of any EU citizens’ panel and the many demands on EU decision-making, an EU Citizens Convention is bound to remain at some distance from any actual constitutional reforms. Ultimately, then, as appealing as the idea of a Citizens Convention may be in terms of openness and deliberation, it is difficult to claim much for its legitimacy if its work would not resonate with the EU public at large.

VII. Discussion and conclusion

If we consider the different models for EU constitutional reform, the post-functionalist dilemma is clearly apparent. The model that is most viable at present, informal intergovernmentalism, is the least legitimate. The model that is actually provided for by the treaties (and that may therefore be considered most legitimate in a formal sense), the European Convention, is actively resisted by the EU member states. The prospect of a Citizens Convention may appear to be a model that can overcome the post-functionalist dilemma. In terms of viability, there may be (or, at least, may have been) a window of opportunity as most EU political actors cannot deny that there remain reasons to pursue a further completion and democratization of the EU and citizens’ panels appear as an appealing new format to address such issues in a participatory way. At the same time, in terms of openness and deliberation potential, a Citizens Convention may be considered to be even more legitimate than a European Convention of elected politicians.

Still, the implications of a Citizens Convention remain uncertain. In terms of legitimacy, the big question is whether the conditions can be created in which the internal legitimacy of the openness and deliberation of an (EU) citizens’ panel can be converted into a credible external claim to legitimacy that is recognized by the EU peoples at large. Also, in terms of viability, it is uncertain whether there is really a window of opportunity and whether EU member states would allow a Citizens Convention to really define the agenda for EU constitutional reform that they ultimately have to adopt.

If we look at the overview from a longer-term perspective, two trends can be discerned in the way the EU approaches constitutional reform. One is a trend towards informalization, which is reflected in the fact that, ever since 2009, it is the informal intergovernmental model that has prevailed in practice. The other trend is the tendency to expand the scope of actors involved as we have moved from the IGC to the European Convention and now might be considering the option of a Citizens Convention. In a way, these trends run in opposite directions: informalization reduces legitimacy, while the increasing scope of inclusion furthers it. However, from the perspective of the EU peoples at large, it can also be argued that the two trends reinforce each other in diluting the process of EU constitutional reform and the focus on key issues of political contention.

If our concern is about credible mediators and their ability to communicate with the people at large, both identified trends appear undesirable. Rather than being increasingly informalized and including ever more actors, EU constitutional reform would seem to be best served by a procedure that is well formalized and that involves a limited, but well-defined and recognizable, set of actors. In a way, this conclusion reaffirms the primacy of the IGC as the primary site of EU constitutional reform. The IGC is the process on which European integration has relied from its beginning, and for now it remains an indispensable element of any formal treaty revision process. It also clearly features the member state governments as the key actors of the process, who are accountable to clearly delineated constituencies with whom they maintain a privileged line of communication. Indeed, as party-ideological differences remain incoherent and are often suppressed (Marks 2004), no other dimension is more effective in capturing political disagreement in the European Union than the disagreement between the member state governments.

If there is no way to bypass the IGC and if the demoi-cratic understanding of the European Union gives compelling reasons to continue to recognize the member states as the primary masters of the treaties, the focus of any reconsideration of the EU model of constitutional reform should not be so much on alternatives to the IGC, but rather on how additional arrangements are best employed to increase the viability and the legitimacy of the IGC. Here, the example of the 2002/03 European Convention can be instructive. This Convention hit a defining moment mid-way when Ministers for Foreign Affairs (most notably, Joschka Fischer for Germany and Dominque de Villepin for France) stepped in to directly represent their governments in the Convention’s deliberation. At the time, this development was lamented as an infringement of the open and deliberative nature of the European Convention (Pollak and Slominski 2004: 217; Schönlau 2007: 78). Undoubtedly, some of the debates became less consensus oriented and more expressive of underlying disagreements (Crum 2012: Ch. 4). But at the same time, media attention of the Convention increased and, ultimately this was also the moment when the member states became pre-committed to the Convention’s conclusions. Such a development stands in sharp contrast to the way member state governments have kept their distance from the CoFoE and its Citizens’ panels. The European Council’s position was half-hearted at the inauguration of the CoFoE, and it seems committed to the most minimalistic response now it has been concluded (European Council 2022). Thus, in contrast to the European Convention, the CoFoE appears to have rather pushed governments out of any commitment to constitutional reform than to have drawn them in.

Another notable factor that may affect the commitment of the member state governments has to do with a sense of urgency. In the early 2000s, it was felt that the European Convention offered an indispensable opportunity for constitutional reform as the enlargement with ten new member states was impending. The number and size of the challenges that the European Union is currently facing – the Russian war in Ukraine above all – certainly invite no less a sense of urgency. Notably, however, this sense of urgency has not been translated into the case for fundamental EU reform; the fact that the EU has accommodated a series of preceding crises by way of informal adjustments has also probably contributed to this.

The upshot of this analysis can be captured in three theses. First, legitimate EU Treaty change requires the triggering of a truly transnational debate in Europe across (national) public spheres. Second, such a debate appears to be more likely by focusing it on a limited number of well-defined key actors rather than by the tendency to widen the range of actors involved. Third, as long as we recognize the European Union to be a demoi-cracy in which political deliberation takes place in national public spheres first of all, these considerations lead us back to the classical model of the IGC and the member state governments as its main protagonists. Hence, alternative models of EU constitutional reform should not so much be evaluated on their potential to substitute the IGC but rather on their ability to catalyse the process and to pre-commit the member states. They may also offer suggestions for making the IGC procedures themselves more amenable to legitimacy – for instance, by holding the key sessions in public and ensuring that all relevant documents and exchanges are transparent to the EU public(s).

In all, aside from other uses to which citizens’ panels may be put in the EU (Abels et al 2022), constitutional reform and treaty change appear as the less appropriate foci. Similarly, as a European Convention remains a preparatory forum for an IGC, its added value also needs to be evaluated in those relative terms rather than as a self-standing institution. If the European Union remains in need of constitutional reform that is adopted according to well-specified procedures and that can capture the attention of the EU public(s) at large, the focus must be on how EU politicians can commit to an IGC again and how this model can be improved rather than on ways to circumvent it.