Is the European Union a militant democracy? Democratic backsliding and EU disintegration

Abstract

There has been much recent debate over whether the European Union is or should be a ‘militant democratic’ actor in order to respond to democratic backsliding in EU member states. This article argues that the EU is a militant democracy in a specific and limited sense, but that this may be normatively undesirable from a democratic perspective. I first develop a definition of militant democracy that focuses on the militant democratic paradox. I argue that the strongest justifications for militant democracy require that two conditions are met: an ‘existential threat condition’ and a ‘necessity condition’. Next, I analyse four ways in which the European Union has been said to be empowered to act in a militant democratic fashion to combat democratic backsliding in EU member states. I show how some, though not all, of these warrant the label ‘militant democracy’. Moving from the descriptive to the normative analysis, I then consider whether the necessity condition can ever be met since there is always the possibility of non-militant responses through forms of EU disintegration. If we accept this argument, EU actors should prioritize robust non-militant measures where possible while pro-democratic member states should disassociate from frankly autocratic member states where non-militant measures fail.

I. Introduction ¹

Over the past decade, the European Union has been forced to face the fact that, far from being an ‘ever closer Union’, EU member states are on divergent paths regarding fundamental values and constitutional arrangements. One EU member state, Hungary, is now typically no longer classified as a democracy by political scientists, and several more are regressing on liberal democratic indicators. ² As a supranational Union formally committed to fundamental values including democracy and the rule of law, as listed in Article 2, of the Treaty on European Union (TEU), the European Union has come under pressure to respond to democratic backsliding in EU member states to try to act to reverse democratic backsliding and insulate European politics from autocratic actors. Among the types of responses to democratic backsliding are ‘militant democratic’ responses. There has been much recent debate over whether the European Union is or should be a militant democratic actor in order to respond to internal democratic threats such as those posed by authoritarian governments in EU member states. ³ However, the question of whether or how the European Union is a militant democracy has never been answered systematically, and the normative justifications for militancy at the EU level are also under-developed. This article addresses these gaps.

This article asks two core questions. First, is it descriptively accurate to categorize the European Union as a militant democracy? To answer this question requires, first, a theoretically sound articulation about what precisely is meant by militant democracy and, second, an evaluation of the tools the European Union has to respond to anti-democratic threats to see whether they meet the relevant criteria. Using a narrow definition of militant democracy focused on the militant democratic paradox, I argue that the European Union is a militant democracy in the narrow sense that it has tools to respond to anti-democratic threats that themselves challenge democratic fundamentals. The second question I pose concerns the justifiability of militant democratic responses to anti-democratic threats in the European Union. I extrapolate the relevant minimal criteria for the justifiability of militant democratic responses from the theoretical literature and argue that the use of militant democratic tools by EU actors is illegitimate given the voluntarist nature of EU membership. If we accept this argument, EU actors should limit their responses to democratic and rule of law backsliding to robust measures that do not prima facie undermine democratic values.

Jan-Werner Müller notes that those who argue against the use of militant democratic measures by the European Union have generally ‘taken to invoking constitutional identity as a kind of trump card against outside interference’. ⁴ As will become clear, this is far from my view. EU member states have pooled their sovereignty such that if one member becomes autocratic, the democratic character of all others is tainted too. Indeed, as I have argued elsewhere, ⁵ I think EU actors should impose wide-ranging sanctions against member state governments veering towards authoritarianism as well as massively scaling up positive democracy protection measures such as supporting free media, prodemocratic civil society actors, the independence of the judiciary and critical voices in academia. Specifically militant measures are proscribed in my view not because they are ‘too strong’, but because a more thoroughgoing response that does not pose a challenge to democratic values is possible: disassociation from autocratic member states.

The term ‘militant democracy’ has been used in a variety of different ways, but in this article I understand it to refer to those responses to anti-democratic actors that prima facie undermine democratic values. I make the case that this understanding of militant democracy, focused on the ‘paradox’ of militant democratic responses, ⁶ is consistent with both the classical definition and defence of militant democracy developed by Karl Loewenstein and with key contemporary articulations of militant democratic theory. ⁷ The other definitional criterion I use is that militant measures must be suitable to responding to anti-democratic threats and that they must be measures of self-defence – interventions to protect democracy elsewhere raise different normative issues.

The remainder of the article is structured as follows. In Part II, I develop a definition of militant democracy that focuses on the militant democratic paradox and examine the conditions under which moderate defences of militant democracy consider that militant responses are justifiable. In Part III, I map the ways the European Union might be said to act militantly to combat democratic backsliding in EU member states – showing how some do and some do not warrant the label ‘militant democracy’, understood in this way. In Part IV, I examine whether EU institutions are ever warranted in acting militantly against member states, considering that the European Union is a voluntarist association of states. While anti-democratic politics in the European Union can constitute an existential threat to the democratic character of the European polity, I argue that militant democratic responses are proscribed because of the possibility of a robust response to democratic backsliding that does not undermine democratic fundamentals: dissociation from frankly authoritarian member states.

II. A Narrow conception of militant democracy

The conception of militant democracy developed here is centred around the notion of the ‘militant democratic paradox’. That paradox holds that militant democratic measures are different (i.e. paradoxical) from other responses to anti-democratic threats in that they pose prima facie challenges to core democratic values such as civil and political equality. ⁸ Understanding militant democracy in this way is heuristically useful as it identifies a set of responses to anti-democratic actors that raise a very particular normative question: when (if ever) it is justified to act against democratic values in order to protect democracy? Besides being useful to isolate this normative puzzle, this conception of militant democracy also coheres with Karl Loewenstein’s classical exposition of the theory, and with major contemporary contributions to militant democratic theory such as the theories of Alexander Kirshner and Bastiaan Rijpkema. ⁹

This conception I adopt of militant democracy is comparatively narrow – the term is sometimes used more generally to describe democratic self-defence, including through ordinary criminal law or constitutional prohibitions of some kinds of anti-democratic activity. The narrow conception can be distinguished from the broader conception in that narrowly militant democratic responses target the equal participation in democratic politics of specific people or groups of people considered to threaten democratic government. ¹⁰ This actor-focused conception can be contrasted with broader action-focused responses that include, in general terms and via more standard sanctions (fines, prison sentences, etc.), the prohibition of certain actions perceived to threaten democracy. We can think here, for example, of prohibitions on Holocaust denial or of fascist symbols, which are sanctioned with fines and even prison terms in several jurisdictions. The narrow conception of militant democracy is delimited to those types of democratic self-defence that target the equal civil and political rights of actors perceived to threaten democracy. Standard examples here include banning extremist parties and stripping extremists of their right to vote or to stand for election.

This definitional question – whether to define militant democracy in a narrow or broader sense – is separate from the justificatory questions of when and how various types of democratic defence can be justified. Suffice to say, many democratic theorists including, as we will see, most militant democrats, consider that narrowly militant democratic responses warrant a higher threshold of justification than broader, non-militant responses to anti-democratic threats. This is because the values to which democrats are committed ordinarily require equal civil and political rights among members of the democratic polity; deviations from this democratic norm thus require special justification. Further, while some democratic theorists consider narrowly militant democratic responses to be either illegitimate or unwise, most think that democracies are warranted in acting militantly in self-defence, especially in extreme circumstances where the continued existence of the democratic polity is at stake. ¹¹ The challenge is clarifying the precise scope and justification for militant responses.

The narrow sense of militant democracy used here is also typical of the original exposition of militant democratic theory. In the first of Karl Loewenstein’s two founding articles on militant democracy in the 1930s, he argues that when democracies face existential threats from extremists, democrats are ‘at war’. ¹² In such exceptional circumstances, Loewenstein argues, democracies are not bound by the ordinary democratic norms like guaranteeing their opponents equal civil and political rights. Rather, they should ‘no longer restrain from restrictions on democratic fundamentals, for the sake of ultimately preserving these very fundamentals’. ¹³ Like the narrow definition of militant democracy I use here, Loewenstein’s approach thus limits militant responses to those that restrict what are ordinarily fundamental democratic rights. Militant responses are therefore ordinarily impermissible in democratic politics but, arguendo, become justified when they are necessary to stave off an existential threat.

Justifying militant democracy: The existential threat and necessity conditions

Militant democratic responses that target the equal civil and political rights of anti-democratic actors have a higher justificatory threshold than non-militant responses that do not prima facie challenge democratic norms. Without developing a full-fledged theory of the justifiability of militant democratic action, this section argues that justifications of militant democracy are most convincing when two precise conditions are met. First, the threats to democracy must be existential (the existential threat condition). Second, militant action must be necessary to contain this threat (the necessity condition). I show these conditions to be normatively plausible and also show that they cohere with the influential contemporary militant democratic theories proposed by Kirshner and Rijpkema.

The ‘existential threat condition’ captures the intuition that anti-democratic activity that does not pose an existential threat to the democratic character of a polity does not warrant debasing democratic values. When faced with minor threats, democracies should wield the standard tools of criminal and constitutional prohibitions, sanctioning anti-democratic actions with fines or imprisonment rather than targeting the equal civil and political rights of anti-democratic actors. While Kirshner’s influential theory of militant democracy does not explicitly formulate an ‘existential threat condition’, something like it is implied in his ‘principle of limited intervention’, which holds that (narrowly) militant democratic tools like party bans are only permissible ‘to block antidemocrats from violating the rights of others’. ¹⁴ Like the existential threat condition, Kirshner’s principle of limited intervention means the threats cannot be hypothetical – there needs to be a real and significant risk of harm to warrant a militant response. An existential threat, however, has a weightier threshold than rights violation. If democracies are permitted to act militantly every time there is a mere risk of an anti-democrat violating an individual’s rights, we may still be using a sledgehammer to crack a nut: the bluntness of militant responses in such circumstances risks creating more democratic harms than not acting at all. Ordinary rights violations at the individual level can be addressed using ordinary criminal law sanctions, rather than via restrictions on the offender’s political and civil freedoms. ¹⁵ For this reason, the strongest defences of militant democratic actions limit their legitimate use to those militant responses necessary to neutralize the risk of anti-democrats harming the democratic character of the polity. This picks up a suggestion from Bastiaan Rijpkema’s theory of militant democracy as ‘self-correction’. Rijpkema grounds militant democratic interventions in the need to maintain democracy’s ‘unique characteristic … [that] decisions can always be reversed’. ¹⁶ Militant interventions are thus justified, in Rijpkema’s theory, only when ‘the self-corrective capacity of the democracy is threatened’ ¹⁷ – in other words, when the continued existence of democracy is subject to an existential threat.

The ‘necessity condition’ captures the intuition that militant democratic restrictions on the equal civil and political rights of anti-democratic actors are only justifiable if they are necessary to neutralize the threat these actors pose to the democratic order. Militant democratic actions are relevantly necessary if only militant democratic responses could plausibly contain an existential anti-democratic threat. If non-militant responses that do not pose such a prima facie challenge to core democratic values could succeed, then these ought to be preferred to militant responses. As Alexander Kirshner writes, militant democratic action must ‘to the greatest degree possible, [be] consistent with our reasons for embracing self-government’. ¹⁸ Kirshner defends a principle similar to the necessity condition, the ‘principle of democratic responsibility’. This principle focuses on the democratic costs of militant action (their challenge to fundamental democratic values). ¹⁹ Because of these costs, Kirshner argues, militant interventions must be temporally limited and ought to be shaped by the goal of eventually reinstating anti-democrats’ equal political standing. While Kirshner does not formulate the necessity demand explicitly, it is implicit in his insistence that democrats must minimize anti-democratic activity. ²⁰ Furthermore, Kirshner insists that ‘if a society restricts participation when a more democratic response is possible, its efforts will have been self-defeating, inflicting unnecessary damage on its political institutions’. ²¹ It is to minimize such unnecessary damage that the strongest normative defences of militant democracy require a necessity condition.

The existential threat and necessity conditions are jointly necessary to justify militant democratic responses on these exacting standards. So, even supposing only militant democratic action would be effective to contain specific anti-democratic actions, the existential threat condition proscribes militant responses if the threat is not sufficiently serious. In such circumstances, a democratic polity must accept the existence of anti-democrats as a consequence of its commitment to equal civil and political rights for all. Similarly, if the threat to democracy is existential, but if militant democratic action would be ineffective to contain the anti-democratic threat, the necessity condition proscribes militant responses. It is better to remain true to democratic values than debase them with no hope of success. It is also worth reiterating that the existential threat condition and the necessity condition operate in the context of a conception of militant democracy focused on self-defence, the subject of the next section. ²²

III. Is the European Union a militant democracy?

With the narrow conception of militant democracy clearly worked out, we can now ask whether the European Union is a militant democracy in terms of the legal tools it possesses to respond to anti-democratic threats. Note that this question does not interrogate the efficacy of measures used by, or those that could be used, by EU actors to respond to democratic backsliding in EU member states. Nor does it interrogate the justification of EU responses (I address this in the next section). Rather, I focus here on mapping out the various responses legally available to EU actors to respond to anti-democratic threats and on assessing whether the response meets the criteria of the narrow conception of militant democracy. Specifically, to meet the descriptive criterion of being a militant democratic response in the narrow sense, the measure must:

1. 1. be suitable to responding to existential threats to the democratic character of the European Union

2. 2. sanction anti-democratic actors by limiting their equal civil or political rights rather than via ordinary and generalized sanctions proscribing anti-democratic actions. ²³

Given that the task at hand is ascertaining whether the European Union is, descriptively, a militant democracy, I will only consider existing measures that seem to raise the militant democratic paradox. Since militant democratic responses are defined in light of their sanction posing prima facie challenges to fundamental democratic values, EU responses to anti-democratic threats that do not include a sanction (such as the measures for monitoring democratic violations) are excluded from this analysis. Specifically, I focus on four existing tools. First, I look at the possibility of the European Commission or member states bringing ‘systemic’ infringement procedures against member states violating democratic values, including under Articles 258–260 of the Treaty on the Functioning of the EU (TFEU). ²⁴ Second, I examine the recent Rule of Law Conditionality Regulation, ²⁵ adopted in late 2020, which allows the European Commission to propose suspensions or reductions in EU funding to a member state with a ‘generalised deficiency as regards the rule of law’. ²⁶ Third, I analyse a less-known procedure for deregistering European political parties or foundations for breaches of the requirement that they uphold EU fundamental values in their programmes and activities. ²⁷ Finally, I evaluate the Article 7 TEU procedure for sanctioning backslidden governments by stripping their right to vote in the Council. ²⁸

Infringement actions against backsliding member states: Articles 258–260 TFEU

The first kind of measure that can be used to sanction anti-democratic activity is infringement actions brought against an EU member state under Articles 258–260 TFEU. Infringement procedures can be brought against an EU member state by the Commission (Article 258 TFEU) or by another member state or states (Article 259 TFEU). The procedures are the standard way for the European Union to try to resolve situations where a member state violates their obligations under EU law and for the Court of Justice of the European Union (CJEU) to sanction such violations (Article 260 TFEU). Infringement procedures are therefore a very broad sanctioning tool that can be used for all sorts of violations of EU law. Yet, as argued by Kim Lane Scheppele, Dimitry Kochenov and others, these procedures could be used to respond to the autocratization of member states, the gravest threat to democracy in the European Union, especially where regression on democratic norms occurs in a systemic fashion. ²⁹ Sanctions under the infringement procedure can be weighty, as the European Court of Justice can impose both lump sum and daily fines if states fail to comply with the CJEU. Indeed, a recent decision of the CJEU does exactly that; daily fines of €1m were imposed on Poland in October 2021 for failing to suspend the activities of the Disciplinary Chamber of the Polish Supreme Court, ³⁰ a body for sanctioning judges ruled to undermine the rule of law. ³¹

In this way, infringement actions can be used to sanction member states violating democracy and the rule of law (protected by Article 2 TEU). However, infringement actions target the anti-democratic action through ordinary sanctions rather than targeting the equal civil and political rights of the anti-democratic actor. This point is easily misunderstood; clearly, as a matter of legal fact, any infringement action must be brought against the government of a specific member state. Furthermore, when used to address member states that are violating their obligations under EU law to uphold EU fundamental values such as democracy and the rule of law, infringement actions would specifically target anti-democratic actors. So why are they not targeted to actors in the relevant sense? As in the national context, it is possible at the EU level to defend democracy with general rules or norms that bind everyone equally. Invernizzi Accetti and Zuckerman, who also use the narrow sense of militant democracy that I use here, give the example of banning paramilitary uniforms as a non-militant measure of democratic self-defence. Such a ban would apply to all, but could very well be practically geared towards undermining the activities of specific anti-democratic actors. ³² However, as long as this ban is enforced through standard criminal law penalties such as fines and prison terms, this does not raise a prima facie challenge to democratic fundamentals requiring everyone to have equal civil and political rights. Infringement actions against member states for violating their obligations to uphold democracy and the rule of law function in the same way. The EU fundamental values of Article 2 TEU are binding on all member states. When these values are violated, Articles 258–260 TFEU can be used to empower the CJEU to demand rectificatory measures and, failing rectification, to sanction any backsliding member states. Crucially, the sanction the CJEU can implement for failing to rectify a deficiency is a financial penalty (a lump sum or daily fine) imposed on the sanctioned member state (Article 260 TFEU). Such a sanction does not undermine that state’s political participation, and thus does not raise the paradox of militant democracy.

In sum, rather than being a putatively militant measure targeted to the political participation of a backsliding actor, the possibility of using systematic infringement actions to respond to violations of democratic values is rather analogous to an ordinary criminal law provision in the national context. While suitable for responding to anti-democratic activity at the member state level (democratic backsliding), such infringement actions are not militant democratic measures as they are not targeted to the political participation of an anti-democratic actor. Note that this does not count against the use of such infringement actions. Rather, infringement actions seem to raise less-demanding legitimacy concerns than militant democratic responses: their use can be legitimate even if the existential threat and necessity conditions are not met. ³³

The rule of law conditionality regulation

The second measure is the Rule of Law Conditionality Regulation 2020/2092, adopted in late 2020. The term ‘rule of law conditionality’ gives away some of the original ratio legis of this Regulation, which in theory allows the Commission to propose that EU funding to a member state is suspended or reduced if there are ‘breaches of the principles of the rule of law’. ³⁴ Given the sorts of amounts transferred in the EU budget, such a conditionality mechanism would seem to be a powerful response to democratic and rule of law backsliding. However, much of the ambition of the initial Regulation was watered down over the course of negotiating the mechanism. Specifically, the adopted measure only allows the Commission to propose budget conditionality when the integrity of the EU budget itself is at stake, and when the link between the breach of the principle of the rule of law undermines the ‘sound financial management of the Union budget or the protection of the financial interests of the Union in a sufficiently direct way’. ³⁵ In other words, and as emphasized in the disputed European Council Conclusions, it is not enough for an EU member state to systematically undermine democracy and the rule of law; they must be doing so in a matter that directly implicates EU finances – the ‘rule of law conditionality’ in fact amounts to a sanction on fiscal corruption and grift. ³⁶ Regardless of this narrowing down of the Rule of Law Conditionality Regulation, it may still be considered appropriate as a tool to sanction anti-democratic activities in some cases: where a member state government undermines democracy by corrupting the rule of law in such a way as to put the EU budget at risk.

However, as with the possibility of systemic infringement actions, the Regulation does not sanction the equal political participation of the anti-democratic actor (backsliding member state governments) but proscribes a specific illicit action (the corrupt use of EU funds) within an ordinary and generalized sanction (withholding EU funding). ³⁷ However grave the sanction may otherwise be, withholding funds from the EU budget does not target the equal political participation of the anti-democratic actor. Making EU funds conditional on member states maintaining the rule of law in a manner sufficient as to protect the EU budget’s integrity is a general demand made by the regulation of all member states. Further, while a weighty sanction, it does not obviously target core civil or political processes in a way that may clash with the democratic norms of equal civil and political rights. Therefore, the Rule of Law Conditionality Regulation does not raise a prima facie challenge to fundamental democratic values. Again, this is not a critique of such measures – far from it. All the reasons why we may worry, normatively, about militant measures – whether pragmatic or principled – are unlikely to have bite in these cases. ³⁸ Responding to anti-democratic threats via the Regulation, as in the case of systemic infringement actions, does not need to meet the high legitimacy thresholds required of militant democratic responses. If such non-militant measures could be effective, then they clearly ought to be preferred by those committed to democracy, as long as they are used in a manner otherwise coherent with a democratic ethos.

Deregistering anti-democratic European political parties and foundations

The third measure under consideration is the possibility under EU Regulation 1141/2014 for European political parties and foundations to be deregistered. ³⁹ Regulation 1141/2014 generally governs the funding and status of European political parties, but also sets out the legal framework for an Authority for European political parties and European political foundations ‘for the purpose of registering, controlling and imposing sanctions on European political parties and European political foundations’. ⁴⁰

Both the registration and the deregistration roles are interesting. To register a political party or associated political foundation, the Authority must assess whether that party or foundation observes the fundamental values of the European Union formulated in Article 2 TEU ‘in its programme and in its activities’. ⁴¹ However, Article 9(3) states that the Authority shall consider as ‘sufficient’ a mere declaration by the political party or foundation. The procedure for deregistration in Article 10 is more thorough, as the Authority is tasked with verifying that political parties and foundations continue to respect EU fundamental values, and authorizes the Authority to deregister a party or foundation if it is deemed to violate those values. ⁴² Considering that democracy and the rule of law are amongst the fundamental values in Article 2 TEU, it seems that the sanction of deregistering European political parties and foundations under Regulation 1141/2014 could be used to respond to anti-democratic threats in the European Parliament.

The deregistration of political parties and foundations also seems to be actor-focused in the relevant way. While the Regulation does not only target anti-democratic threats – European parties and foundations can also be deregistered on the basis of the party failing to (continue to) meet other more formal conditions of registration – the sanction can be imposed directly on political parties and foundations for undermining Article 2 TEU values in their programmes or activities. The only further substantive demand is that these activities and/or programmatic alignment breach Article 2 TEU values in a manner that is ‘manifest and serious’. ⁴³ Furthermore, the sanction of deregistration has profound effects undermining political parties’ and foundations’ equal political rights, as deregistration would mean that the European political party or foundation is not entitled to a proportionally equal share of European funding. The scale of this restriction should not be underestimated, as registration under the Regulation allows European political parties to claim up to 90 per cent of their expenses from the EU budget and foundations to claim up to 95 per cent (Article 19(3)). Excluding a European political party or foundation from these sorts of budget contributions would clearly undermine the equal democratic playing field vis-à-vis other parties and foundations that are financed from the EU budget.

While sympathetic to the legitimacy of the Regulation, arguing that ‘the deregistering of European political parties by no means entails an absolute “exclusion from the democratic game”’, Fabio Wolkenstein concedes that the Regulation is a militant democratic measure and that the sanction ‘would mean a big blow to any European party … reducing its capacity to make its voice heard in the European Parliament’. ⁴⁴ Ludvig Norman, in an article tracing the development of the Regulation, agrees with this assessment, arguing that the Regulation ‘demonstrates a clear shift from a democratizing logic to a more protective one’, ⁴⁵ a logic he defines by commitment to the militant democratic principle that ‘democracy must be protected, even at the cost of sometimes infringing on democratic principles’. ⁴⁶ We can therefore conclude that deregistration of an anti-democratic political party or foundation under Regulation 1141/2014 constitutes a militant democratic sanction. Under the stringent conditions developed above, that means this Regulation can only be justifiably used if putatively sanctioned political parties and foundations form an existential threat to the democratic character of the European Union, and if deregistration would be necessary to contain this threat.

Disenfranchising backsliding member states in the Council: Article 7 TEU

The last measure under consideration is the well-known mechanism of Article 7 TEU. This article provides a mechanism to sanction member states violating EU fundamental values listed under Article 2 TEU by, inter alia, stripping them of their right to vote in the Council of the European Union. ⁴⁷ Much has already been written on Article 7 TEU, so we can be brief. For our purposes, we must focus initially on two things: its suitability for responding to anti-democratic threats; and whether it does so by targeting anti-democratic actors’ equal political participation. In each case, the mechanism clearly does meet the criteria for militancy.

Much more clearly than the Rule of Law Conditionality Regulation or the possibility of using systemic infringement actions, the ratio juris of the Article 7 TEU procedure clearly lies in responding to violations of EU fundamental values including violations of the value of democracy through democratic backsliding by member state governments. Article 7 TEU is intended to do nothing more than respond to such violations. The sanction detailed in Article 7(3) also clearly targets the equal political participation of anti-democratic actors. Disenfranchising a member state government from voting in one of the European Union’s primary legislative bodies raises thorny normative questions about the democratic legitimacy of continuing to hold such a sanctioned member state subject to legislation co-decided in the Council (as per Article 7(3)). ⁴⁸ A standard demand of democratic legitimacy is that those subject to putatively democratic rules have a proportionally equal stake in making those rules. Whether or not one holds that anything of democratic value would be lost were an anti-democratic member state government to lose its right to vote in the Council, clearly the citizens of that member state would not enjoy equal representation in the Council once such a sanction was imposed. ⁴⁹ They would be subjected to EU legislation without having an equal stake in co-authoring (or co-authorizing) that legislation. This analysis of Article 7 is not, in my view, vulnerable to the objection that member states signed up to the European Union voluntarily, and therefore gave their consent to be bound to the Article 7 procedure. The claim that it is anti-democratic to hold a state subject to laws without an equal stake in co-authoring those laws does not turn on such consent – much as disenfranchising a group of citizens at the national level would not be legitimate if a referendum were held and received a majority of votes. ⁵⁰

We can therefore conclude that, like Regulation 1141/2014, Article 7 TEU is a militant democratic measure. ⁵¹ It is suitable for responding to anti-democratic threats ⁵² and targeted to anti-democratic actors’ political participation. Consequently, Article 7 raises the militant democratic paradox, so is legitimate only if it meets the high thresholds given by the existential threat and necessity conditions.

Would a militant European Union be acting in self-defence?

We have identified several measures that EU actors can take vis-à-vis antidemocratic challengers that raise the paradox of militant democracy. However, before we can conclude that the European Union is a militant democracy we must consider one final question, namely whether the European Union is acting in democratic self-defence. As a supranational union of countries pooling some of their sovereignty in myriad but limited ways, it is not immediately obvious that EU actors responding to democratic backsliding in an EU member state or European political party or foundation are acting in self-defence. Recall that the strongest defences of the legitimacy of militant democratic action require the anti-democratic challenge to pose a real threat to the continued existence of the polity as a democratic community. We must therefore reflect on the extent to which the European Union’s own constitutional identity matters in this context.

Although it is tempting to address the question of democratic self-defence by asking whether or not the European Union is ‘a democracy’, this is a mistake. Whether or not one considers that the European Union is a sort of democratic federation, a supranational confederation or something in between does not matter, in my view. Regardless of one’s perspective on the European Union’s constitutional identity, ⁵³ the European Union acts in self-defence when it seeks to contain anti-democratic threats for two reasons. First, the European Union is centrally organized and shaped by its member states. Member state politics are in that sense internal to EU politics, even where a policy domain is the exclusive competence of the member state. The European Union acting to protect democracy in a member state is therefore responding to threats to democracy that are in the European Union, rather than elsewhere. Second, the European Union’s character as a polity committed to democracy can be threatened by anti-democratic actors. The European Union’s commitment to democracy is not merely an abstract ideal, but is implicated in the European Union’s legal coherence and in the democratic legitimacy of EU law and governance. Ulrich Wagrandl makes this point simply: ‘European citizens today are not only governed by themselves but, via European institutions, also by other states and their citizens, whom they and their state govern, in turn.’ ⁵⁴ If a member state government is no longer democratic, its representative co-legislating in the Council has no legitimate standing to co-decide EU law and its head of government can no longer legitimately claim to represent its citizens in the European Council. ⁵⁵ The democratic authority of EU institutions whose mandate is delegated by member states (such as the European Commission, the European Court of Justice or the European Central Bank) would also be tarnished if an autocratic member were to be permitted to weigh in on this delegation. Only the democratic mandate of Euro-parliamentarians seems untarnished, provided that European elections continue to be organized in a manner that is both free and fair.

In my view, we must therefore conclude that the European Union is acting in self-defence of its democratic legitimacy when it responds to try to combat democratic backsliding. Taken together, the fact that (1) the European Union does dispose of tools that pose prima facie challenges to democratic values by sanctioning anti-democratic actors’ equal political participation and (2) the European Union does act in democratic self-defence mean it is descriptively accurate to refer to the European Union as a ‘militant democratic’ actor.

IV. Is it legitimate for the European Union to act militantly in responding to democratic backsliding?

Having established that the European Union is a militant democratic actor, the question arises if it is justified in acting in using militant democratic responses to anti-democratic threats. My task here will be modest in that I ask whether the European Union can meet the minimal criteria for the justifiability of militant democratic action that I articulated in Part II of this article. I did not present a defence of the justifiability of militant democratic action from first-principles there, so my analysis here is necessarily conditional. Either way, if it is not legitimate for the European Union to act militantly to protect democracy in Europe, then the militant tools the European Union currently has at its disposal should not be used and, where possible, should be reformed in order to dissolve the militant democratic paradox. Instead of using militant democratic tools to respond to anti-democratic threats, the European Union should prioritize responses that do not raise challenges to democratic fundamentals, such as fines for infringing upon obligations under EU law and budget conditionality (see above). If it is legitimate for the European Union to act militantly, this speaks in favour of it doing so, at least when militant measures are necessary to contain existential threats to the European Union as a supranational community committed to democracy.

It would seem natural to start the investigation into the legitimacy of EU militant democracy by examining the justifications given in the literature for militant democracy at the national level. Indeed, in Part I of this article, I examined some of the major contributions to militant democratic theory, identifying the existential threat and the necessity conditions as minima for robust defences of militant democracy at the national level. However, before we can confidently apply these minimal conditions of justifications of militant democracy at the national level, we must see whether the European level is relevantly similar. And indeed, I argue that there are important limits to the analogy between national democracies and the context of the supranational EU.

Specifically, the ‘existential threat’ condition discussed in Part I plays out differently at the transnational European Union level. Core elements of state level politics, namely the existence of an ‘ineluctable and inescapable political authority’ wielding tools of ‘coordination and compulsion’, ⁵⁶ do not exist at the level of the European Union. Whereas national politics is characterized in part by the inescapable nature of political authority, EU member states retain the right to withdraw from the European Union ‘in accordance with [their] own constitutional requirements’ (Article 50(1) TEU). This is important, because it highlights the fact that the European Union is, at its core, a voluntarist association. Member states that no longer wish to conform to the demands set collectively via the Treaties, including a commitment to democracy and the rule of law, or who no longer wish to contribute to collectively addressing common problems in Europe, or who otherwise wish to secure a higher level of independence (real or imaginary) from their European peers, have the option of withdrawing from the European Union.

This may seem like a merely academic point. After all, many anti-democratic actors in the European Union are not opposed to the existence of the Union, nor to their states’ EU membership. While, of course, some strongly eurosceptic European political parties exist, and some of them are anti-democratic, other anti-democratic actors see the benefits of continued membership in a reformed European Union. At the level of member states, the governments of Hungary and Poland, the two member states most drastically regressing on democratic norms, seem fairly strongly committed to continuing EU membership at this time, in line with broad support for EU membership among their citizens. A possibility to voluntarily withdraw from the EU via Article 50 seems irrelevant if backsliding member states insist on continued association despite clearly signalling their unwillingness to commit to common European standards on, for example, judicial and media independence.

On the other hand, it is quite clear that the continued inclusion of states violating democracy and the rule of law has the potential to corrode the European Union’s claims to democratic legitimacy. Indeed, European Court of Justice President Koen Lenaerts recently warned that rule of law backsliding in Poland ‘threatens the survival of the European project in its current form’. ⁵⁷ He was commenting especially on challenges to the competence of the European courts to settle legal disputes – a competence put to the test in the winter of 2021 when the illegitimately composed body known as the Polish Constitutional Tribunal argued that key elements of the European legal order were at odds with the Polish constitution. More simply, as discussed above, if a non-democratic member state wields a vote (especially a veto) in key EU institutions such as the European Council and the Council of the European Union, those institutions can no longer be considered democratic, even in a derivative way. All decisions made in these institutions that are taken with the support of authoritarian actors are tainted by their inclusion, much like a parliament whose composition includes unelected bishops, mullahs, or hereditary aristocrats cannot call itself a democratic parliament.

None of this is to say that democratic backsliding in the European Union always meets the ‘existential threat’ condition; if we look at the first of the militant measures analysed above, for instance – the possibility under EU Regulation 1141/2014 for antidemocratic European political parties and foundations to be deregistered – it seems hard to justify on the existential threat standard. Deregistering European political parties and foundations – essentially de-funding them – would clearly undermine their ability to compete with other European parties on an equal playing field. It can also be used to sanction anti-democratic actors. For these reasons, we have classified it as a militant democratic mechanism. However, the activities of anti-democratic political parties and foundations do not at this point plausibly threaten the continued existence of the European Union as a polity committed to democratic government, so the existential threat condition seems illusive. The fundamental threat to democracy in Europe comes, at this point in time, from governments of EU member states that are pulling their nations towards autocracy, not from European political parties and foundations (though the latter may be complicit in the former). This point notwithstanding, it does seem clear that democratic backsliding of member states at the scale that has been observed in, for instance, Hungary and Poland, clearly puts pressure on the European Union’s character as a polity committed to democracy. So, while we may debate the precise threshold at which the European Union’s democratic character is existentially threatened, it is clear that democratic backsliding by member states can, in principle, meet that threshold (and may already have done so).

Militant democracy or European disintegration?

When we look at the second criterion of robust defences of militant democracy, the ‘necessity demand’, the case for EU militant democracy seems weaker. As we saw above, a common theme of militant democratic theories is the need for restraint. If militancy is not necessary to pacify anti-democratic threats to the continued democratic character of the polity, then such measures are illegitimate. Polities committed to democracy should ordinarily combat anti-democratic threats in ways that uphold democratic fundamentals such as equal civil and political rights. That may mean accepting anti-democratic political activity to some extent as a regrettable but tolerated excess resulting from the free exercise of civic and political rights. Or it may mean proscribing certain anti-democratic activities (say Holocaust denial, or the formation of extremist militias) generally through criminal law prohibitions and sanctions. Only when anti-democratic activities jointly (1) form an existential threat that (2) requires measures that pose prima facie challenges to fundamental democratic values does militant democratic theory justify their use.

It is in considering the necessity demand that, in my view, the normative justification of EU militant democratic action seems suspect. I have argued that, in a crucial way, the democratic authority of the European Union is derivative in a large part to the democratic mandates of the governments of its member states. This is irrespective of our answer to the question of whether we consider the European Union to be ‘a democracy’ or not. I have also argued that, given the possibility of dissociation, the European Union should be understood as a supranational Union grounded on voluntarism. An implication of these arguments is that EU member states do not have a right to EU membership. Instead, EU membership is conditional both on each member state’s continued acceptance of EU membership and on the continued willingness of other member states to be bound to them supranationally. ⁵⁸

If dissociation from authoritarian member states – a form of limited EU disintegration – insulates the EU polity from anti-democratic politics without posing prima facie challenges to democratic values, then these steps should be preferred to militant democratic responses, even when the European Union is faced with an existential threat to its democratic character. In my view, the same voluntarism that allows member states to leave the European Union creates possibilities for dissociating with an authoritarian member state. The nature of the European Union as a voluntary Union therefore undermines the legitimacy of EU militant democracy.

Let us take each possibility for dissociation in turn to see whether they would permit member states committed to democracy to contain anti-democratic threats. In an article on the democratic legitimacy of European disintegration, Markus Patberg makes a helpful classification of ways in which the EU polity could disintegrate. ⁵⁹ He analyses five forms: ‘retreat’, ‘revocation’, ‘exit’, ‘expulsion’ and ‘dissolution’. With ‘retreat’, individual member states try to negotiate exemptions from certain rules or obligations to which they were previously subject, and that others continue to be bound by. ⁶⁰ ‘Revocation’ is a more generalized form of disintegration, in that all member states agree to turn back a specific aspect of European integration. ‘Exit’ occurs when one or more EU member states withdraw from the European Union (as with Brexit), while ‘expulsion’ refers to member states deciding that one or several member states must have their membership withdrawn. ‘Dissolution’, the most radical form of disintegration, would occur if member states agreed to dissolve the European Union entirely.

‘Retreat’ and ‘revocation’ provide theoretical possibilities of dissolving the threat that backsliding member states pose to the democratic character of the European polity, but only when taken to the extremes of these types of disintegration. Retreat occurs when member states ‘opt out’ of certain rules or obligations. This form of renegotiation would require other member states’ agreement. However, only the theoretical possibility of a backsliding member state negotiating an opt-out for the quasi-totality of collective rules and obligations would resolve the threat posed to the democratic character of the European Union by backsliding states. An autocratic state continuing to associate with democratic member states in some ways that pool its sovereignty means it would continue to pose a threat to the democratic legitimacy of those elements of association. Furthermore, retreat puts the ball in the court of the backsliding state(s) and therefore does not seem to be an apt alternative to militant democratic responses in that it cannot be a pathway for democratic self-defence by pro-democratic actors.

‘Revocation’ could theoretically extinguish the threat posed by authoritarian member states to other member states, but does not seem to be a possibility for addressing the threat posed by backsliding member states at the level of the EU polity. As with retreat, only a radical version of revocation would theoretically contain authoritarian member states. For only if democratic member states negotiate exemptions from all areas of EU integration where they currently pool sovereignty with an autocratic member state would they be fully insulated from supranational association with that member state. The European Union, in this hypothesis, would become a merely intergovernmental organisation. Any revocation less than this step would mean that the democratic legitimacy of each democratic member state continues to be tarnished by supranational association with an authoritarian EU member state, as with retreat. While revocation is not, like retreat, the sole prerogative of the backsliding state, it would require their agreement, and thus also seems to capture the notion of democratic self-defence poorly, if at all.

With ‘dissolution’, the threat of authoritarian politics to individual pro-democratic member states may be protected because such states would no longer pool their sovereignty with autocratic member states were the European Union to be dissolved. However, as with the radical form of revocation discussed above, this type of EU disintegration would neutralize the threat of autocratic politics by getting rid of the European Union entirely. Furthermore, and similarly to radical revocation and retreat, dissolution would require the agreement of the EU member state that has become autocratic, thereby making it ill-suited as an alternative to militant democratic measures of self-defence by pro-democratic actors in the European Union.

‘Exit’ and ‘expulsion’ seem more adequate for neutralising the threat of anti-democratic politics to the democratic character of the EU polity. ‘Exit’, however, could address the existential threat of authoritarian politics at the level of the EU polity only if authoritarian member states used Article 50 TEU to withdraw from the European Union. This would mean that the authoritarian state(s) no longer participated in the supranational legislative and decision-making processes at the EU level, thus removing the putative existential threat they might pose to the democratic character of those processes. While this would seem an effective solution, there are two problems. First and most importantly, the exit route relies on the willingness of the autocratic state to withdraw from the European Union unilaterally, given that Article 50(2) requires the withdrawing state to notify the European Council of its intention to withdraw. Second, Article 50(1) requires that member states withdraw in accordance with its own constitutional requirements (my emphasis). This seems to imply that an overreaching, authoritarian executive may not always be legitimately authorized to exit the European Union, even if they wanted to. ⁶¹ Finally, given that democratic self-defence focuses on the agency of EU institutions and pro-democratic member states, the fact that exit lies entirely in the hands of the backsliding state means it does not really pose an alternative to militant democratic measures focused on democratic self-defence.

‘Expulsion’ in this sense seems to better meet the requirements of neutralizing the threat of authoritarian politics to the legitimacy of EU law and politics. Removing EU membership from a state whose government has become frankly autocratic would effectively contain the influence of autocratic politics from all EU legislative and political processes. Furthermore, unlike with the radical revocation considered above, expulsion has the potential of protecting the democratic character of the EU polity itself (which in the scenario of radical revocation would be dissolved as a supranational-democratic association). One may argue that the main democratic cost of expulsion to the European Union is the democracy of the expelled member state itself, but that would be to ignore the fact that, by stipulation, that state would already have regressed so far on democratic standards as to no longer to be a democratic polity. Nevertheless, as noted by Patberg (and many others), there is no Treaty provision for expulsion, so for this measure to be used as an alternative to militant democratic responses, the Treaties would need to be reformed. ⁶²

There may be another way to functionally expel a frankly autocratic member state without Treaty change, but it would require the close coordination of pro-democratic member states. As I have argued in greater detail elsewhere, the Article 50 procedure can be used not just by an individual member state to leave the European Union (as was the case with Brexit), but it can also be used in a coordinated fashion by a group of member states. This means it could theoretically be used to exclude member states with severe democratic deficits from continued European cooperation. ⁶³ One way in which this could work is for member states committed to democracy to collectively exercise their right to withdraw from the European Union via Article 50 TEU and refound a new European Union 2.0. ⁶⁴ Indeed, if enough member states were to do this, they could negotiate their collective withdrawal from the European Union in a manner that transfers the resources of the current European Union to this new supranational association. While the autocratic member state or states would continue to enjoy de jure membership of the old ‘European Union’, the substantive rights and benefits of European integration would be denied to them; it ‘would leave autocratic Member States in the empty useless shell of the original EU’. ⁶⁵ Crucially, given the absence of a right to supranational association, such a move to disassociate from an autocratic member state or member states would not violate democratic fundamentals (although, of course, it would come with weighty democratic costs to the citizens of those states). ⁶⁶

In sum, if we accept the view that the European Union is a voluntarist association, it seems there are possibilities of dissociating from autocratic member states. Such possibilities do not pose prima facie challenges to democratic fundamentals, as there is no democratic right to EU membership. If dissociation from a frankly autocratic state resolves the existential threat to the democratic character of the European polity, then militant democratic responses cannot meet the necessity demand. I have argued that all five types of European disintegration give theoretical ways of dissociating from a frankly autocratic state, but that only expulsion would protect the EU polity’s democratic character and empower pro-democratic actors to respond to existential threats in self-defence. The possibility of expulsion therefore undermines the legitimacy of EU militant democracy.

V. Conclusion

This article made two arguments which are independent from one another. First, taking a narrow definition of militant democracy, I systematically analysed different legal tools to respond to democratic threats in the European Union. The goal was to assess whether, descriptively, the European Union is indeed a militant democracy – that is, whether it possesses tools to respond to anti-democratic threats that are targeted to the anti-democratic actors’ political participation in such a way as to pose a prima facie challenge to fundamental democratic commitments. I argued that the European Union is indeed a militant democracy in this precise way: the sanctions described in Article 7 TEU and EU Regulation 1141/2014 meet these criteria (notwithstanding the fact that these sanctions have never been imposed). So, once we focus on what, in my view, is the heuristically most interesting and useful elaboration of militant democracy – namely a conception that acknowledges the militant democratic paradox – we can see that the European Union is indeed a militant democracy. Furthermore, I argued that the question of whether or not the European Union is ‘a democracy’ is a red herring. We do not need to settle the question of the ultimate constitutional character of the European Union to assess whether the European Union can act in democratic self-defence. Either way, the legitimacy of the European Union’s institutional structure depends on direct and indirect democratic legitimation; it thus can be said to be acting in democratic self-defence when responding to internal anti-democratic threats.

The second argument of this article considered the normative justification of militant democratic responses to anti-democratic threats in the European Union. I addressed this question by testing whether the existential threat and necessity conditions could be met in the context of anti-democratic challenges to the European Union’s character as a polity committed to democracy. I argued that anti-democratic politics can pose an existential threat to the democratic character of the European Union, especially in the context of democratic backsliding by EU member states. However, the necessity condition is undermined by the possibility of isolating a frankly autocratic member from European integration. Working from a voluntarist conception of EU membership, I considered different pathways of EU disintegration that could neutralize the threat of a frankly autocratic member state. I argued that expulsion in particular would be a form of limited disintegration that would contain the threat of autocratic member states to the European Union and would not violate democratic fundamentals. If this argument is convincing, then EU actors are never be justified in acting militantly. Either the necessity threshold is not met because anti-democratic actors fail to pose an existential threat to democracy in the European Union (as I argued is the case regarding the deregistration of European political parties and foundations) or there are robust responses available that would not raise the paradox of militant democracy.

Given the voluntarist nature of EU supranational association, the European Union is limited to acting in a manner coherent with the fundamental values it purports to hold. We can think here for instance of mechanisms to support pluralist democracy and civil society in member states where they are threatened, to politically oppose authoritarian actors in their projects to undermine democracy in EU member states, to hold backsliding state governments accountable to their legal obligations to uphold domestic democracy and organize free and fair elections, and to sanction them with fines and budget conditionality where they fail to do so. ⁶⁷ It is probably just as well that EU actors should refrain from acting militantly; the absence of institutional resources giving effective form to EU pouvoirs régaliens suggests that the European Union would be structurally unsuited to most kinds of militant democratic action, which tend at the state level to mobilize the coercive apparatus of the state. The European Union simply does not dispose of the sorts of power that most forms of militant democratic action require. It is perhaps not surprising from this perspective that the two militant democratic sanctions available to the European Union have not yet been used. Where non-militant measures fail, democratic members of the European Union should dissociate with authoritarian member states to preserve the democratic character of the Union and, consequently, of their own polities.