Voluntary association, not state consent: why the EU’s stance on secession rests on the wrong concept of legitimacy

ABSTRACT

The EU’s Prodi doctrine stipulates that a new state formed through secession from an EU member state will be treated as a third party vis-à-vis the Union. This article engages with debates on secessionism, self-determination, and democracy in the EU to discuss what the doctrine entails for the democratic legitimacy of the EU. We argue that the doctrine is only compatible with one source of political legitimacy that the EU partly depends on, the state consent model. However, it is not compatible with the voluntary association model which is a moral basis that is increasingly relevant for a supranational union of democratic states. The prevalent practice of organizing referendums on EU accession shows that securing popular support is today an important feature of the politics of legitimacy in the EU. We illustrate our argument with the case of Catalonia and contrast it with cases from the history of European integration.

KEYWORDS:

• European Union
• democracy
• secession
• self-determination
• state consent
• voluntary association

Introduction

How to explain, understand and assess the European Union’s (EU) stance on secessionism in its member states has been subject to intense scholarly debate (Bauböck 2019; Bremberg and Gillespie 2022; Closa 2017; Gounin 2013; Holesch and Jordana 2021; Keating et al. 2019; Kenealy 2014; Massetti 2022; Walker 2017; Weiler 2017). The approach taken by the EU in this context is commonly referred to as the ‘Prodi doctrine’, named after Romano Prodi, former President of the European Commission. The doctrine dates to at least 2004, when Prodi, in a response to a Member of the European Parliament (MEP), stated that:

[w]hen a part of the territory of a Member State ceases to be a part of that state, e.g. because that territory becomes an independent state, the treaties will no longer apply to that territory. In other words, a newly independent region would, by the fact of its independence, become a third country with respect to the Union and the treaties would, from the day of its independence, not apply anymore on its territory. (Prodi quoted in Guidi and Casula 2020, 176)

The gist of the doctrine is that a new state formed through secession from an EU member state becomes a third country with respect to the Union. This means that the new state ends up outside of the EU and its citizens would cease to enjoy the rights they currently possess as EU citizens (e.g. the rights of free movement and residence within the Union). For them to enjoy these rights again, the newly independent state would need to apply for EU membership and go through a protracted accession process. They also risk being denied EU membership since the admission of a new member state requires unanimous consent from existing EU member states, including from the member state it has just left.

The political support for the Prodi doctrine appears rather solid in the EU. In response to initiatives taken by secessionist movements in EU member states, representatives of the European Commission have time and again reasserted the continued validity of the doctrine. For example, at the time of the Scottish independence referendum in 2014, José Manuel Durão Barroso, then President of the European Commission, stated that an independent Scotland would have to re-apply for EU membership (Keating 2017a). The Commission issued a similar statement on the unconstitutional referendum in Catalonia in 2017, adding that any Catalan referendum would need to be held in line with the Spanish Constitution to be recognized by the EU (European Commission 2017). The support for the Prodi doctrine seems to be similarly solid among EU member states, especially in France, Italy and Spain – where secessionism is a particularly sensitive issue. The European Parliament appears more divided. Whereas most MEPs of the largest parliamentary group, the conservative European People’s Party (EEP), seem to support the doctrine, several MEPs from the Greens/European Free Alliance (EFA) have expressed a more understanding attitude towards secessionism in the EU (Reuters 2017). It should also be noted that the Court of Justice of the EU (CJEU) has not, as of yet, had reason to test the legal standing of the doctrine.

This article takes these recent political developments as a starting point for engaging with on-going scholarly debates on secessionism, self-determination and democracy in the EU (see e.g. Arrighi and Stjepanović 2019; Bremberg and Norman 2023; Cheneval, Lavenex, and Schimmelfennig 2015; Coppieters 2010; Fossum 2017; Keating 2017b; Reinikainen 2019). Our aim with the paper is to critically scrutinize the effects of the Prodi doctrine on the democratic legitimacy of the EU. To do so we assess the doctrine in the light of two models of legitimacy in international political theory, namely the ‘state consent model’ and the ‘voluntary association model’ (Christiano 2010; 2012). We make two points relating to these models. The first is that the Prodi doctrine fits with the state consent model. This is the traditional basis for the legitimacy of international organizations. However, as the process of European integration has expanded and deepened, the state consent model has become less effective as a means to legitimize the power and competences of the EU, since the EU has moved towards becoming something more akin to a supranational union of democratic states (Bauböck 2019). The prevalent practice of organizing referendums on EU accession shows that securing popular support, rather than only state consent, is today an important feature of the politics of democratic legitimacy in the Union (Boomgarden and de Vreese 2005; Morel 2007; Oppermann 2013; Tierney 2019). We contend that ensuring broad democratic legitimacy is today equally important for the political legitimacy of the EU, as the need to secure the consent of its member states.

Our second point is that the voluntary association model makes us better equipped to assess the basis for the democratic legitimacy of the present-day EU, and it also clearly points to tensions that the Prodi doctrine gives rise to. The voluntary association model is similar to, but not identical with, the notion of demoicracy in the EU (Cheneval, Lavenex, and Schimmelfennig 2015; Nicolaïdis 2013; Ronzoni 2017), which is why it is important to discuss how these two perspectives on democratic legitimacy in the EU diverge and what it entails for assessing the Prodi doctrine. The main difference is that the voluntary association model proceeds from a more unequivocally individualistic moral foundation. Strictly speaking, it solely grounds the legitimacy of an organization like the EU in the interests of individuals. It is precisely this moral foundation that the Prodi doctrine disregards. The upshot, we argue, is that the doctrine undermines the democratic legitimacy of the EU. The voluntary association model presupposes that people in a secessionist unit in an EU member state should be able to make a free choice as to their future EU membership based on their own evaluation of whether this membership is in their interest or not. But this is exactly what the Prodi doctrine seeks to prevent. As a basis for our assessment, we depart from an understanding of secession ‘as a move to change the status or affiliation of a territory within a wider constellation of polities’ (Bauböck 2019, 229). Methodologically, we follow Rawls and others by aspiring to ‘go back and forth between principles and practices until we reach a reflective equilibrium’ (Bauböck 2019, 236; see also Rawls 1971).

The article proceeds as follows: the next section sketches the conceptions of political legitimacy that we take to be most relevant in an assessment of how the Prodi doctrine affects the EU’s democratic legitimacy. The subsequent section discusses the state consent model and how the Prodi doctrine aligns with this moral basis for the EU. Then we introduce the voluntary association model and discuss the democratic legitimacy of the EU based on this model. This is followed by a section in which we discuss how this model differs from demoicracy, and we also show that organizing popular referendums has become an increasingly important practice to ensure democratic legitimacy in the EU. We then discuss how the Prodi doctrine affects the Union’s democratic legitimacy and illustrate its implications in relation to Catalonia, which is one of the most salient cases of secessionist aspirations in an EU member state in recent times. We also discuss the normative underpinnings of doctrine in contrast to historical examples in which the EU has responded differently to cases of self-determination. The paper ends with a discussion of potential objections against our argument before we present our conclusions.

Two notions of legitimacy in international political theory

The political legitimacy of the EU can be assessed from various points of views (cf. Bellamy 2019; Bellamy and Castiglione 2002; Keating 2023; Lord 2023; Majone 1998; Moravcsik 2002; Scharpf 1999). We focus on two fundamental notions in international political theory that we refer to as ‘judicial legitimacy’ and ‘democratic legitimacy’. These two notions are meant to capture the different types of legitimacy that international organizations need to achieve in order to have the moral authority to impose laws and rules on their member states.

Judicial legitimacy, or legitimacy as legality, is about the legitimacy of the laws as well as of the judicial system (cf. Clark 2009; Føllesdal 2006, 445–446). If laws are legitimate in this sense, then both the laws as such and the agents of the laws will be rightful holders of authority. This means that these agents will be entitled to apply and administer the laws at the same time as the subjects will be obliged to comply, either for content-independent reasons or since they have a moral duty to comply.¹ Føllesdal is not entirely clear on how to assess whether EU laws are legitimate in this sense but he suggests that it is important that the member states ‘have revocably transferred limited parts of their sovereignty by treaty’ (Føllesdal 2006, 445). We believe Føllesdal is on the right track if he means to say that a transfer of sovereignty by treaty is the condition for this type of legitimacy in the EU context. By this we mean that EU laws will be legitimate in this judicial sense if and to the extent that both the laws and the EU’s judicial system are sanctioned by a transfer of sovereignty by treaty from its member states. A transfer of sovereignty by treaty is not only lawful in the sense that it is compatible with the norms of international law (Bodansky and Watson 1992; see also Applbaum 2010). It is also compatible with EU member states conditionally delegating parts of their sovereignty to the EU while retaining the right to take it back. This means that these states also retain the right to reclaim their full sovereignty. The latter right is crucial here as we see it. The consequence is that the member states retain their right to consent – which we take to be the ultimate criterion for this judicial type of legitimacy.

Democratic legitimacy is also about authority; more specifically, the authority of institutions for collective decision-making. Christiano explains that the basic condition for democratic legitimacy is that we can fall back on ‘a fair system of collective decision-making when[ever] there is disagreement on the substance of the decisions and fundamental interests are at stake’ (Christiano 2010, 120). In such situations, democratic legitimacy provides the answers to the question ‘by what right has this decision been made and imposed on all?’ (C hristiano 2010, 120). We follow Christiano in assuming that whether it is morally right to make and impose a decision on all in such a situation depends on the institutions for decision-making being ‘publicly committed to the equal advancement of the interests of the persons who are affected by those institutions’ (Christiano 2010, 121). The underlying requirement for political legitimacy here is, thus, that ‘democratic institutions are legitimate to the extent that and because they publicly realize the equal advancement of the interests of the members of society’ (Christiano 2010, 121). If and when this requirement is fulfilled, we should assume that there is a content-independent reason to comply with a decision. The reason is that the fulfillment of the requirement yields ‘a right to rule in the sense that persons have strong reasons to go along with the decision just because it was democratically made and despite the fact that they might disagree with the content of the decision’ (Christiano 2010, 121–122).

The Prodi doctrine and the state consent model

As already mentioned, the argument of this article is that the Prodi doctrine relies on a conception of legitimacy this has become increasingly inadequate for the present-day EU – the so-called state consent model. In this section we seek to explain the logic of the Prodi doctrine as an expression of state consent-based legitimacy. More specifically, the section clarifies the basic ideas of the state consent model and describes how the doctrine can be reconciled with these ideas. The subsequent section explains why the state consent model fails to suffice as a basis for the legitimacy of the EU and explicates why the doctrine is detrimental for the democratic legitimacy of the Union.

The state consent model is the traditional understanding of the legitimacy of the international system (Christiano 2010; 2012; see also Lobo de Souza 1995, 521). The basic idea of the model is that a state only can be bound by international law or decisions made by international organizations if it has consented to be so. This means that the moral and political authority of the international system is altogether dependent on states and their consent. The state is the only source of legitimacy, and this is the case regardless of how democratic a state is (Christiano 2010, 124–125). What matters for legitimacy is, thus, not consent from a state that qualifies as democratic, but consent from a state that is recognized as a state. Another key assumption is that consenting to be a party to an international treaty constitutes the proper way to consent (Buchanan 2010, 90; see also Bodansky and Watson 1992).² Consenting to be party to a treaty means that a state signs and ratifies a treaty, and, through that concrete act, expresses its willingness to undertake the legal rights and obligations contained in that treaty. A third key assumption is that the consent of a state is valid more or less unconditionally. This means that a state’s consent should be taken at face value and be assumed to bind the state regardless of the state’s bargaining position.³ In the case of international law, a state consenting/signing a treaty should, thus, be assumed to bind the state to respect the content of the treaty even if the act of signature was a result of pressure from more powerful states. In the case of international organizations, consent from a state/its signing of a treaty similarly legitimates the power and competences that the organization then wields over the state regardless of the circumstances for its consent.

In contrast to Christiano, we do not deem it necessary to revise this model in all cases where an international organization wields power and authority over states. Our assumption is instead that traditional state consent normally is sufficient as long as we are dealing with intergovernmental organizations. A case in point is the forerunner of the EU, the European Coal and Steel Community (ECSC). This organization was established through the signing of the treaty of Paris by representatives of six European states in 1951. There was certainly a democratic element in the process of establishing the ECSC in the sense that it was founded by democratic states and that the treaty was ratified by the parliaments in all the would-be member states. However, the democratic element was still quite limited since the decision to join the ECSC was not preceded by broad public debate nor did any of the signatory states hold a referendum on the Paris treaty (Laursen 1996). However, it was quite sufficient to legitimate the establishment of the ECSC since the sovereignty of the member states remained more or less unaffected by their membership in the organization (Rosamond 2018). This changed to a certain extent when the successor of the ECSC, the European Economic Community (EEC), was established in 1957 through the signing of the Treaty of Rome, also without any preceding referendums despite the EEC entailed a larger degree of devolution of sovereignty from the member states.

Once we appreciate the importance that the traditional state consent model attaches to a state signing a treaty, we also grasp the moral logic of the Prodi doctrine. If we look at the EU from the point of view of this model, then the Union appears as a product of the member states’ consent, in the sense of them having signed and ratified various EU treaties. From the same point of view, we also see their signing of these treaties as commitment to undertake the legal rights and obligations contained in the treaties. The conclusion that logically follows from this model is that only states that have signed and ratified EU treaties can claim EU-specific legal rights. A region like Catalonia has never signed and ratified any EU treaties. Consequently, a future, independent Catalonia would, on the moral logic of this model, need to express its ‘consent to be bound’ by the EU treaties in order to come to enjoy the rights and obligations contained in them. By the same token, all existing member states of the EU, Spain included, would need to consent to grant EU membership to Catalonia. This means that the Prodi doctrine, and its explicit requirement that an independent Catalonia would need to apply for EU membership and go through the regular accession process, follows from the normative underpinnings of the state consent model.

The voluntary association model and the democratic legitimacy of the EU

Importantly, the state consent model is not the only model in international political theory of relevance to understand how the Prodi doctrine affects the legitimacy of the EU. Another and at least as relevant theory is the voluntary association model. We propose that the voluntary association model should be understood as a modification of the state consent model. The models are similar in the sense that they both require consent from a state that is about to join an international organization. They both grant states an unconditional right to exit organizations they have joined. The main point of deviation is that the voluntary association model is based on a different view of the process leading up to a decision to join or exit an international organization. On this score, the voluntary association model requires a much more democratic basis for any such decision.⁴ While the state consent model implicitly gives this right to state representatives (e.g. the government or the sovereign), the voluntary association model grants it to the people as citizens of a state.

This does not mean that a state needs to hold a referendum every time it considers joining or exiting an international organization. As we have argued, it is reasonable to see the ECSC as legitimate although it came into being without any referendums in the member states. However, a referendum is critical as soon as the organization that a state is about to join (or exit) has a certain degree of autonomy in its power and competences to pass laws and make decisions. The reason we believe that the latter capacity calls for a referendum is that this will limit the sovereignty of a member state, and that people in the member states, consequently, risk experiencing a deficit of democracy. Failing to secure popular consent would depreciate the overall level of democracy, at the same time as it undermines the organization’s democratic legitimacy.

This is precisely the development that the EU has undergone the last decades. The fact that the present-day EU restricts the sovereignty of its member states is, thus, the main reason for its need for democratic legitimation (Brack, Coman, and Crespy 2019; Schiemann 2007). The Union’s increasing power and competences have also put the spotlight on the need for its institutions to ensure a democratic control of the decision-making processes in the EU (Føllesdal and Hix 2006). However, the voluntary association model should not be seen as a quick solution to the problem of the EU’s democratic legitimacy. What the model does, strictly speaking, is to grant the citizens of a state the right to decide whether their state should be a member of the EU or not. But this is far from an insignificant right, morally speaking. By giving them this right, the model is able to let these citizens decide if the power and competences of the EU should be delegated to the Union in the first place. Furthermore, it also leaves it to them to decide if they accept that the EU exercises these powers. If they decide to join the EU and also refrain from calling for a new referendum on leaving the EU, we should assume that the EU is democratically legitimate despite its democratic deficiencies. It is so since their choice to remain in the EU entails a tacit consent to be members and to conform to the laws and decisions of the EU (Christiano 2010, 136; Pettit 2010, 156).

The evolving politics of legitimacy in the EU

If we look at the EU from the point of view of the voluntary association model, it is striking how much the Union, in practice, has come to rely on this model for its democratic legitimacy. The continuously expanding and deepening integration of the EU has been accompanied by an increasing democratization of the principle of state consent. The trend is that governments in both presumptive and existing EU member states increasingly have chosen to hold referendums where they have left it to the citizens to decide whether the government should consent to their state’s subjection to the power and competences of the EU (Boomgarden and de Vreese 2005; Morel 2007; Oppermann 2013; Tierney 2019). This is particularly pronounced when it comes to EU accession. A study commissioned by the European Parliament asserts that holding membership referendums has become ‘the general norm for legitimising [EU] membership’ (European Parliament 2017, 21). As shown in Table 1, 14 of the 16 new EU member states in the last three decades have held membership referendums. Cyprus and Bulgaria stand out as the only states becoming EU members without a preceding referendum on EU accession since the 1990s.

Table 1. Referendums in the EU on accession and treaty reform.

Year	Country/Region	Issue	Status	Outcome	Turnout
1972	Denmark	Accession EEC	Required	YES (63.3%)	90.1%
1972	Ireland	Accession EEC	Required	YES (83.1%)	70.9%
1972	Norway	Accession EEC	Advisory	NO (46.5%)	79%
1972	France	EEC enlargement	Advisory	YES (68.3%)	60.3%
1975	United Kingdom	Remain in EEC	Advisory	YES (67.2%)	64%
1982	Greenland	Withdrawal from EC	Advisory	YES (53%)	74.9%
1986	Denmark	Single European Act	Advisory	YES (56.2%)	75.4%
1987	Ireland	Single European Act	Required	YES (69.9%)	44%
1992	Denmark	Maastricht Treaty I	Required	NO (49.3%)	83.1%
1992	Ireland	Maastricht Treaty	Required	YES (68.7%)	57%
1992	France	Maastricht Treaty	Advisory	NO (51.1%)	70%
1993	Denmark	Maastricht Treaty II	Advisory	YES (56.7%)	86.5%
1994	Austria	Accession EU	Required	YES (66.6%)	82%
1994	Sweden	Accession EU	Advisory	YES (52.3%)	70%
1994	Finland	Accession EU	Advisory	YES (56.9%)	83%
1994	Norway	Accession EU	Advisory	NO (47.8%)	89%
1994	Åland Islands	Accession EU	Advisory	YES (51.9%)	61.2%
1998	Ireland	Amsterdam Treaty	Required	YES (61.7%)	56%
1998	Denmark	Amsterdam Treaty	Required	YES (55.1%)	76.2%
2001	Ireland	Nice Treaty I	Required	NO (46.1%)	35%
2002	Ireland	Nice Treaty II	Required	YES (62.9%)	49%
2003	Czech Republic	Accession EU	Required	YES (77.3%)	55%
2003	Estonia	Accession EU	Required	YES (66.8%)	64%
2003	Hungary	Accession EU	Required	YES (83.7%)	46%
2003	Latvia	Accession EU	Required	YES (67.0%)	73%
2003	Lithuania	Accession EU	Required	YES (91.1%)	63%
2003	Malta	Accession EU	Advisory	YES (53.6%)	91%
2003	Poland	Accession EU	Required	YES (77.5%)	59%
2003	Romania	Accession EU	Required	YES (89.7%)	56%
2003	Slovakia	Accession EU	Required	YES (92%)	52%
2003	Slovenia	Accession EU	Required	YES (89.6%)	60%
2005	Spain	Constitutional Treaty	Advisory	YES (76.7%)	42%
2005	Luxembourg	Constitutional Treaty	Advisory	YES (56.5%)	89%
2005	France	Constitutional Treaty	Advisory	NO (45.3%)	69%
2005	Netherlands	Constitutional Treaty	Advisory	NO (38.2%)	63%
2008	Ireland	Lisbon Treaty I	Required	NO (46.6%)	53%
2009	Ireland	Lisbon Treaty II	Required	YES (67.1%)	59%
2012	Croatia	Accession EU	Required	YES (66.7%)	43%
2016	United Kingdom	Withdrawal EU	Advisory	YES (51.9%)	72.2%

Note: Referendums held in EU member states on e.g. Euro-membership (Denmark, Sweden), opt-outs (Denmark), bail-out (Greece), EU-Ukraine association agreement (the Netherlands), EU refugee quotas (Hungary) are not included. Source: data compiled from Beach (2018), European Parliament (2017), Lyck (1998), Ålands statistik- och utredningsbyrå (1994).

Moreover, when we consider the ratification of the Lisbon Treaty in 2009, which stipulates a procedure for member state withdrawal from the EU (art. 50 TEU), taken together with the Brexit referendum in 2016 and the subsequent departure of the UK, a similar trend is crystallizing for the exit from the Union. The origin of this trend can be traced back to 1975 and the UK referendum on remaining a member of the EEC. The number of referendums of this kind is too few to suggest that this is an established practice. Nevertheless, the extent to which it is part of an emergent norm of what the democratic legitimacy of the EU rests on is appreciated if we asked ourselves what would happen if a government of an EU member state were to decide to withdraw from the Union without a preceding referendum. The legitimacy of such a decision would most likely be harshly criticized. Furthermore, it would most likely be so regardless of whether it would be legally permissible for the government to make such a decision.

There are several scholars who have discussed the normative entailments of the trend towards an increasing use of referendums to legitimate EU membership (see e.g. Cheneval 2007; Lacey 2017). The most relevant contributions from the point of view of our argument relate to the notion of demoicracy (Bellamy 2019; Cheneval, Lavenex, and Schimmelfennig 2015; Scherz 2018; Wolkenstein 2018). Proponents of demoicracy suggest that the EU is neither a federation based on supranational democracy, nor an intergovernmental arrangement solely based on state consent. It is a ‘Union of peoples who govern together, but not as one [people]’ (Nicolaïdis 2013, 351). To be democratically legitimate, the political authority of such a Union needs to be exercised so that the multiple peoples in the EU can govern themselves together, not as one European people, but within a common European framework. This presupposes that the demoi of Europe recognize that they unavoidably affect one another’s democracies in multiple ways, and that this also generates reciprocal obligations (Ronzoni 2017, 211). Institutionalizing reciprocal obligations would strengthen the demoicratic character of the EU and reduce the EU’s democratic deficit, as well as secure freedom as non-domination for the demoi of Europe (ibid).

The demoicractic model of the EU’s democratic legitimacy is in many ways similar to the idea that voluntary association is the basis of the democratic legitimacy of the present-day EU. However, there is one important point of deviation, at least with the more ‘people-based’ and state-centric understandings of the idea of demoicracy. In some versions of demoicracy, peoples or demoï, rather than individuals, are seen as the primary normative subjects of the EU. As Ronzoni puts it, the ‘democratic equality that must be achieved is that among the different demoï or peoples, rather than individuals, of the EU’ (Ronzoni 2017, 211).⁵ On this score, the voluntary association model proceeds from a much more unambiguously individualistic normative premise, but not in the sense of aiming to realize some form of individual right to consent. The voluntary association model aims to grant each individual in a state a right to vote on the state’s membership of all independent international organizations. The ultimate aim is to give each individual an equal chance to promote her or his interests in this context.

Christiano explains that the voluntary association model aims ‘to allow individuals to choose for themselves what ventures they wish to engage in and to tailor those ventures to their particular interests’ (Christiano 2010, 136). The democratic equality sought for here is, thus, not an equality of peoples or demoï, but ‘equality among persons in the advancement of their interests’ (Christiano 2010, 124). It is true that the voluntary association model also gives states and their peoples a privileged role. However, it does so for strictly instrumental reasons. The reason is that ‘the modern democratic state is an extremely sophisticated system for the identification and advancement of the interests of a very broad proportion of its population’ (Christiano 2010, 124). This means that the ‘moral foundation of the voluntary association model of legitimate international institutions and law [still] consists in […] the interests of persons’ (Christiano 2010, 124).

The voluntary association model has an individual-centered moral foundation which implies far-reaching implications for how we should think of the nature of the EU. One such implication concerns the permanence of the peoples or demoï of the EU. If the democratic legitimacy of the EU ultimately hinges on the interests of individuals, we cannot see the existing demoï in Europe as set in stone once and for all. Instead, the question of whether it is desirable to maintain, say, the Spanish demos or if it should be split – or merged – into several demoï (e.g. Spanish, Catalan, Basque, Galician, etc.) should rather be assessed on the basis of how the Spanish demos is able to advance the interests of the individuals it consists of. Another implication concerns how to view the EU in political terms. Rather than seeing the EU as a Union of peoples or of member states, we should think of EU membership as also encompassing an individual dimension. This model, namely, bestows each citizen in a presumptive and existing EU member state the power to evaluate whether EU membership is in her/his interests, and it grants her/him a right to vote on the state’s EU membership based on this assessment. This mean that the right to associate with the EU not only can be seen as a right pertaining to states. It is also, and ultimately, a right that should pertain to individuals.⁶

The precarious democratic legitimacy of the Prodi doctrine

The fact that the voluntary association model is based on an individualistic moral foundation is also the main reason for arguing that the Prodi doctrine is undermining the democratic legitimacy of the EU. We illustrate the argument here by way of applying the model to the case of Catalonia. We choose to discuss Catalonia in this context since it is one of the most salient cases of a region of an EU member state with secessionist aspirations. The popular support for Catalan independence has increased in recent years and the Catalan government organized an unconstitutional referendum on independence in 2017, which was heavily suppressed by Spanish authorities. It is not necessary to support Catalan independence to understand that the case is relevant for discussing the Prodi doctrine’s view on secession in the EU in relation to the democratic legitimacy of the Union, since although a majority of Catalan citizens do not support independence for Catalonia, an overwhelming majority seems to be in favour of granting Catalans the right to decide on the issue (Bremberg and Gillespie 2022; see also López and Sanjaume-Calvet 2020; Muro and Vlaskamp 2016). Thus, the voluntary association model emphasizes the advancement of the interests of persons, which implies that the EU cannot reasonably deny persons in a hypothetically independent Catalonia the right to freely associate with the EU without this having adverse consequences for the democratic legitimacy of the Union. If the legitimacy of the EU is to remain unaffected by an independent Catalonia being placed outside the EU, it needs to be the citizens of the independent Catalonia who place themselves there.

Similar to previous discussions on secession in the EU as a form of ‘internal enlargement’ or ‘territorial rescaling’ of the Union (see e.g. Bauböck 2019; Keating 2023; Kenealy 2014; Walker 2017), we argue that there is a difference between an entirely new applicant state lacking any history of previous association with the EU, and an applicant state formed through secession from an EU member state. The difference is that people in the latter case already would be members of the EU. They would not only be so in the sense that they live in an EU member state. They would also be EU citizens and, thus, be entitled to freedom of movement and residence in the EU. This means that these people also have a tangible and personal interest in EU membership that people in an entirely new applicant state do not. This fact makes it much less reasonable for the EU to fall back on its usual procedure for accession in the Catalan scenario. It is perfectly reasonable to stick to this procedure in the case of a new applicant state. However, doing the same in the Catalan type of cases would mean nothing short of an ultimatum. What it means in these cases is that people in the secessionist unit would need to choose between either becoming independent or retaining the rights that they currently possess as EU citizens. To be sure, this is precisely what the Prodi doctrine is meant to signal to all people in parts of EU member states with secessionist movements.

From the point of the voluntary association model, however, any such ultimatum is deeply problematic. The problem is that it is difficult to square with the basic requirements of this model. The voluntary association model presupposes that individuals should be able to choose between international organizations based on their evaluation of whether membership in a given international organization is in their interest or not. This is obviously something completely different from setting ultimatums pressing people to relinquish their membership of an organization. You are not free to choose if you are pressed in some direction. For the EU to live up to the standards of a voluntary political association also vis-à-vis people in Catalonia, it would need to let them make an unconstrained choice about their future EU membership, at least as a first step in a negotiating process.

The history of European integration provides examples as to how the EU could act in the type of cases that we are discussing here. What the Union could do is to grant people in secessionist units in EU member states the same right that it once granted the inhabitants of Greenland and the Åland Islands. In the case of Greenland, the EEC did so when the Greenlanders held a regional referendum on their continued EEC membership in 1982. Furthermore, the EEC also proved that this flexibility was not a façade when it turned out that the majority of the Greenlanders voted for leave (see Table 1). In that situation, Greenland left the EEC despite the rest of Denmark (of which Greenland still forms an autonomous region) continued being a member (European Parliament 2017). The EU showed a similar type of flexibility about a decade later, when the Åland Islands (an autonomous region in Finland) held a referendum on a slightly adjusted treaty in connection with the Finnish referendum on EU accession in 1994 (ibid). This solution was perhaps even more remarkable in the sense that it suggested that the Åland Islands could have joined the EU on their own in case of a yes-vote there and a no-vote in the rest of Finland. It also allowed for the possibility that the Åland Islands could have remained outside the EU in case of a no-vote there and a yes-vote in the rest of Finland.

It is possible to object here that a similar solution for a future, independent Catalonia one-sidedly would favour the Catalans at the expense of the Spanish state and its right to consent. The solutions for Greenland and the Åland Islands were sanctioned by the states controlling these regions (i.e. Denmark and Finland), and the issue at stake was rather self-determination rather the secession. This means that those solutions also were compatible with the traditional state consent model. By contrast, opting for the same solution in the case of a future, independent Catalonia would mean that neither the existing state (i.e. Spain) nor any other EU member states would be able to consent. It would be a flagrant violation of the principle of state consent.

However, this is not an objection against our position. Our position is that the EU depends on both state consent and voluntary association for its political legitimacy. It depends on state consent for its judicial legitimacy and on voluntary association for its democratic legitimacy. Furthermore, we also claim that it is possible for the EU to respect both these grounds for its legitimacy and still be much more flexible vis-à-vis a newly independent region than the Prodi doctrine allows for. The EU could do so simply by postponing accession negotiations with a seceding region until the member state controlling the region agrees to a referendum on its independence, and thereby consents to the region’s independence. If the EU would do so, then a treaty on EU accession could be waiting to be approved by a referendum if the people of the region were to opt for secession in their preceding referendum on independence. This solution would not only be compatible with the requirements of voluntary association. It would also respect state consent and, thus, be compatible with the traditional understanding of that model. However, this type of flexibility is exactly what the Prodi doctrine is designed to prevent.

Another possible objection is that we erroneously claim that the Prodi doctrine presents the people in a future, independent Catalonia with an ultimatum entailing that they are pressed to renounce their EU citizenship. According to this reasoning, the doctrine does nothing of the kind. It is the people in this region who would choose to place themselves there. These individuals enjoy EU citizenship by virtue of their citizenship of an EU member state. This is stipulated in the EU treaties (art. 9 TEU and art. 2 TFEU). This means that people in Catalonia must be aware that they will lose their EU citizenship if they leave Spain. If they nevertheless choose to leave Spain, they will also make an active and conscious choice to renounce their EU citizenship.

This objection, however, assumes that the Prodi doctrine is an unavoidable legal necessity, rather than recognizing it as a political stance. It is true that EU citizenship is based on national citizenship in an EU member state. However, it is not true that the EU therefore must be maximally inflexible and rigidly legalistic. There is nothing compelling the EU to uphold the Prodi doctrine, neither in its laws nor in the traditional state consent model. As we have seen, it would instead be perfectly possible for the EU to adopt a more flexible approach and still respect state consent. This means that a more flexible approach also would be judicially legitimate.

Conclusions

Two conclusions are worth underlining based on our assessment. The first is that the European Commission’s pursuit of the Prodi doctrine is a political choice. The doctrine is neither an unavoidable consequence of the principle of state consent nor an inevitable legal fact. The doctrine is compatible with the state consent model, and it may be claimed to express a certain understanding of its implications. However, it is an understanding that draws this model in a maximally inflexible and secession-deterrent direction. The same is true of the doctrine’s relation to existing EU treaties. It can of course be argued that the doctrine is compatible with these legal frameworks but doing so relies on a quite rigid interpretation of the treaties.⁷ Referring to a narrow interpretation of the EU treaties to solve an inherently political problem becomes a circular argument that tends to exacerbate the problem rather than paving the way for a negotiated agreement.

A second and related point is that the doctrine’s character of a political choice means something for how we should understand its consequences in terms of the EU’s democratic legitimacy. What any supporter of the doctrine who also supports the EU needs to understand is that the doctrine is a threat to the EU’s democratic legitimacy despite being compatible with the state consent model. The choice to adhere to the doctrine affects both the kind of political order that the EU promotes as well as the type of moral authority that the EU can invoke. Upholding the doctrine is a choice in favour of ‘Europe of the States’ instead of ‘Europe of the Citizens’. It is the existing member states who benefit from the doctrine, since upholding it makes it very costly to challenge the territorial status quo within EU member states.

It can certainly be argued that a majority of citizens in the EU member states might actually prefer an intergovernmental EU over the voluntary association of democratic states that we advocate (cf. De Vries 2018). That might be true, but it would be bad news for anyone who thinks that the EU should be able to interfere if a member state government starts to dismantle democratic institutions on a national level. An intergovernmental EU means a Europe of sovereign states with more extensive powers to both consent and veto collective decision-making on the EU level. In such European political order, the moral authority of the Commission and the CJEU to uphold the principle of rule of law and protect the rights of all EU citizens, i.e. also in cases like Poland and Hungary, would be significantly diluted (cf. Kelemen 2020). This, we argue, is the wider implication of upholding the Prodi doctrine for the democratic legitimacy for the EU, beyond the more immediate effects on the unfolding secessionist dynamics within EU member states.

Acknowledgements

We would like to thank the editor and two anonymous reviewers at Regional and Federal Studies for critical yet constructive comments. Andreas Føllesdal, Claes Granmar, Ulf Mörkenstam, Jonas Tallberg, and participants at the Global and Regional Governance seminar, Department of Political Science, Stockholm University, as well as the Research Seminar, Swedish Institute of International Affairs, have provided helpful feedback on earlier drafts of this article.

Disclosure statement

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

Additional information

Funding

Funding for this article has been partly provided by the Marianne and Marcus Wallenberg Foundation through the research project ‘National independence and European integration: Scotland’s and Catalonia’s democratic challenges to the political order in the EU (MMW 2017.0027)’.

Notes

1 A content-independent reason to comply is weaker than a political obligation, which we take to entail a moral duty to obey. For political obligations in the latter sense to arise, we assume laws also need to be just (see Buchanan 2004; Stilz 2009; Wellman 2005). If so, an agent of the law would also have a right to be obeyed (Buchanan 2004, 235).

2 As pointed out by Buchanan, the state consent model is also compatible with indirect consent through global governance institutions created by state consent, as well as with a state’s implicit consent to customary international law (Buchanan 2010, 90).

3 Christiano correctly points out that the traditional state consent model is perfectly compatible with a powerful state pressing a weaker state and taking advantage of its weaknesses (Christiano 2010, 125–126).

4 Christiano claims that the voluntary association model presupposes that states are representative of their populations (Christiano 2010, 124–125), including minorities in the state (Christiano 2010, 127). While representativeness is urgent for a number of reasons, we do not take it to be as important as a democratic basis for a decision to join or exit an organization. In order for the voluntary association model to work, the latter is the critical condition as we see it.

5 However, there are theorists who advocate a more republican version of demoicracy that attempts to balance the rights of states/peoples against the rights and duties of individuals as normative subjects (Wolkenstein 2018; Cheneval, Lavenex, and Schimmelfennig 2015; Cheneval 2007).

6 Paraphrasing Gauthier, each citizen in a presumptive and existing EU member state can be thought of as possessing a right ‘to enter into and continue [EU] association with those with whom she wishes to associate […] and to avoid or exit from association with those with whom she wishes not to associate’ (Gauthier 1994, 360).

7 For an explicit example of this kind of legalistic reasoning, see Piris (2017).