Parliamentary Democracy by Default: Applying the European Convention on Human Rights to Presidential Elections and Referendums

This paper is concerned with the Convention’s “democracy clause,” that is Article 3 of Protocol No. 1, which provides for the right to free elections. Why should it be described as a “democracy clause” and what is its significance for today? The paper first sketches out the drafting history, which reveals that while the framers were keen to preserve their inherited domestic institutions, they also thought it crucial to promote democracy. The Convention invokes but does not define democracy. It is within the Court’s competence to elucidate its meaning. The Court holds that pluralist democracy is the only system compatible with the Convention. This paper argues that Article 3 of Protocol No. 1 further presupposes a representative legislature. It does not require member states to introduce a specific system of elections and representation, but it obliges them to conform to parliamentary democracy. Thus, the Court’s model of democracy rests on two pillars: pluralism and parliamentarism. It subscribes to pluralist democracy theories but contradicts the monolithic conception endorsed by Carl Schmitt. The Court’s model has the potential to offer a robust account of democracy. Yet, Article 3 of Protocol No. 1 has never lived up to its potential. The Court’s relevant jurisprudence is inadequate to address the contemporary antidemocratic shifts that are underway in certain member states. Hence, the paper suggests that the Court’s power to apply this clause is not limited to general elections but also extends to presidential elections and referendums.

1 Introduction "(T)he ECHR cannot be correctly applied and effectively respected by a dictatorial or an illiberal regime. Only a genuinely democratic regime is compatible with the philosophy and the spirit of the Convention." 1 That is how the president of the European Court of Human Rights (henceforth, the Court) described the close link between human rights protection and democratic institutional design when he addressed a conference organized to mark the 70th anniversary of the establishment of the Convention in May 2020.
The criterion of democracy pervades the Convention. Its preamble emphasizes that human rights "are best maintained on the one hand by an effective political democracy and on the other by a common understanding and observance of the human rights upon which they depend." Thus, safeguards on human rights are inseparable from the institutions of an effective political democracy. The notion of "democratic society" is also invoked in the limitation clauses of Articles 6 and 8-11. Yet, the Convention does not define democracy. It is left to the Court to elucidate its meaning.
Today, there is a consensus in academic circles that the Court is the single most effective regional judicial body. 2 The term "effective" is mostly used to emphasize that its judgments are mandatory for the member states and that the Court provides a far higher standard of human rights protection than was ever intended. 3 It has an impressive and unique human rights record, but it is effective in another sense too. The Court has had an enormous impact on the nature of democracy in the member states. For instance, it has played an essential role in stabilizing democracies in post-war and post-cold war Europe. 4 The Court promotes democracy by insisting that democracy is the only political model contemplated by the Convention and the only system compatible with it. 5 Over time, it has adopted a model of democracy that is centered on the idea of a pluralistic society, and this approach is remarkably prominent in the case law on the rights ensured by Articles 10-11. 6 Nevertheless, it is only Article 3 of Protocol No. 1 (henceforth, Article 3) that specifies that the Convention's model of democracy is a parliamentary democracy.
Thus, the paper devotes special attention to the provision of what it calls the "democracy clause." 7 It argues that Article 3 is at the very foundation of what the preamble describes as an "effective political democracy" and what the limitation clauses designate "democratic society," because it is predicated on the existence of a democracy with free and fair elections and a deliberative parliament. The paper further argues that consequently, the Court's model of democracy has two pillars: pluralism and parliamentarism. This model is compatible with the pluralist democracy theories of modern political thinking but is incompatible with the monolithic version of "democracy" as developed by the influential theoretician Carl Schmitt.
Although the Court seems generally responsive to the idea that the Convention protects parliamentary democracy, the case law of Article 3 does not reflect this mission coherently. Taking an example from Turkey, where a nationwide referendum resulted in constitutional amendments that transferred legislative powers to the directly elected president, the paper demonstrates that the jurisprudence associated with Article 3 is inadequate to protect parliamentary democracy and address the contemporary antidemocratic systemic shifts certain member states are facing. Hence, the paper proposes to read Article 3 as a guarantee of parliamentary democracy that focuses not only on issues that concern parliamentary elections but also on referendums 8 and presidential elections that clearly and substantially affect parliamentarism in a member state.
The paper proceeds as follows. The first part presents a brief outline of the history of Article 3 with a view to exploring the drafters' approach to democracy. The travaux préparatoires provide fascinating insights into the clause's origins and substantiate the assumption that, in inserting this article, the drafters' aimed at strengthening domestic parliamentarianism and preventing backsliding by new democracies. 9 Then, in the second section, the focus shifts to the Court's jurisprudence. The third part addresses the following question: what theory of democracy supports the drafters' intentions and the Court's case law? In answering the question, the paper draws a distinction between the pluralist and monolithic versions of democracy and explores the strength of the Court's pluralist democracy model. The fourth section connects the Court's pluralist model to Article 3 and considers the limits of the current Article 3 jurisprudence, as this is not capable of addressing major systemic changes that may happen as a result of presidential elections and referendums. Hence, the fifth part argues in favor of a change in the Court's interpretation so that the Court applies Article 3 to presidential elections and referendums that clearly and substantially affect the situation of parliamentarism in a member state. This section also demonstrates how the suggested shift in the Court's interpretation would make it possible to recognize the outcomes of presidential elections and referendums as legitimate both within the country and abroad. The conclusion summarizes the findings.

The Drafters' Approaches to Democracy
In the late 1940s and early 1950s, the mission of the Convention meant different things to different drafters. For some, it was designed to strengthen already existing domestic democratic institutions. They took for granted that member states enjoyed democratic legitimacy and that human rights were already well protected; hence, for them, the role of the Court was merely to correct isolated illegal domestic acts and to overturn aberrant decisions when all domestic remedies had been exhausted. 10 For others, the main aim of the Convention was to ensure that member states stayed true to their democratic aspirations and to prevent backsliding by the new democracies. 11 Thus, they assigned the Court the task of serving as an "alarm that would bring … large-scale violations of human rights to the attention" 12 of Europe by making a "'pre-emptive strike' against the menace of new tyrants." 13 Overall, the Convention can be 8 There is no consensus in the legal and political science literature on the meaning of the term "referendum." Ulieri 1996, 3. I understand a referendum as a general vote by the entire electorate on a particular political question that has been referred to them for a direct decision. 9 At that time, the category "new democracies" contained those that had been "continuous democracies only since a date between 1920 and 1950: Austria, France, Italy, Iceland, Ireland, and West Germany." Moravcsik 2000, 233 10 Bates 2010, 26 11 Moravcsik 2000 Harris et al. 1995, 2 13 Wildhaber 2011 seen as a compromise between these two views, 14 though the voice of those who argued that the Convention system was meant only to overturn aberrant decisions when all domestic remedies had been exhausted was and still is stronger. 15 Yet, the first proposal for a European convention was more representative of the first view. The draft, which was crafted by a political organization called the European Movement, 16 aimed at creating "a type of collective pact against totalitarianism." 17 When the Council of Europe was established and its two institutions, the Consultative Assembly and the Committee of Ministers, were set up, 18 the Assembly started to reformulate the so-called European Movement Convention. David Maxwell-Fyfe, who was head, and Pierre-Henri Teitgen, who was rapporteur of the Assembly's Committee on Legal and Administrative Questions, played prominent roles in this process. They altered the Convention's text, but as Teitgen elegantly noted in an Assembly sitting, the primary aim and the most consistent public justification for the Convention remained the same: "We are less concerned to set up a European juridical authority capable of righting isolated wrongs, isolated illegal acts committed in our countries, than to prevent, from the outset, the setting up in one or other of these countries of a regime of the Fascist or Nazi type. That is the essential element of our purpose. We are seeking an international procedure capable of active intervention right from the start." 19 He was aware that "evils progress cunningly, with a minority operating, as it were, to remove the levers of control," 20 so he found it necessary for the European institutions to intervene before it was too late. He used every occasion to stress the importance of including the basic guarantee of democracy in the forthcoming text. The situation was a delicate one. There was a prolonged and challenging negotiation between the Consultative Assembly and the Committee of Ministers. The contentious issue was whether the Convention should include a clause guaranteeing democracy as such and not just a list of a select number of civil and political rights. The representatives of the assembly and the ministers struggled to reach a consensus on this specific provision.
The first assembly draft set out a list of civil and political rights to be protected by the states. The rights were merely listed, and the document contained a pledge to "respect the fundamental principles of democracy in all good faith and in particular, as regards their metropolitan territory, to hold free elections at reasonable intervals, with universal suffrage and secret ballot, 14 But see Michael O'Boyle, who emphasizes that the case law built on Article 3 echoes this disagreement as to what the role of the Court should be in the area of electoral rights: The Court has already contributed to ensuring the democratisation of electoral systems in the former Soviet-satellite states (see, e.g., the pilot judgments) and occasionally, it has also reminded consolidated democracies of particular shortcomings in the protection of electoral rights. O'Boyle 2009O'Boyle -2010 Sadurski 2020, Ulfstein 2020. 16 The aim of the International Committee of the Movements of European Unity (or the European Movement for short), which had eminent statesmen among its members, was to guarantee peace in Europe by creating a sort of European unity. 17 Bates 2010, vii 18 In the early years, the Council of Europe consisted of these two institutions. The deliberative body was the inter-parliamentary Consultative Assembly, which discussed European matters; the conclusions it forwarded to the Committee took the form of non-binding recommendations or opinions. As Raymond Aron aptly put it, the assembly did not have "a shred of sovereignty." Aron 1949. The intergovernmental Committee of Ministers was the decision-making body, which was free to reject the proposals of the assembly. For more on the relationship between the two bodies, see Robertson andMerrils 1994, 12-13 19 Consultative Assembly (1950) 20 Consultative Assembly (1949Assembly ( ) 1158 so as to ensure that Government action and legislation are, in fact, an expression of the will of the people." 21 In addition to this, the text included a right to express criticism and to form an opposition. Thus, this draft listed the rights to free elections, to express criticism, and to form an opposition as fundamental principles for the functioning of democracy.
Experts advising the Committee of Ministers argued that including such an article would not be unproblematic because it was impossible to reach agreement on the meaning of the fundamental principles of democracy. 22 Furthermore, the UK government opposed having a reference to "universal suffrage," the right to vote being, in many founding member states, subject to restrictions: inhabitants of the colonies and protectorates did not acquire citizenship. As an alternative, they suggested inserting a provision to ensure the protection of the member states' already existing democratic institutions. 23 The British were eager to preserve the ancien régime by protecting the domestic institutions they inherited and their overseas empire, so they insisted that the provision should focus explicitly on institutions. As a response, the Committee of Ministers presented a counter-draft that deliberately omitted this particular clause on democracy altogether. In the Assembly, this caused widespread outrage. Assembly members strongly urged the ministers to include a provision requiring member states to respect "the political liberty of their nationals and in particular, with regard to their home territories, to hold free elections at reasonable intervals by secret ballot under conditions which will ensure that the government and legislature shall represent the opinion of the people." 24 So while the first draft protected the democratic principles, and the second draft sought to protect the democratic institutions, this third draft, representing a compromise, aimed at protecting individuals' political liberties. 25 Ultimately, the Convention was signed on 4 November 1950 without a clause guaranteeing either democratic principles, institutions, or political liberties, because the ministers were not able to reach agreement on its content. However, they did decide to add a protocol incorporating, among other things, the right to free elections. When drafting the protocol, the UK government proposed a modification of the text to emphasize that holding free elections must "ensure the free expression of the opinion of the people in the choice of legislation and government." 26 This proposal was motivated by a fear that the Assembly draft's requirement of elections "under conditions which will ensure that the government and legislature shall represent the opinion of the people" 27 might give rise to an obligation to adopt a proportional representation system. 28 The Belgian government wanted to delete the word "government," arguing that the choice of the government was not necessarily made directly by the people. 29 Both proposals were adopted, and two other changes were also made in the text. The term "political liberty" was omitted because it was not sufficiently precise, and the reference to "home territories" was also removed, because the protocol did not apply to overseas territories. On 20 March 1952, the protocol was finalized and signed by the ministers. Article 3 stipulates: "The High Contracting Parties undertake to hold free elections at reasonable intervals by secret ballot, under conditions which will ensure the free expression of the opinion of the people in the choice of the legislature." Although the text of Article 3 was watered down, and consequently, it became relatively modest, the provision had a systematic ambition. The drafting documents reveal that the founding member states were acting in their own best interests to protect their privileges as colonial empires, but at the same time, they were united in their commitment to prevent systemic shifts. The preparatory works suggest that there was a consensus among the drafters on the importance of democracy, which was forged by their shared anti-totalitarian commitment. The Second World War delivered a widely understood historical lesson; subsequently, politicians who aimed to prevent backsliding ("never again") created a European human rights regime whose members were committed to protecting democracy. They agreed that the new European order should be based on parliamentary democracies and that a special "democracy clause" was needed to fight significant systemic changes. Thus, the core concept of Article 3 is associated negatively with anti-totalitarianism and positively with parliamentary democracy.
In the early 1950s, it was easier for the member states to reach a general consensus on the importance of parliamentarism than it would be today. At that time, all the founding member states were democracies with functioning parliaments. 31 They institutionalized their democratic systems differently: they used different systems of government (e.g., unitary, federal) and constitutional forms (e.g., republican, constitutional monarchy), and they had different electoral systems (e.g., proportional representation, first-past-the-post). However, there was a consensus among them on the core feature of a democratic system: the national parliaments were at the heart of their democratic structures. For them, parliament embodied democracy, and the representatives in these parliaments gained their mandates in free, fair, and competitive elections between multiple distinct political parties.

The Court's Model of Democracy
In the early years, the Court did not expressly address the issue of democracy in those terms. It was only in the late 1970s and early 1980s that a general consensus emerged in the Court on certain core features of a democratic system. 32 Since then, the Court has applied "pluralism" as an umbrella term that is essential to understanding the Court's model of democracy. 33 By connecting the idea of pluralism with that of democracy, the Court has acknowledged that a society is only democratic when there are competing value systems and everyone's values are equally respected.
The Court's understanding of pluralism applies to various aspects of the democratic society. First, it applies to the plurality of political parties. 34 As the Court puts it, to promote "dialogue and exchange of views necessary in a democracy, it is important to ensure access to the 31 Belgium, Denmark, France, Ireland, Italy, Luxembourg, the Netherlands, Norway, Sweden, and the UK. 32 This, in practice, meant that Greece could rejoin the Council of Europe only after the restoration of democracy in 1974, and the Parliamentary Assembly seated the Turkish members only after the demise of the military dictatorship in Turkey. Tomuschat 2013, 472 33 ten Napel 2009 United Communist Party of Turkey and Others v Turkey political arena for oppositional parties on terms which allow them to represent their electorate." 35 Connected to this, the Court has recognized the separation of powers as a distinctive feature of democracy 36 and noted that parliamentary control requires an opposition that can serve as a check on the parliamentary majority.
Second, the Court has spoken of pluralism in public debate and education. 37 The case law strongly connects freedom of speech and democracy; the Court has stressed that "freedom of political debate is at the very core of the concept of a democratic society," 38 and "democracy must be based on dialogue and a spirit of compromise." 39 Since the Handyside case, the Court has repeatedly emphasized that without pluralism, tolerance and broadmindedness, there is no democratic society. 40 Moreover, the case law links the plurality of opinions with parliamentarism; for the Commission, the purpose of parliamentary elections "is to ensure that fundamental changes in prevailing public opinion are reflected in the opinions of the representatives of the people." 41 These statements should not be understood as factual recognitions of the existing societal diversity and disagreement; rather, they convey the message that pluralism serves democratic society. 42 From this, we can conclude that the Court regards pluralism as a condition sine qua non for a democratic society and emphasizes the importance of a pluralist political process within a pluralist society. 43 Yet, only Article 3 requires the procedures and institutions of parliamentarism to be necessary features of the member states' basic democratic structures. 44 It encompasses the procedures of parliamentarism: the secret ballot and individual voting rights. The clause prescribes that the legislative power should rest with a body constituted as a result of free elections and confirms the importance of the individuality and secrecy of voting. Furthermore, Article 3 explicitly requires member states to have an institution called a legislature. The term "legislature" is a generic term for a constitutionally designated institution that gives assent on behalf of a political community to public policy measures. 45 The principal task of the legislature is to prescribe general, binding rules for each member of society, that is, lawmaking. By using the term "choice," Article 3 implies that being directly elected is a defining characteristic of the legislature; hence, the people should elect at least one chamber of the legislative body to meet the conditions of free and fair elections. Thus, the representative principle 46 is also part of Article 3. In addition, Article 3 envisages a legislature performing a deliberative task. Parliamentary democracy under Article 3 presupposes the idea of discussion and openness. The underlying assumption is that the legislature is an institution where the political will of the people is constructed in a structured procedure, that is, in an open 35 Nieuwenhuis 2007, 373 44 There is a burgeoning literature on Article 3. For the purposes of this article, there is no need to delve into the vast body of case law that addresses all the relevant aspects of this provision. For a detailed examination, see e.g., van Dijk et al. 2018. Rainy et al. 2017. 45 Norton 1990, 1 46 See e.g., Mill's analysis of parliament's function in Mill 1910, 239. discussion on all matters of public interest; laws arise out of a conflict of opinions. Article 3 explicitly refers to the people's free expression of opinion, thereby accepting that an informed electorate should choose the legislature. Here, one can recognize Edmund Burke's ideas 47 and John Locke's suggestion that social progress may be realized through representative institutions and public discussion. 48 The Convention bodies recognized the significance of Article 3 at a very early stage. The Commission interpreted the provision as one that "presupposes the existence of a representative legislature, elected at reasonable intervals, as the basis of a democratic society." 49 In its first judgment dealing with Article 3, the Court also emphasized that "since it enshrines a characteristic principle of democracy, Article 3 of Protocol 1 is accordingly of prime importance in the Convention system." 50 Yet, the unique and open-ended formulation of Article 3 leaves the Court considerable scope for interpretation. According to a literal interpretation of Article 3, it is framed as an undertaking on the part of the high contracting parties to hold elections. For a long time, the phrase "the High Contracting Parties undertake" used by Article 3 was taken to mean that this article solely gives rise to state obligations and not individual rights. However, when first asked to determine complaints under Article 3, the Commission, the Court's predecessor stated that it ensures the right to free elections: "whatever the wording of Article 3, the right it confers is in the nature of an individual right since this quality constitutes the very foundation of the whole convention." 51 The Court later endorsed this view and emphasized that the formulation is due to a "desire to give greater solemnity to the commitment undertaken." 52 Since then, European citizens have looked to the Court to guarantee their rights.
As shown above, the wording of Article 3 is the product of a particular set of historical circumstances, and the difficulties of the wording may cause problems in interpretation. Some election principles-free elections and the secret ballot-are explicitly enshrined in the document. Others-universal, equal suffrage-cannot be found expressis verbis in the text, but that does not mean that they are not protected at all. The Court has interpreted the right to free elections to mean that it allows individuals with grievances against the state regarding both enumerated and unenumerated election principles to challenge their treatment. The Convention bodies have delivered decisions on the requirement for universal suffrage, 53 on the content 54 and deprivation of voting rights, 55 and on the organization of elections within the country 56 and abroad. 57 Moreover, in some cases, case law has gone further than in most domestic legal systems with regards to the right to vote of convicted prisoners 58 and incapacitated persons. 59 The Court has acknowledged that, despite the unique formulation of its text, Article 3 provides for an individual right to vote and encompasses unenumerated election principles as well. So while in the early 1950s, ensuring an individual right to vote might not have been in the drafters' mind, when the Court was called to interpret it as a state obligation or an individual right, it opted for the latter, because that interpretation was in accordance with the original underlying principles.
Likewise, the case law of Article 3 suggests that the provision may not just apply to general elections but also to other elections. Today, the scope of Article 3 includes regional parliamentary elections 60 and European parliamentary elections. Remarkably, the Court has set aside the argument stressing that Article 3 intended to address national legislations alone by arguing that the European Parliament assumed the powers and functions of national parliaments. 61 The Court thus characterized the European Parliament as a legislature in the sense of Article 3, because national legislatures had transferred legislative powers to it. Hence, the fact that the framers of the provision did not and could not have had European Parliamentary elections in mind did not mean that today Article 3 ought not to apply to situations which, according to the drafting history and the preparatory works, were not among the debated issues in the 1950s.
In sum, over time, the Court developed a pluralist model of democracy and acknowledged the significance of Article 3. It recognized that Article 3 ensures an individual right to free elections and that it even protects unenumerated election principles. The Court also understands Article 3 as a provision that applies to elections other than general elections if those elections concern institutions that assume legislative powers. Thus, the Court's pluralist model of democracy, together with the protections in Article 3, may provide a robust account of democracy.

Pluralistic Versus Monolithic Versions of Democracy
What kind of democracy theories supports the Court's model of democracy and what democracy theories are incompatible with it? The Court's model of democracy seems to go against the "monolithic version of democracy" as developed by Carl Schmitt. 62 This section sharply contrasts the pluralist and monolithic versions of democracy and very briefly outlines Schmitt's conception, precisely because this is the most influential theoretical antipode to the pluralistic model.
The two pillars of Schmitt's monolithic version of democratic theory are the two principles of representation and identity. They form complementary poles, both of which are necessary to constitute a state. 63 For Schmitt, the representation of the people based on an election in a discursive, autonomous parliament is a fundamentally aristocratic idea. 64 Taking the complex process of representation that is characteristic of pluralist democracy as his point of contrast, he emphasizes a "genuine" representation of the will of the people understood as a substantial unity, a full identification (if not literal identity) between leaders and led. For him, democracy is an attempt to establish this "genuine identity" between rulers and the ruled. 65 The ruler may 60 X, Y and Z v. the Federal Republic of Germany, Decision of the Commission of 18 May 1976 61 Rudolf 1999, 685 62 Schmitt 2000, 22-50. Schmitt does not call his system "sovereign dictatorship," rather he reinterprets "democracy," because he accepts the inevitability of democracy. 63 Bielefeldt 1996, 393 64 Arato 2000, 942 65 Scheuerman 1995, 138 be a directly elected unitary sovereign, who acts as an authentic representative of the people. Schmitt does not offer a procedure adequate to test whether the sovereign assumes his powers in the name of the people. Instead, he emphasizes that even the silent agreement of the people is always possible and easy to recognize. 66 For Schmitt, the ruled are the people who exist in their ethnic and cultural "oneness." 67 This oneness ensures the strict internal homogeneity of the community. The homogenous people themselves cannot take part in discussions or deliberations; they can only engage in acts of acclamation, voting, and saying yes or no to simple questions posed to them by the unitary sovereign. 68 Hence, the monolithic version of democracy espouses a mass-based executive-centered plebiscitary regime where the people "possess a rather modest, even passive role." 69 This non-representative, immediate democracy without free elections, political parties, and parliaments, where representation is replaced by the consensus of the people in their "immediate presence," and dissent is not publicly visible, 70 is not compatible with the Court's model of democracy. A plebiscitary-style elected executive that rules by decree without a legitimate parliamentary mandate is also incompatible with it. 71 Moreover, the Court's model deliberately challenges the Schmittian characterization of the principles of discussion and openness as outmoded and counters his assertion that basic liberties, such as free speech, would lose their "rationale" with the demise of a deliberative parliament. 72 For all these reasons, the monolithic version of democracy is irreconcilable with the Court's model of democracy.
The question that then arises is what concept of democracy supports the Court's model. As the previous sections demonstrated, the Court's model has two pillars: pluralism and parliamentarism. Pluralism forms a key to many theories of democracy. There are a great many pluralist theories of democracy in modern political thinking, from the works of Robert Dahl 73 to Thomas Christiano. 74 Although the Court certainly does not follow any specific theory, its model is compatible with all theories that emphasize the importance of the pluralist political processes within a pluralist society, including multiple political parties. 75 A common denominator of the Court's model and many pluralist theories is that they emphasize free and fair elections with ongoing public discourse. In the Court's model, the jurisprudence built on Article 3 explicitly excludes the possibility that democracy can rely on instruments other than a legislature based on pluralist popular elections.
In short, the Court's model of democracy, which is built on pluralism and parliamentarism, is reconcilable with various pluralist conceptions but is irreconcilable with the monolithic version of democracy endorsed by Schmitt. The next section will consider how the Court's current Article 3 jurisprudence fits the Court's model of democracy. 66 Schmitt 2008, 139 67 Preuss 1995, 153-4 68 Schmitt 2008, 131 69 Scheuerman 1996, 313 70 Schmitt 2000. As Heiner Bielefeldt reminds us, Schmitt first recognized the principle of representation and free mandate, but later he became the propagandist of "a pure plebiscitarian democracy." Bielefeldt 1996, 393-394 71 Schokkenbroek 2006, 913 72 Schmitt 2000, 37 73 Dahl 1959. 74 Christiano 1996 See, e.g., Zysset, who assesses the Court's democratic society jurisprudence based on Christiano's egalitarian theory of democracy or Letsas, who discusses the Convention in the context of Ronald Dworkin's work. Letsas 2004, 280 5 The Limits of the Court's Article 3 Jurisprudence Since its early years, the Court has been very cautious in its assessment of the scope of Article 3. 76 Following the suggestion of the travaux préparatoires, Article 3 still "applies only to the election of the 'legislature,' or at least one of its chambers if it has two or more." 77 The Court interprets the term "legislature" in light of the prevailing constitutional structure of the state in question 78 and does not characterize the concept of the "legislature" as "autonomous," 79 unlike other notions of the Convention (e.g., "criminal charge," "civil rights and obligations"). 80 Thus, the extent of its coverage remains narrow because the Court considers the cases concerning presidential elections and referendums as not within the ambit of Article 3. 81 Applications dealing with presidential elections have repeatedly been declared inadmissible. 82 For instance, the Court has held that Article 3 did not apply to the appointment of the federal president of Austria, 83 the election of the president of Azerbaijan, 84 and the president of the Former Yugoslav Republic of Macedonia. 85 However, in the latter case it left open whether Article 3 might relate to elections for the head of state 86 on the grounds that it is a function with significant legislative power.
When it comes to referendums, the Court has been more reluctant to declare applications admissible. According to the established case law, Article 3 does not apply to referendums, 87 because the convention does not guarantee a "general right of consultation to the population" 88 and it does not require that the people be consulted via referendum. 89 The Court has declared complaints concerning referendums inadmissible, mostly on the grounds that the complaint in question is incompatible ratione materiae with the provisions of the Convention. 90 There have also been cases in which the Court has declared the complaint inadmissible ratione personae. For instance, in the Ouardiri case, the Court held that a Swiss Muslim and Muslim associations were not entitled to claim against the vote on the Swiss ban on minarets, because they could not sufficiently prove that they were victims of a violation of the Convention as they did not intend to build minarets. 91 Although the Court had not excluded the possibility "that a democratic process described as a 'referendum' … could potentially fall within the ambit of Article 3," 92 it always insisted that the decisive factor was that the referendum in question was not an election concerning the choice of the legislature. Even those referendums in which the 76  people were effectively voting to determine the type of legislature that they would have (e.g., EU accession or membership referendums, independence referendums) fell outside the Convention's sphere of protection. 93 In brief, the Court has repeatedly refused to review applications relating to referendums and presidential elections. The Court's firm stance on referendums and its ambiguous position on presidential elections have had wide-ranging effects. The paper illustrates this with the case of Turkey, where a switch to a presidential system was put to a referendum. The case is, therefore, instructive in setting out conditions that might make the application of Article 3 to both referendums and presidential elections defensible.

Lessons from Turkey
For a long time, reforming the 1982 Constitution had been high on the Turkish agenda. Following the general elections in June 2015, the ruling AKP party had made establishing an executive presidency central to its campaign promises. After the failed coup of 15 July 2016, the president introduced a state of emergency and proposed to extensively amend the constitution in order to adopt what the Turkish authorities described as a "Turkish-style" presidential system. 94 The constitutional amendment facilitated a transfer of competences from the legislature to the president. It abolished the post of the prime minister and enabled the directly elected president to do a range of things, including to appoint cabinet members without any parliamentary oversight or approval, to rule by default through executive decrees, and to declare an emergency. 95 Additionally, it gave the president control over the judiciary at the expense of the legislature. For instance, twelve members (out of fifteen) of the Constitutional Court and one-fourth of the top administrative judges would be appointed by presidential decrees under the envisaged system. 96 This constitutional amendment failed to get the backing of two-thirds of the MPs in parliament and hence the proposal was submitted to a referendum. 97 Shortly before the referendum, the Monitoring Committee of the Parliamentary Assembly of the Council of Europe requested the opinion of the Venice Commission on the proposed constitutional amendments. The Commission delivered its report a month before the referendum took place. It pointed out that the referendum "is taking place during the state of emergency when very substantive limitations on freedom of expression and freedom of assembly are in force. In particular, the extremely unfavorable environment for journalism and the increasingly impoverished and one-sided public debate that prevails in Turkey at this point question the very possibility of holding a meaningful, inclusive democratic referendum campaign about the desirability of the amendments." 98 Irrespective of this preliminary report, the referendum was held, and 51.4 per cent of the Turkish electorate voted to switch to a presidential system. The OSCE observed the referendum process and subsequently published a report. The report held that the legal framework 93 X v. the UK concerned the referendum on whether the UK should remain in the EEC; Nurminen and Others v. Finland concerned Finland's accession to the EU; Z v. Latvia concerned the referendum on Latvia's accession to the EU; Niedzwiecz v. Poland concerned the referendum on Poland's accession to the EU; Moohan and Gillon v. the United Kingdom concerned the Scottish secession referendum. 94  was "inadequate for the holding of a genuinely democratic referendum," 99 "fundamental rights were restricted by extraordinary state of emergency powers" and that the referendum "took place on an unlevel playing field and the two sides of the campaign did not have equal opportunities," 100 with the result that the Turkish electorate were not adequately informed on the matters to be decided.
One of the most controversial issues of this process was the decision of the Turkish National Electoral Commission, delivered 15 minutes before the end of the voting, to count ballot envelopes and ballot papers that lacked the obligatory seals as valid. Since this decision could not be appealed in the domestic legal order, it was challenged by the Republican People's Party before the European Court of Human Rights. 101 The applicant party argued that the Electoral Commission had delivered a decision that completely disregarded Article 3 of the Convention.
Scholars also presented arguments supporting the claim of the applicant party. Constitutional law professor Ali Acar was of the view that when an authority in a member state (in the given case, the Turkish election authority) unilaterally extends the scope of the right to free elections protected by the Convention, the Court should be able to consider this interpretation. 102 Riza Türmen, a former judge of the European Court, presented another argument. He referred to the McLean and Cole case, in which the Court declared the two applications inadmissible because they were, among other things, about a referendum, which was not considered to be elections to the "legislature" within the meaning of Article 3. Türmen argued that the Court's reasoning in this case, namely that "there is nothing in the nature of the referendum at issue in the present case which would lead the Court to reach a different conclusion here," 103 suggested the possibility of applying Article 3 if the referendum is different in quality and has a significant impact on the composition and power of the legislature. 104 Although both the Venice Commission and the OSCE found that the holding of the Turkish referendum was, in many respects, contrary to good international practice, the Court rejected the applicant's claim solely on the grounds that the referendums do not come under the scope of Article 3. 105

Applying Article 3 to Presidential Elections and Referendums
As the Turkish case demonstrates, a shift from a parliamentary democracy to a massbased executive-centered plebiscitary regime may happen through a referendum. Often, an essential step in this shift is a presidential election where the executive is directly elected by a popular vote. Indeed, not all presidential systems are per se contrary to Article 3. This only applies in a Schmittian mass-based executive-centered plebiscitary regime, where the president, being directly elected, claims to be a direct representative of the people and exercises law-making power by issuing decrees 99  without parliamentary authorization or abrogating statutory measures. This sort of regime is not a thing of the past. Today, Schmitt's hostility to parliamentarism and conflict resolution by negotiation 106 seems to be all the rage. His approach to democracy attracts many contemporary autocratically inclined leaders, who take advantage of his ideas. 107 And as the Turkish example illustrates, there are even experiments in which politicians have attempted to make the Schmittian alternative to parliamentary democracy a reality.
Even though a systemic shift may be institutionalized through presidential elections and nationwide referendums, the Court's process-based review does not include these institutions; it only covers parliamentary elections and parliamentary procedure. How can the Court change its interpretative practice to broaden the scope of Article 3?
One way to do so would be for the Court to stress the importance of domestic democratic processes under Article 3. It could argue as follows: Since member states are self-governing polities, the Court respects domestic democratic processes 108 by using the doctrine of the margin of appreciation, 109 which allows the member states' democratic institutions to consider their domestic circumstances. The margin of appreciation is the means by which the Court defers to domestic democratic processes within member states. 110 However, as Jean-Paul Costa put it in his 2008 speech, "as we approach the core operation of democracy, such as the right to vote and the right to form political parties or the right to participate in free political debate, so the margin of appreciation contracts almost to a vanishing point." 111 The Court may defer to domestic processes as long as the member states do not violate the underlying principle, that is, parliamentary democracy. What it may not do is defer to the basic domestic process when that very process is under attack 112 or already captured by anti-democratic forces. Put another way, the Court may only defer to domestic processes within otherwise essentially democratic states.
In order to protect democratic processes, the Court should recognize that Article 3 establishes a very clear connection between law-making and basic democratic voting processes. Article 3 thus requires member states to be committed to parliamentary democracy when structuring and organizing basic democratic processes that affect legislative matters or institutions enjoying legislative power. Free and fair parliamentary elections are essential voting processes, but they are not the only ones. For instance, if the president is elected directly by the people, the same question regarding the method used to elect the president as applies in a parliamentary election would apply here, namely, whether the opinion of the population has been adequately represented. Likewise, a referendum that decides on legislative matters or determines the type of legislature in a given country also raises similar concerns from the point of view of representation.

Presidential Elections
In its judgments concerning presidential elections, the Court has interpreted the word "legislature" in light of the constitutional structure of the member state. The Court has reviewed the domestic constitutional rules defining the extent and nature of the head of state's involvement in the legislative process 113 and the strength of the link between presidential powers and the legislative process. 114 Consequently, Article 3 may apply to presidential elections if "the Head of State has been given the power to initiate and adopt legislation or enjoys wide powers to control the passage of legislation or the power to censure the principal law-making authorities," 115 because, in that case, the office of the head of state could arguably be considered to be a "legislature." So at least theoretically, the organization of voting in a presidential election can count as a democratic process under the Convention. Nevertheless, the Court has not to date examined a case regarding whether a presidential election complies with Article 3.
One of the reasons might be the difficulties arising when examining the constitutional structure of a member state. Such assessments require the Court to be utterly familiar with the almost fifty different constitutional structures. Given the Court's subsidiary role, it would be more fitting for it to characterize the concept of the "legislature" as autonomous, while it would also help the Court protect domestic parliamentarism. If the Court did offer an autonomous meaning of "legislature," it would be opting for a functional approach: this would involve understanding the "legislature" as the rulemaking parliament and organs other than parliaments to the extent that they possess rulemaking power. There were some indications in the early case law that the Commission had adopted this approach; for instance, it considered whether a county council possessed "an inherent primary rulemaking power." 116 Following this example, the Court may examine whether the directly elected president has a substantial role in the legislative process by assuming the powers and functions of national parliaments (e.g., the power to initiate legislation or to adopt general legal measures).

Referendums
Another consequence of adopting such an argument would be that a referendum that decides on legislative matters or determines the type of legislature a country would have would also be subject to the Court's procedural review. There are striking examples of how international bodies may review domestic referendums. For instance, in the Gillot 117 case, the United Nations Human Rights Committee expressed the view that Article 25 of the International Covenant on Civil and Political Rights on the right to vote applied to referendums. The case was about referendums on self-determination in New Caledonia. Gillot and twenty other French citizens who did not meet the length-of-residence requirement claimed that they had been discriminated against based on their ethnic origin because of how persons eligible to vote were determined. The committee examined the criteria used by the authorities and did not find a violation of any covenant article, because the criteria differentiated between residents as regards their relationship to the territory based on the length of the residency requirement, whatever their ethnic or national origin. 113 Boskoski v. the FYROM, para 1 114 In the Guliyev case the Court held that the president's powers are only "accessory" to the parliament's legislative power. Article 25 Section (a) of the covenant ensures every citizen's right and the opportunity "to take part in the conduct of public affairs, directly or through freely chosen representatives" and "to vote and to be elected at genuine periodic elections." By contrast, Article 3 requires the holding of regular elections to the legislature and does not mention direct participation by citizens, let alone the possibility of holding a referendum.
Today, the only rationale for excluding the Court's involvement in referendum cases is rooted in a literal reading of Article 3. 118 As the UK Supreme Court judge Lord Neuberger insisted in the Moohan case, the "word 'elections' is not a word that naturally covers a referendum, which does not involve electing anyone to any post," and it would be absurd to argue that governments have an obligation to hold referendums "at reasonable intervals." 119 While it would indeed be absurd to argue that governments have such an obligation, there are alternative arguments for the inclusion of a referendum under Article 3. For example, one may argue that the Convention presupposes the right to self-determination. If the population is called upon to take a decision itself, it must be entitled to the benefit of the same guarantees as are applicable regarding the choice of its legislature. 120 One may also argue that, although today no right to a referendum may be derived from Article 3, 121 the provision should be interpreted as ensuring this right. This paper suggests adopting another argument.
In Council of Europe member states, the main rule is the exercise of power through representation. The competence of the representative bodies is complete and general. Compared with this, the direct exercise of power by the people, usually through referendums, is exceptional. The member states have a wide margin of appreciation in regulating and holding referendums. There are countries in Europe with only minor elements of direct democracy (e.g., Germany). In other Council of Europe member states, however, referendums are part of domestic institutional frameworks and political practice (e.g., Switzerland). Here, occasionally, people are allowed to decide directly on specific issues that otherwise fall within the competence of the legislature. Should this be the case, the legislature is obliged to ensure that the referendum result is implemented. Furthermore, the legislature should also refrain from any act that would influence or frustrate the realization of a referendum.
Theoretically, the direct exercise of power in a parliamentary system can only be complementary: the referendum can complement and influence the exercise of representative power. Nevertheless, in some member states, a referendum itself can be legislative in the sense that it may change the laws of the country and perhaps even the constitution. And in some cases, the referendum results may even determine the type of legislature the country would have, as happened in the Turkish case. Thus, once domestic law provides for the process of a referendum, there is no reason to exclude ab ovo the review of this exercise of public power from the remit of the Convention. Should the Court recognize a referendum as a democratic process under Article 3, this would not mean that any referendum would automatically fall within the Convention's remit. On the contrary, this would open the way for the Court to review a claim that concerns a referendum that may clearly and substantially affect the situation of parliamentary democracy in a member state. 118 Schabas 2015Schabas , 1021 Moohan and another v. The Lord Advocate, Judgment of the UK Supreme Court of 24 July 2014, para 44-45 120 De Meyer 1993, 556 121 Grabenwarter 2014 During this review, the Court might consider the argument that a referendum, as one of the many tools of direct democracy, is an expression of the sovereignty of the people and that, in general, it aims to serve democratic purposes. However, the Court might also consider that referendums might be fundamental contradictions of parliamentary democracy. Unitary sovereigns may come to power by direct appeal to the people, and once they are in power, the practice of referendums remains in place. For instance, in the Schmittian model of a massbased executive-centered plebiscitary regime, unitary sovereigns test their popular support by organizing "referendums." And very often these referendums result in overwhelming majorities voting in favor of the executive. So, as Max Weber pointed out, the logic of the referendum can be to "legitimate executive domination" 122 and recent European examples show that even in established democracies, a government-led referendum may increase "executive power, as the government seeks the freedom from Parliament and legal constraint in order to implement what it sees as the popular will." 123 Furthermore, national referendums may take place under conditions that are usually less democratic than general elections. For instance, very often, referendums do not come at the end of a complex deliberative process; instead, the questions are pre-set, the procedures do not meet the test of open deliberation, and the people only enter at the very end of the process. Indeed, a referendum can at best only state a decision of principle; it cannot determine a course of public action. However, the results of these referendums are, in most cases, difficult to change and usually irreversible. 124 The political context often makes it impossible to go against the referendum result, particularly when the executive itself has initiated the referendum.
Recognizing presidential elections and referendums as democratic processes under Article 3 would mean understanding that the aim of the Convention mechanism built upon Article 3 is to protect parliamentary democracy and to ensure that when organizing voting processes, this principle is not arbitrarily abridged by national governments. In what follows, the paper looks through existing supranational review procedures concerning referendums and presidential elections and reveals whether there are gaps in the Convention's rights protection mechanism.

Existing and Lacking Supranational Review Procedures
Prior to a referendum, Council of Europe advisory bodies may review the question to be put to a referendum. For instance, the Venice Commission may examine the conformity of the question submitted to a referendum with its Code of Good Practice from a procedural and a substantive point of view. 125 The former concerns the wording and content of the question. Under substantive validity, this Code understands the requirement that the referendum question should comply with all the superior law, including international law and the Council of Europe's statutory principles: democracy, human rights, and the rule of law.
Thus, the prior review process exercised by the Council of Europe's advisory and monitoring bodies already exists. In addition, a subsequent review procedure is also available before the Court. European citizens may challenge the outcome of a nationwide referendum if the outcome already constitutes part of domestic law. In this sense, a referendum is similar to parliamentary decision-making because the democratic process may improve the prospects that an outcome will be rights compliant, but the process itself does not make the outcome per se rights compliant. Both the parliamentary process and the voting process of a referendum can be defined in contrast to the outcome: they occur in different forms and consist of a range of activities.
Nevertheless, while the Court often assesses the quality of the parliamentary process underlying the impugned legislation, 126 it has indicated its reluctance to conduct a subsequent procedural review of the domestic voting process in a referendum. The Venice Commission may provide a non-binding opinion, as it did after the controversial 2014 referendum on Crimea, which was not internationally recognized. 127 However, only the Court can be the venue for individuals to contest their rights to take part in referendums or presidential elections. In order for the Court to be able to undertake a subsequent procedural review, a shift in interpretation would be required. The Court should recognize that the very same procedural conditions should apply to voting in parliamentary elections, voting in presidential elections, and voting in referendums. This paper calls for procedural conditions that require procedural fairness in every decision on the right to vote.
First, the most fundamental procedural condition all these three types of voting processes must meet is that members of the electorate should be treated as free and equal persons by the state authorities. The basic election principles-equal, universal, and free suffrage and the secret ballot-are derived from that fundamental requirement. Any vote that is held in the territory of the member states must be held freely and fairly; this applies to the organization of the vote, the counting of the votes, the recording and the transmission of the results, and the effectiveness of the appeal system. Any voting process must also ensure that voting rights can be accessed and exercised; this includes questions of who is entitled to vote, questions of universal suffrage and the secret ballot (or the lack thereof) as well as questions related to campaigning and organizing the vote. There are concerns about discrimination in connection with all types of voting processes: for instance, the problem that a polling station might not be accessible for voters in a wheelchair, 128 the prohibition of national minorities from taking part in political life, 129 or the loss of voting rights on account of placement under partial guardianship. 130 In the Moohan and Gillon case, the rules that prevented convicted prisoners from voting in the Scottish independence referendum were challenged, albeit unsuccessfully, before the Court. 131 Similarly, the Court refused the complaint of Liucija Baskauskaite, who was deprived of the possibility of being a candidate in the Lithuanian presidential elections. 132 Second, the issue of the inclusiveness of citizens' participation is also crucial in any voting procedure. 133 When organizing a vote, member states are responsible for ensuring that citizens have a full participative role: i.e., that they have access to the relevant information in order to form an opinion freely without any pressure. 134  information of all kinds during a campaign period 135 is relevant when voting in an election or a referendum. That means that campaigns should treat competing perspectives fairly and that unbiased public information materials should be provided. The Court's view on the member states' positive obligation is also pertinent: under Article 3, states must ensure that coverage by regulated media is "balanced and compatible with the spirit of 'free elections. '" 136 Overall, the legitimacy of the voting process in parliamentary elections, presidential elections, and referendums would depend on how these basic procedural conditions are observed. One of the main advantages of adopting a purposive approach to Article 3 to cover referendums and presidential elections would be to enable the voting procedure and the outcomes of these votes to be accepted as legitimate, both within the country and abroad.

Conclusion
Article 3 is a unique provision of Protocol 1 of the Convention. It is formulated as an undertaking on the part of the member states to hold elections. The wording is the product of a particular set of historical circumstances, and the difficulties of the wording have had consequences for its application. This paper discusses Article 3 in a number of different ways. First, the paper draws on knowledge of the drafting debates to understand the principle underlying this provision. The preparatory works demonstrate that although the text of Article 3 was watered down during the drafting process, the framers shared the same systematic ambition: to maintain parliamentary democracy through the protection of the democratic process and prevent systemic changes. Then, the paper analyzed, with reference to the case law, the Court's engagement with the concept of democracy. It found that the Court's model of democracy is built on pluralism and parliamentarism and supports all modern democracy theories that take these principles seriously. However, this excludes the monolithic version of democracy as famously developed by Carl Schmitt.
The Court's model has the potential to provide a robust account of parliamentary democracy, but the case law built on Article 3 is not yet sufficient. There are gaps in the Convention's protection system. Currently, the Court does not recognize presidential elections and referendums as democratic processes of domestic decision-making, even though, as the example of Turkey demonstrates, a systemic shift from a parliamentary democracy to a mass-based executive-centered plebiscitary regime can be institutionalized through a referendum. Likewise, a transfer of power may happen through a presidential election.
The paper has proposed reading Article 3 as a safeguard for parliamentary democracy that not only focuses on parliamentary elections but also on referendums and presidential elections directly and materially affecting the situation of parliamentarism in a member state. It has demonstrated that the only reason for not involving the Convention mechanism in these cases is the literal understanding of Article 3. However, if the underlying principle of Article 3 is parliamentary democracy, it is reasonable to accord protection to all domestic voting processes that clearly and substantially affect the situation of parliamentarism in a member state. As the historical account indicates, the original adjudication mechanism of the Council of Europe assumed that parliamentary democracy might be endangered. Hence, applying Article 3 to presidential elections and referendums would not seem out of step with the original mission of the Court, which was to protect the already existing domestic parliamentary systems and democratic governments from sliding into non-democracy. On the contrary, it would be a natural development of the case law if the Court provided an autonomous meaning of the concept of the "legislature" and adopted a functional approach by understanding the "legislature" as an organ possessing rulemaking power. This "organ" may be a directly elected president assuming legislative powers or a referendum deciding on legislative matters or determining the type of legislature a country would have. The Court would then be able to conduct full-fledged judicial inquiries into an allegation of non-compliance with the Convention and discredit measures that betray the principle of parliamentary democracy.