Regulating Parties by Constitutional Rules in Liberal Democracies

Abstract When establishing constitutional rules that regulate political parties, liberal democracies struggle between civil liberties—thus tolerating anti-democratic parties—and potential threats of democratic breakdown, which can be reduced by prosecuting and prohibiting anti-democratic parties. We suggest that liberal democracies must balance false positives and false negatives by combining ex ante and ex post regulatory mechanisms. By making use of a unique dataset of thirty-seven liberal democracies collected by the authors, we find empirical results consistent with our positive theory. An extensive review of the normative debate and case law provides additional qualitative support.


A. Introduction
This article offers empirical, positive, and normative arguments on the regulation of political parties.There is surely vast literature from the standpoint of election law in political science and in comparative constitutional law.As there is surely vast literature on political parties, their campaigning, financing, and internal organization in comparative politics.However, apart from studies on specific party ban practices, 1 the comparative literature about modes of regulating political parties, 2 the main institutional actors of elections, is less developed. 3Modes of regulation of political parties is usually a footnote in comparative election law debates, perceived to be on the fringe of a minor issue.We attempt to bridge the divide between the hitherto disparate literatures.
The concern about political parties undermining democracy while tolerated by democracy is not new.Historically, the scrutiny has been on fascism before and during WWII as well as communist parties during the Cold War.Lately, the disquieting rise in populism on the right and on the left in Europe, the US, and in other areas of the globe has reintroduced the debate.This is, therefore, an auspicious time to get a handle of the state of the debate.There are so many important questions: Are and should political parties be mentioned or protected by the

B. Normative Theory of Party Regulation
Regulating political parties is a perilous exercise. 9The experiences of the World Wars brought political and legal consequences regarding political parties and the envision of democracy per se.One of the most vicious faces of Nazism, Joseph Goebbels, ridiculed the Weimar legal arena, while declaring that "[t]his will always remain one of the best jokes of democracy, that it gave its deadly enemies the means by which it was destroyed." 10The first president of the Spanish Constitutional Court epitomizes the democratic state of the last century as the "state of parties," in his book by the same title. 11Still, democracy is a difficult concept to define. 12Moreover, there is uncertainty regarding what constitutes an anti-democratic party. 13

I. Democracy and Its Enemies: The Militant Democracy
All political parties aim at conquering power. 14However, some use the legal arena to later fulfil their illiberal or antidemocratic agendas.In the first half of the last century, Karl Löwenstein wrote his famous "militant democracy" (wehrhafte/streitbare Demokratie) 15 approach to fascism, arguing that "fire should be fought with fire." 16In this sense, legal orders should ensure that the enemies of democracy will not be able to "exploit the freedoms inherent in democracy." 17omewhat ironically though, the idea of militant democracy has roots on Carl Schmitt's constitutional theory, when he addressed the notion of a "constitutional core." 18But it goes way back in history.The French revolutionary Saint-Just wrote the famous expression "pas de liberté pour les ennemis de la liberté" (there is no freedom to the enemies of freedom). 19How can the ultimate democratic goal be ethically defended through undemocratic means?Such enthralling paradox is hard to unveil. 20et, Löwenstein was not alone in his militant war of ideas before and during WWII.Joining the "never-again" mentality, Karl Mannheim offered his sociological diagnosis in the reflection "The Third Way: A Militant Democracy." 21As he clearly wrote, that "the principle of laissez-faire will not help us any further" 22 and that "our democracy has to become militant if it is to survive." 23öwenstein and Mannheim were two German scholars forcibly exiled in the United States of America for their Jewish origin.These famous pieces of "literature in exile" revealed that several intellectuals were not aligned with the Nazi world at home. 249 Zachary Elkins, Militant Democracy and the Pre-Emptive Constitution: From Party Bans to Hardened Term Limits, 29 DEMOCRATIZATION 174, 183 (2022).
According to John Rawls, "an intolerant's freedom can be restricted ultima ratio, when there is real danger to 'the institutions of liberty.'" 25 Notably, it was Karl Popper that addressed his famous paradoxes of freedom, tolerance, and democracy.Under his understanding, unlimited tolerance to those who are intolerant will ultimately destroy not only the tolerant, but also tolerance itself. 26o, where does this leave us?Will militant democracy hinder or strengthen democracy itself?Does militant democracy taper democracy to a specific orientation?Would it mean banning political parties?One can certainly expand militant democracy beyond the strict domain of political parties. 27Both emergency clauses 28 and eternity clauses 29 belong to the "constitutional arsenal" of militant democracy. 30The militant democracy approach is visible today outside the confinements of political organizations.Taking this idea even further, András Sajó considers that wearing religious garbs in public creates the same emotional mechanisms of paramilitary uniforms. 31istorically-speaking, jurisdictions that have not experienced totalitarianism and authoritarianism might raise some eyebrows and be warier of the militant democracy motto. 32In fact, American scholarship considered the concept of militant democracy one of the most "startling aspects" of European constitutionalism. 33One possible explanation for such wariness could be that the original German expressions wehrhafte/streitbare Demokratie were translated to English as "militant democracy" instead of being more literally translated to "defensive democracy."If, on the one hand, the word "militant" is more captivating, on the other hand, though, it can be perceived as politically biased or politically compromised.
At a more fundamental level, many noted that the concept of militant democracy for a long time lacked consistent theorization. 34However, recent scholarly efforts tried to densify it in the context of the contemporary constitutional theory. 35To Müller, militant democracy is the preemptive willingness to adopt "prima facie illiberal measures" to thwart the destruction of the democratic regime. 36To put this another way: The idea of militant democracy is that in order to play the game, political parties need to stand by some constitutional principles.In the aftermath of WWII, democracy can no longer be relativist or politically agnostic, as it ought to resist totalitarian attempts. 37That is to say, if fascism employed "emotional mobilization" as opposed to reason, militant democracy opposes such emotionality. 38n the other side of the spectrum, others fail to see the validity of militant democracy in contemporary societies.The concept of militant democracy raised acute critique, mainly due to its exceedingly legalistic reaction to extremism. 39As we will see, there is some doctrinal support for this perspective.
Conceptually, militant democracy might be labelled as paternalistic, because it treats the electors as politically unqualified and eliminates public discussion of all political perspectives. 40ilitant democracy brings a "re-politicization of the question of membership in the demos." 41or the sake of this argument, it diminishes the people, who are left "deprived of effective citizenship." 42his idea resonates with the call for better arguments. 43Bearing this in mind, instead of excluding problematic parties, some argue that they should be welcomed in the democratic arena, with hopes that they will become more moderate. 44f some believe that the concept of militant democracy can be invoked in a non-arbitrary way, 45 others argue for "an irreducible element of arbitrariness" when deciding about the enemies of democracy. 46Such arbitrariness might even be biased or selective, purposely excluding from the democratic game target political competitors. 47Accordingly, the danger of a cascading effect rises, because banning could be used to tackle any expression of dissent. 48rom the beginning, strong reasons grounded within philosophical liberalism advocate against outlawing political parties only on the basis of their ideology, unless they commit specific criminal acts punished by criminal law. 49Issacharoff, while acknowledging the need for self-preservation, stresses that beforehand the most significant thing is to build strong institutional protections. 50urthermore, many claim that the concept of militant democracy was narrowly designed as a response to fascism, which no longer comprises a critical threat. 51To Müller, militant democracy cannot be legitimated or normalized, as it requires the use of illiberal measures. 52Quite paradoxically, democracy while fearing its death, would commit suicide.In other words, to defend itself, democracy betrays its own foundations. 53till, others consider either adhering to the original militant democracy thesis or to other variations, that some democratic defensiveness is needed.To borrow Barak's thought, "defensive democracy: Yes; uncontrolled democracy: No." 54 Building on Sajó's distinction between militant democracy and "militant constitutionalism," 55 Gutmann and Voigt pushed this line of thinking even further, arguing that militant constitutionalism is a "constitutional design feature that aims at safeguarding constitutionalism."To them, militant democracy is a distinct concept, as it operates in a preventive way and merely addresses the government. 56With a very interesting argument, Elkins states that militant democracy takes "the low road," as it preserves democracy while engaging in anti-democratic measures.Instead, constitutionalism takes "the moral high road" as it adheres to "a higher democratic commitment." 57ome claim that "defensive democracy," which is inspired by the idea of self-defense against democratic threats, 58 is a more comprehensive concept when compared with "militant democracy." 59Others downplay the relevance of such distinction, because both terms are used to portray anti-democratic actions aiming to combat threats to democracy itself. 60The stakes in this debate then are high.Doomen, after rejecting the concept of militant democracy, 61 pivots attention to a "mitigated democracy," which consists of protecting the rule of law and certain rights "against democracy." 62Likewise, Schupmann advocates a "constrained democracy," which is the adoption of constitutional mechanisms such as eternity clauses that prevent democratic parties from amending liberal constitutionalism out of the constitution. 63Emphasizing the social dimension, Malkopoulou and Norman reject militant democracy's "anti-participatory and elitist logic" and instead merge proceduralism's adherence to dissensus with a social-democratic logic in the design of democratic constitutions. 64hile some scholarship equates democracy with liberal democracy, 65 others stress that democracy does not equal liberal democracy.Following Joseph Raz, 66 some scholars remind us that "not every country that can plausibly advertise itself as a democracy is a liberal democracy." 67esistance to dictatorship and fear of historical setbacks gave rise to concepts such as militant democracy, in the period of the World Wars, or of a "limited democracy" (democradura) 68 or "constitutional patriotism," 69 in the transitioning democracies of the end of the 20 th century.
Historically, restrictions on the freedoms of individuals or groups took several forms and appeared in the form of hybrids, transitioning or even façade democracies such as "illiberal democracy," 70 "protected democracy," 71 or "tutelary democracy." 72However, militant democracy bears distinctive traits.In fact, while militant democracy was an attempt to strengthen the democratic core values, the abovementioned subtypes diminished them.Moreover, militant democracy "has all the mobilizational power of a vanguard movement." 73n either event, contemporary scholarship that revisits and enriches the theoretical grounds of militant democracy offers new perspectives.Whether reconceptualizing militant democracy "as a matter of degree" or replacing the label of militant democracy with "new vocabulary," the discussion is very much still alive. 74As Tyulkina advocates, militant democracy "is not an isolated, old-fashioned, abstract idea from postwar Europe." 75n sum, untangling the democratic paradox within the penumbras of populism is a difficult exercise.Some argue for a more robust conception of deliberative democracy, 76 while others substantively narrow militant democracy to "help attain an intermediate end." 77Be that as it may, cautiousness is recommended.

II. Banning Political Parties
It is undeniable that political parties are crucial to the democratic system. 78Consistent with this general understanding, if political parties are leading figures of politics, then they carry enough significance to justify legal regulation by the state. 79Still, at its core, it is of paramount importance not to confuse parties and the state. 80From a purely pragmatic perspective, at best, party bans are the most visible form of preemptive defense to the political rights of future generations. 81At worst, bans are "the mark of tyranny." 82ccording to Niesen's definition, party bans refer to "all juridical forms that effectively prevent the founding and continued operation of political parties, whether in the form of dissolution, 68  substantive registration requirements, temporary suspension, or prohibition of and prosecution for party formation." 83Proscribing political parties is the most aggressive measure available 84 and will have key implications for "electoral stability, electoral and parliamentary fragmentation, and government formation." 85otwithstanding the possible harm to the overall quality of the democratic system, 86 dissolution of political parties collides with political competition and the right to free participation in the political arena. 87Furthermore, it is inevitable to conclude that the decision as to what truly threatens the democratic order "is necessarily an exceptional decision." 88hy should a democratic government act in a "totalitarian way?" 89 The main effects of party banning were addressed in the available literature. 90One possible effect is the further radicalization of the banned party. 91Proscribing a certain party might romanticize their cause or even trigger the martyr effect. 92Accordingly, the party ban would be a short-term measure and the party could even re-emerge under a distinct label. 93The protection of minority rights and democratic pluralism is also a preoccupation, if minority parties are targeted. 94o Fox and Nolte, when considering party bans, legal traditions are relevant.They distinguish between 'tolerant' and 'intolerant' democracies, focusing on whether a given constitutional tradition is a procedural democracy or a substantive democracy.On the one hand, a procedural democracy will prioritize a party's external respect for democratic rules, notwithstanding its ideology/practice. 95o hinder the "danger of democide," procedural models hinge on institutional checks and balances and sanctioning criminal acts of leaders and members of antidemocratic parties. 96On the other hand, substantive democracy also focuses on the party's ideology/practice being democratic. 97alkopoulou proposes to rename the procedural model as "the criminal model," as the expression "procedural" might ignore the substantive values. 98Bourne built on a third element to this distinction: Whether party proscriptions only boycott "anti-system behavior," such as association with violent groups, or if it also sanctions "anti-system ideology." 99To give an example, if the American model allows undemocratic ideas, the German model sanctions undemocratic ideas. 100 democracies "are more likely to be 'tolerant democracies' that not only eschew party bans but avoid using existing legal provisions to ban parties." 101ersel distinguishes the "external democracy" of political parties from its "internal democracy." 102In this sense, external democracy reflects the relation between parties and society or the state.This is the more visible part when courts consider proscribing political parties.By contrast, internal democracy relates to the organization of the party itself.A party can profess democratic values but, at the same time, impede the replacement of the party leadership. 103Should internal democracy mimic the external form?Although Mersel argues that internal democracy should be as relevant as external democracy when assessing party proscription, when the internal part functions are at stake, he sustains that only the "essentials of democracy" must be imposed on political parties. 104Such "minimum democracy" requirements relate with the representation, participation, and the liberty of individual party members. 105In these situations, parties enjoy a wider margin of discretion and courts should intervene only as a last resort. 106hen viewed more closely, there is a clear evolution on the reasons for justifying party banning.Examples include, but are not limited to, the following: (1) If, during WWII, party banning focused on lato sensu fascist parties, including Nazi or Nazi-inspired parties; (2) if, during the Cold War, it focused on communist parties; and (3) if, following the 9/11 terrorist attacks, it focused on religious extremism. 107In the current times, apart from technical reasons-such as failing to comply with membership or financing rules-parties can also be banned for substantive reasons.The main substantive reasons are the following: Endangering the democratic and secular foundations of the state, its territorial integrity, as in Turkey, engaging in a violent or racist agenda, as in the Netherlands, or neglecting democratic internal procedures, as in Portugal or Spain. 108o sum up, Bligh distinguishes what he calls the "Weimar" paradigm-which relates to the abovementioned Löwenstein's conception of militant democracy-and the "legitimacy" paradigm. 109If the former related to targeting anti-democratic ideologies per se, such as fascism or communism, the latter is narrower and threatens certain fundamentals within the liberal constitutional order, namely equality and non-discrimination, or secularism. 110he question that follows can be stated thus: In which situation(s) can a party be prohibited?There appears to be a straightforward answer to this question.As Casal Bertóa and Bourne bluntly put it, the effective application of ban provisions is "relatively rare." 111 Because the prohibition of political parties ought to be truly exceptional, then its misuse or instrumentalization should be prevented. 112As we will address in Section D, the European Court on Human Rights (ECHR) employs self-restraint techniques when deciding on the dissolution of political parties.In fact, activities disputing the organization of the state can be tolerated, inasmuch as they do not undermine democracy itself. 113

III. Democratic Backsliding
In the contemporary era, the fear of democratic backsliding anchors on the fact that some constitutional democracies are being threatened by illiberal movements. 114As Ginsburg and Huq note, as the three main predicates of democracy-rule of law, competitive elections, and rights to free speech and association-decline, the danger of democratic reverting becomes a reality. 115ixon and Landau imagery of "abusive constitutionalism" also entails the idea of a "minimum core" of rights and institutions "necessary for a true constitutional democratic order". 116urthermore, towering social and economic inequalities, toxic polarization, underrepresentation of minorities and women, religious fundamentalism, xenophobia, and social unrest are caveat signs towards gradual democratic erosion.
In this sense, democratic backsliding assumes many forms.On the one hand, it can be triggered by a sudden event, such as a coup d'état or backsliding "by surprise". 117On the other hand, it can be the outcome of more discrete and gradual measures, such as legal reforms, or even derive from a level of democratic neglect that ultimately will unravel liberal constitutional democracy. 118everal scholars have raised convincing democratic concerns about populist rhetoric and "executive aggrandizement". 119In several states, such as Russia, Turkey, Hungary, Poland, Venezuela, Bolivia or Brazil, just to name a few, such aggrandizement aims at circumventing democratic rotation and democratic opposition. 120Thereby, reducing judicial independence, controlling the media, tilting electoral formulae, restricting the activities of the NGOs, censoring dissenting voices in the academia, all steadily contribute to backsliding. 121ome argue that a good constitutional design or strong forms of judicial review can mitigate constitutional erosion. 122The breakdown of democracy and the upsurge of autocracy is a threat to constitutionalism itself. 123Furthermore, the COVID-19 pandemic escalated both the number and quality of threats to liberal democracy worldwide. 124According to V-Dem, the level of democracy enjoyed by the average global citizen in 2021 is down to 1989 levels/ Hence, the last 30 years of democratic advances are now eradicated.The number of liberal democracies is down to 34 in 2021.There have not been so few since 1995, over 26 years ago.Closed autocracies are up from 25 to 30 between 2020-2021.Electoral autocracy remains the most common regime type in the world, as adopted by 60 countries.Together, autocracies now harbor 70% of the world population, approximately 5.4 billion people. 125ch a scenario is worrisome.As Scheppele wrote, constitutions cannot defend themselves. 126he weakening of the separation of powers can only be curtailed by a solid political, cultural, and social defense of constitutional democracy.

C. Positive Theory of Party Regulation
Any form of regulation or legal intervention must address and balance false positives-mistakenly acquitting or not sanctioning noncompliant individuals-and false negatives-mistakenly sanctioning compliant individuals. 127False positives and false negatives are necessarily costly.Unfortunately, enforcement technologies are imperfect and have not been able to eliminate false positives and false negatives in most human activities regulated by a state.However, one needs to emphasize that false positives and false negatives are not necessarily equally costly.In fact, this cost asymmetry inspires and divides many debates in the literature on regulation. 128

I. Dealing with False Positives and False Negatives
One possible framework is to approach regulation as a set of policy mechanisms to minimize the joint costs of false positives and false negatives.Therefore, policies will be prone to sanctioning, for example, depending on whether the cost of false positives are less or more significant than the costs of false negatives.At the same time, regulatory policies may reflect the extent to which there is an intrinsic trade-off between these two costs, which may or may not exist depending on available technologies and specific institutional contexts.
How can we apply these general insights from the regulatory literature to how democracies regulate the activity of political parties?Table 1 summarizes the challenge faced by liberal democracies under this framework. 129Ideally, anti-democratic parties would be fully restrained while democratic parties would not get sanctioned.The main diagonal in Table 1 indicates correct policies given the nature of political parties.However, the minor diagonal in Table 1 identifies the possibility of false positives and false negatives.False positives refer to not restraining antidemocratic parties-which can result in a democratic breakdown.False negatives take place when restraining democratic parties-with excessive limitations to freedom of association.
In a liberal democracy, unlike authoritarian regimes, false negatives are more costly than false positives.In other words, punishing democratic political parties engaged in lawful and compliant activities-but misunderstood otherwise-is socially and politically more costly than not punishing undemocratic parties or parties that promote anti-democratic activities.Therefore, there is a deep asymmetry in terms of regulatory burdens that shifts liberal democracy in the direction of tolerating more false positives than false negatives.We use the term liberal democracy in a broad sense.We acknowledge, though, the more fine-tuned distinction between liberal democracy and electoral democracy.See LARRY DIAMOND, DEVELOPING DEMOCRACY: TOWARD CONSOLIDATION 10 (1999).
It is important to emphasize that we are not suggesting that liberal democracies should seek zero false negatives and tolerate a plenitude of false positives.Depending on available technology and institutional contexts, a mix of presumably more false positives and presumably fewer false negatives is more desirable.Therefore, the expectation is that constitutional rules regulating the organization and activity of political parties reflect this tension in ways to accommodate the best possible mix.
For example, the most effective constitutional rule to eliminate false positives is prohibition of all political parties.Clearly such a rule violates the minimum standards of liberal democracy.Therefore, liberal democracies, unlike authoritarian regimes, are not expected to engage widely in prohibiting political parties.Now consider the opposite example.The most effective constitutional rule to eliminate false negatives is to shield from prosecution-and civil litigation-all political organizations for any sort of activity, including crimes such as murder, kidnapping, or money laundering.However, it is difficult to envisage a situation where a liberal democracy can tolerate such political organizations without embarking in serious (un)democratic consequences.Each liberal democracy, depending on available technology and institutions, which, in turn, reflects social and political preferences, must find a compromise between these two-obviously unrealistic-extremes.

II. Ex Ante Versus Ex Post Regulation
Following the seminal insights by Steven Shavell 130 in the 1980s, we can further investigate how liberal democracies address this tension between false positives and false negatives.Specifically, the appropriate mix can result from combining ex ante and ex post regulatory mechanisms which constitutional rules can codify. 131et us start with an ex ante mechanism, also called regulation in economic literature.Political parties are subject to a strict entry/formation ex ante control by some higher authority, such as an administrative agency, a court, or an independent regulator that is independent from the "regulated industry".In practical terms, there is some sort of licensing procedure by which a political party needs to satisfy some constitutional and legal requirements before being granted legal status.Now consider an ex post mechanism, called litigation in economic literature.In this case, political parties are subject to ex post legal proceedings by some higher authority, a criminal court, or a constitutional court, or even the executive and/or legislative branches of government.In this setting, there is no prior assessment or licensing of political organizations.They are simply liable ex post when violating constitutional rules or not complying with certain democratic principles.In a similar vein-although referring to ex post or ex ante party bans-see Kouroutakis, supra note 13, at 312-14.
It is intuitive that ex ante and ex post mechanisms prioritize distinct concerns.Suppose false negatives are much more costly than false positives.An ex post mechanism seems more appropriate.It is easy to assemble a political party and engage in political activity so that false negatives are avoided.Sanctions operate ex post to deter and punish anti-democratic conduct.Because all political parties are allowed and licensing is unrequired, it is possible that antidemocratic parties are formed and take a role in the public debate even within the institutions of liberal democracy.Therefore, ex post mechanisms emerge as the viable response to such possibility.
Consider, for sake of completeness, the opposite reasoning.False positives are more costly and constitute a threat to liberal democracy.An ex ante mechanism is more adequate.Licensing and some form of pre-approval are more effective in avoiding false positives.At the same time, as long as the ex ante mechanism is reasonably effective, there is a limited scope for ex post litigation or prosecution because all licensed political parties are democratic in their nature.
As we have hypothesized before, jurisdictions look for a mix of regulation, such as some form of ex ante control, and litigation, such as implementing ex post liability.Constitutional rules codify a response to the complex challenge of respecting full freedom of political association as part of civil liberties and addressing the political damage of tolerating anti-democratic parties.

III. First Versus Second Order Regulation
Recent scholarship on the regulation of political parties has made an important distinction between first and second order regulation. 132By first order regulation, one understands the constitutional norms that address party recognition and party bans.By second order regulation, legal scholars have in mind a set of legal norms and practices-mostly, but not necessarily infraconstitutional-that, without addressing the possibility of party bans directly, minimize the effect of false positives.Second order regulatory mechanisms can include election law design or electoral system choices 133 , forms of mandatory voting 134 , access to public financing and other forms of increasing political participation costs for disloyal parties. 135ur positive model has important implications for the debate about first and second order regulation.Ex ante regulation is primarily first order regulation; ex post regulation can be both first order, such as banning parties, and second order, such as legal norms and practices that undermine the viability of such parties.
The trade-offs discussed previously can easily be applied to ex post second order regulation.It can be alleged that false negatives are less costly with second rather than first order regulation because there is no risk of banning truly democratic parties.However, there is still the possibility that some truly democratic parties incorrectly bear additional costs.In fact, for example, ex post second order regulation can be easily used to protect and entrench incumbent parties rather than minimize the influence of disloyal parties. 136Therefore, while potentially protecting democracy from the detrimental consequences from allowing false positives, second order regulation can impose significant costs in terms of limiting electoral choices or providing significant barriers to entry to new democratic parties.In terms of comparing ex post first and second order regulation, our model does not suggest that the latter is always or universally better than the former.Both have costs and benefits.As discussed before, ex post first order regulation deals with banning false positives while minimizing false negatives when comparing to ex ante regulation.Yet, ex post second order regulation avoids banning but develops legal norms and practices that potentially affect all new parties and ossify the old party system, thus, constraining voters' preferences.The balance between ex ante control and ex post liability, first, and between ex post first and second order regulation, second, requires an understanding of specific local determinants.It is important to emphasize that we do not suggest that every jurisdiction should have the exact same mix of ex ante control-false negatives-and ex post liability-false positives-as we also do not expect that all jurisdictions agree on the appropriate balance between ex post first and second order regulation.In other words, there is no unique size that fits all.For example, social and political preferences matter.A country evolving from a totalitarian regime is likely to exhibit less tolerance for false positives than a liberal democracy with no dictatorial past.

D. Exploring the Data
This section provides for a quantitative exploration.We start by describing the dataset.Our first results are derived from clustering analysis.This exercise allows us to identify the main trends in terms of substantive differences across countries.The following step is to consider determinants that can explain these substantive differences.Finally, an application of principal component analysis highlights the relevant distinction between ex ante and ex post constitutional regulation of political parties.

I. Dataset
The authors have collected information about constitutional regulation of political parties as of 2021 in thirty-seven countries. 137This set of countries includes most member states of the European Union and OECD plus a sample of other countries in Latin America.This is not a random sample but one for which information could be obtained by consultation with experts in absence of comprehensive information.They are all democratic countries because considerations about autocratic countries are excluded from our discussion.
For each country in the dataset, the authors have considered a survey of eleven questions that cover constitutional reference and regulation of political parties, mechanisms of ex ante and ex post enforcement, and existing case law. 138These questions reflect both de jure constitutional rules and de facto information about actual litigation.
As summarized by Table 2, the survey starts with a more contextual constitutional framework: 70% have constitutions that mention political parties in some way.Within civil liberties, only 32% include an explicit reference to political parties.However, 81% have both explicit and implicit references.
Details of concern include the fact that 46% prohibit certain parties-in more detail, fascist, and similar parties (16%), communist parties (8%), regional parties (8%), other-including violent-parties (13%)-whereas 11% protect specific types of parties, mainly regional minority parties, 11%.At the same time, 51% include certain aspects of regulating the internal organization and activity of political parties.One can also observe that 92% allow nonparty members to participate in general elections, hence providing some limitation to party monopoly in the democratic process.Specifically, the possibility of non-party members in party lists, 87%, and non-party lists or candidates, 60%.
In terms of enforcement, the survey makes a distinction between ex ante and ex post institutional mechanisms.Only 8% are subject to some form of control before registration, but 84% have specific registration laws for political parties.At the same time, 73% contemplate some form of regular monitoring of political parties' activities-regular monitoring by an independent agency, 16%, an election committee, 38%, constitutional court, 8%, and other public institutions, 41%-and 76% establish the possibility of sanctioning parties for anti-democratic activitiessanctioning by regular courts.51%, constitutional court, 38%, special courts, 8%, executive branch, 22%, and legislative branch, 11%-while 43% have developed case law in recent times.
As a general first conclusion, most of the countries in the sample have constitutional regulation of political parties, including some degree of monitoring and sanctioning.However, few countries implement ex ante enforcement, most rely on ex post mechanisms.A significant proportion has produced case law on the subject.

II. Clustering Analysis
The goal of clustering analysis is to identify from the available dataset, rather than traditional legal analysis, the similarities and dissimilarities across jurisdictions when it comes to constitutional rules about political parties.This approach is atheoretical and, therefore, purely data driven.In this respect, we follow the methodology developed by Chang et al. for property law. 139As these average-linkage clustering is one of several methods of agglomerative hierarchical clustering where the clusters are then sequentially combined into larger clusters until all elements end up being in the same cluster.Each step of the clustering algorithm combines the two candidate clusters separated by the shortest distance.In average-linkage clustering, the link between two clusters contains all pairs of elements, and the distance between clusters equals the average distance between those pairs of elements in distinct clusters.Average-linkage clustering avoids a drawback of alternative agglomerative methods where clusters formed via single-linkage clustering may be forced together due to single elements being close to each other, even though many of the elements in each cluster may be very distant from each other.The average-linkage clustering algorithm requires a constant-rate assumption that the distances from the root to every branch tip are equal.Id.
authors do, an average-linkage agglomerative hierarchical clustering, that is, an unsupervised machine-learning method, is developed to derive a dendrogram that shows how the thirty-seven jurisdictions in our dataset can be categorized into a full family tree.Methodologically, depending on how closely one observes, a family tree can be divided into 2, 3, 4, or even 37 branches.First, we show the complete picture of the family tree.Second, we discuss the results mainly based on a cutoff of five groups, although including twelve subgroups to assist a more qualitative analysis.Notice that more or fewer groupings are equally valid; five and twelve are always arbitrary choices by the authors.The graphical output of agglomerative hierarchical cluster analysis is called a dendrogram.In our case, it graphically presents all thirty-seven countries grouped at various levels of-Gower-distance, based on the eleven questions surveyed.We used the itemized sub questions rather than the aggregate question, when given the choice, to better reflect possible distinctions.
In Figure 1, all the thirty-seven jurisdictions used in the analysis are placed in the dendrogram.As explained by Chang et al., "the horizontal lines extend upward for each terminal grouping, and at various similarity values, these lines are connected to the lines from other observations with a vertical line.Groups of observations continue to combine until, at the top of the dendrogram, all observations group into a single cluster." 140t this point, we can observe that all countries are ranked into some clustering order, starting with Australia, on the left-hand-side, and Germany, on the right-hand-side.The conclusion is that Australia and Germany are the most distinct countries in our study.The disadvantage of the full dendrogram is that we do not get an immediate intuitive explanation for the family tree.Therefore, we present a more detailed summary of the family tree when using cutoffs of five and twelve distinct groups in Table 3.
Combining the information in Table 3, the dataset yields five distinct branches-branch one with twenty-nine countries, going from Australia to Hungary in Figure 1, and itself subdivided in six different subgroups.Branch two has four members: Brazil, Poland, Portugal, and Turkey.Branch three is Israel.Branch four is Mexico and Colombia.Branch five is Germany.Overall, one can conclude that the full family tree suggests twenty-nine countries that are less regulated and eight countries that are more regulated.However, the nature of the regulation mechanism and its implementation differs, thus generating four distinct branches on the right-hand-side of the dendrogram.Notice that the graphical representation does not say that Germany is more regulated than Israel, and Israel, in turn, is more regulated than Portugal.Individual country comparisons of that sort are misleading with this methodology.Still, the dendrogram does convey the message that countries on the left-hand-side are somehow less regulated and constitute a large family, with some subbranches, while jurisdictions on the right-hand-side are more regulated but in varying ways, thus constituting separate branches.

III. Possible Determinants of Clustering Analysis
We investigate possible determinants of the previous clustering analysis.The hypothesis is that countries in the left-hand-side of the dendrogram have an institutional setting more reluctant to limit political activity whereas jurisdictions in the right-hand-side have the opposite context, for example, due to historical reasons or political inclinations.The independent variables are wellknown measures provided by different international public and private organizations.Specifically, we consider the freedom score by Freedom House, rule of law by the World Bank Governance Indicators, and liberal democracy by V-Dem.Given the thirty-seven countries in our dataset, we also make a distinction between common-law and civil-law systems and jurisdictions subject to the European Convention of Human Rights.
Two important additional controls are included in the statistical analysis.They both reflect the proportionality of the election system for the main legislative institution, a standard second order regulation mechanism as discussed in the previous section.The least-squares index provides a direct measure of proportionality-it measures disproportionality between the vote distribution and the seat distribution. 141The effective-threshold index assesses the minimum voting-in percentage-to achieve representation in the main legislative institution. 142Notice that the interpretation of both variables is similar-a lower number means more proportionality and less percentage of votes to get elected whereas a higher number means less proportionality and more percentage of votes to get elected.
In Table 4, the descriptive statistics are introduced for both the dependent variable (order in the dendrogram, with Australia being one and Germany being thirty-seven) and the independent controls.Notice that 16% are common-law jurisdictions-with Cyprus and Israel categorized as 0.5 rather than 1, civil-law or 0, common-law-and 70% are subject to the European Convention of Human Rights.
Due to the limited number of observations, we report the ordinary-least squares results for individual specifications (I) to (III) and the full specification (IV) in Table 5.A negative coefficient means that the independent variable produces an impact closer to Australia-left-hand-side of the dendrogram-whereas a positive coefficient has the converse interpretation, that is, closer to Germany-right-hand-side of the dendrogram-in the ordered cluster.
As to the statistical quality of the regressions, there is no indication of multicollinearitymeasured by the variance inflation factor, VIF-and of heteroskedasticity-as pointed out by the standard White's test.The goodness of fit indicates that specifications including the independent variable freedom score are better than otherwise.
The statistical results are fully consistent with the hypothesis.In particular, a higher freedom score locates the jurisdiction closer to Australia and further away from Germany, on average.Rule of law and liberal democracy have a similar effect, but only in the absence of freedom scores, in specifications (II) and (III) respectively.This result is indicating that an inclination for more freedom, rather than the quality of democracy or the quality of the legal system, correlates with the relative order in the computed clustering.One can also notice that common-law jurisdictions are more likely to be situated at the left-hand-side of the ordered clustering.The coefficient related to the European Convention of Human Right is negative but never statistically significant.The coefficient associated with the least-squares index is positive but also never statistically significant.Thus, in this case, second order regulation does not seem to be relevant to explain differences across countries.
Unfortunately, concerning the effective-threshold index, we only have data available for 20 countries.Therefore, the inclusion of this control reduces the number of observations.Our findings are reported in Table 6.Goodness of fit and statistical significance are weak by comparison with Table 5.Nevertheless, the results are consistent in terms of the sign and statistical relevance of freedom score.The regression analyses provide statistical support to the intuitive interpretation that family branches on the left-hand-side of the dendrogram combine less regulated jurisdictions-on average, with better freedom scores too-whereas family branches on the right-hand-side of the dendrogram tend to be more friendly to regulation of political parties-on average, with worse freedom scores.

IV. Principal Component Analysis of Enforcement Mechanisms
Based on the collected information for each of the thirty-seven jurisdictions, following the method already discussed by Garoupa and Santos Botelho, 143 we used factor analysis to construct two indicators to measure enforcement rules-ex ante and ex post.Factor analysis provides for a statistical indicator that summarizes information about a set of dummy variables by exploring differences in variance-the information about the statistical determination of all indicators is available from the authors upon request.Both ex ante and ex post reflect in statistical ways the variance across these underlying dummy variables. 144Therefore, the calculated indicators are driven purely by data considerations, rather than subjective conjectures about the nature or additivity of each variable.Specifically, the weight each of the dummies has on a given indicator, ex ante or ex post, is determined by factor analysis and not some ad hoc consideration.These indicators reveal and illuminate the important regulatory aspects that explain differences across countries.
Considering ex ante, we rank countries from low enforcement rules-Austria, Belgium, Germany, Italy, Netherlands, Switzerland-to high enforcement rules-Bulgaria, Israel, Portugal-with a single intermediate category.In relation to ex post, the indicator has some more variance-it operates with three dummies rather than two dummies as the previous indicator-reflected in lowest enforcement rules-Cyprus, New Zealand, Sweden, Switzerlandand highest enforcement rules-Belgium, Bulgaria, Czech Republic, Denmark, Germany, Israel, Portugal, Romania, Slovakia, South Korea, Turkey, USA-and more possible categories in between.
In Table 7, one can observe the distribution of the thirty-seven countries in terms of enforcement mechanisms on a 3x3 matrix.The largest group of jurisdictions in our dataset has some form of intermediate enforcement, about thirty countries, with some variations within the two ex ante and ex post indicators.On the main diagonal, we have two additional important groupscountries with overall weak enforcement rules-Austria, Switzerland-and countries with overall strong enforcement rules-Bulgaria, Israel, Portugal.On the opposite diagonal, we find a small group with strong ex post enforcement and weak ex ante enforcement rules-Belgium and Germany.No country has the opposite enforcement mechanism, that is, consistent with our theory about political parties in democracies, no country has strong ex ante enforcement and weak ex post enforcement rules.In the case of ex ante, the indicator considers questions 6 and 7 in the survey.As to ex post, questions 8, 10, and 11 are included in the indicator.
The results confirm that most jurisdictions have some enforcement rules that combine both ex ante, mostly some form of specific registration, and ex post, mainly some form of judicial control, although many without actual case law.In line with previous considerations, countries in the different extremes of the distribution are quite limited in number.

E. Case Law
In post-war European constitutional history, the "most muscular forms" of militant democracy 145 -party bans-were beforehand an "instrument of symbolic politics." 146As we will see though, as democratic consolidation progressed, some high courts decided not to outlaw parties of marginal political relevance. 147lthough the case law available and the myriad of states that have already banned parties is remarkable, we decided to focus on five constitutional experiences: Germany, Israel, Greece, Spain, and Turkey.All these states are members of the Council of Europe, and therefore it is interesting to compare how domestic jurisdictions and the European Court of Human Rights handled this delicate subject.
Parties' proscription can ultimately be legitimate through its conformity with the rule of law. 148Indeed, as Backes observes, party banning in autocracies might be left solely to the executives, while most democracies give political parties "the privilege of only being able to be dissolved by a high court decision." 149he European Commission for Democracy through Law (the 'Venice Commission') offered some guidelines regarding the prohibition and dissolution of political parties. 150It stresses that the prohibition or dissolution of political parties is "a particularly far-reaching measure [that] should be used with utmost restraint" and urges states to opt for "less radical measures" that can also prevent democratic distress. 151In 2020, the Parliamentary Assembly of the Council of Europe called on the governments of member states to "ensure that measures restricting parties cannot be used in an arbitrary manner by the political authorities." 152In short, both institutions recommended parsimony.

I. Germany
The German Basic Law includes several provisions to safeguard democracy. 153In a fairly "complex" 154 writing, Article 21 Section 1 of the German Basic Law states that: "Political parties shall participate in the formation of the political will of the people [ . . .]." Scholarship and jurisprudence refer to this section as the "party privilege" (Parteienprivileg), which grants all political parties the freedom to organize and mobilize the electorate. 155Article 21 Section 2, however, consecrates that "parties that, by reason of their aims or the behavior of their adherents, seek to undermine or abolish the free democratic basic order or to endanger the existence of the Federal Republic of Germany shall be unconstitutional." 156n 1953, the German Federal Constitutional Court (FCC) declared unconstitutional the Socialist Reich Party (SRP)-a successor of the German Imperial Party with a neo-Nazi orientation-and dissolved it. 157Echoing the abovementioned Schmittian idea of a constitutional core, the FCC considered that exclusion of a political party is justified when "supreme fundamental values of liberal democratic constitutional state" are shaken. 158The Court considered that the SRP revealed an undemocratic and racist political agenda.Still, at that time, the FCC did not expressly mention the concept of militant democracy. 159ontrary to models of negative republicanism, 160 such as Austria or Portugal, in which, respectively, only neo-Nazi 161 or fascist 162 parties are prohibited, in Germany the pendulum swung back and forth to the extreme left and to the extreme right.Such oscillation motion reveals that what was relevant to the FCC was not the righteousness of the political ideas, but their unconstitutional objectives.In accordance, what Article 21 of the Basic Law seeks is to block "any repetition of the one-party state that molded the Third Reich." 163he first formal reference to militant democracy, not only in Germany but also in Europe itself, was made on the case of the Communist Party of Germany (KPD).KPD, amid the Cold War, called for the downfall of the Adenauer's government. 164What was impressive about this case was that, although mentioning militant democracy, it departed from Löwenstein's idea of urgency or "imminence." In 1956, in "the longest by far of all the Court's opinions" 308 pages, 165 the FCC ruled that the ultimate goals of the party-proletarian dictatorship and revolution-were incompatible with the "free democratic basic order." 166There is little doubt that, in this case, the application of Article 21 was more problematic than in the previous case. 167KPD obtained only 2.2% of the popular vote in the 1953 Bundestag election so was not a menace at the time of its banning. 168Still, despite the inexistence of an "empirical danger," the Court focused on the "logical danger" of the KPD fulfilling its unconstitutional goals in the foreseeable future. 169n the Radical Groups Case, 170 in 1978, the FCC followed a "militant dormancy" reasoning. 171At that time, three radical left-wings parties were denied campaign broadcasting in some German states.Overall, the Court held that such denial threatened the principle of equality of opportunity granted to every political party.
In 2003, the FCC considered a petition to ban the extreme right-wing National Democratic Party (NPD), although the proceedings were later dismissed on technical grounds. 172Quite astonishingly, the Court found out that much of the evidence considered in the admissibility stage had derived from very well-placed state agents and informants working within the NPD. 173n 2017, the FCC ruling on NPD disappointed "both opponents and adherents" of militant democracy. 174It failed opponents, as they anticipated a much more significant increasing of the intervention threshold, and it dissatisfied adherents as the decision was not to ban the NPD. 175hile recognizing NPD affinity with National Socialism, the examination of proportionality, using the criteria of necessity and appropriateness, brought the Court to the conclusion that the NPD's actions lacked "potentiality." 176As NPD had a shy parliamentary representation and a declining membership, its actions would not succeed in mobilizing the "right-wing extremist movement." 177This reasoning represented a paramount shift.Uninfluential parties would not be subjected to the extreme political measure of proscription.Still, they could be subjected to other less severe measures-a "soft militant democracy" 178 path-such as the withdrawal of public party financing.
In conclusion, the FCC went from a sheer militant democracy stance to a more mitigated or rhetorical resort to militant democracy.On a surface level, it may seem that the fact that the FCC has now a more tolerant attitude towards antidemocratic or illiberal parties reveals "the evolving maturity of German's democracy." 179However, upon a closer inspection, such conclusion must be interpreted cum grano salis.As unlikely as this may sound, in an empirical study, Bourne and Casal Bértoa unveiled that "against all expectations" states that experienced authoritarianism are equally prone to outlaw parties as those who have not had such historical experiences. 180

II. Israel
Navot delves into the interesting paradox of the Israeli experience regarding the banning of parties. 181On the one hand, Basic Law Section 7A and the Parties Law militantly allow "an 'easy' disqualification of terror-supporting parties," as bans can be approved merely for illegitimate political speech. 182On the other hand though, the Supreme Court of Israel adopts a neutral stance, does not adhere to the legal militant democracy shibboleth, and states that the proscribing of political parties must be reserved when there is unequivocal proof of terroristic actions. 183This invites a second important point: Is Israel the example of a "softer militant democracy" 184 path?Right before the 1965 elections, the Israeli Central Elections Committee 185 refused to approve the Socialist List, arguing that it was an illegal association, which denied the "integrity and the very existence of the State of Israel." 186Adhering to the idea of a defensive democracy, the Supreme Court, in a majority opinion, confirmed the ban. 187wo decades later, the problem reemerged, as the extremist right-wing "Kach"-the Kahana movement-sustained a racist agenda against Israel's Arab population. 188Following Kach's banning by the Central Elections Committee, the Supreme Court, however, approved the party's candidacy, arguing that a democratic society should respect unpopular worldviews. 189This decision was polemic and triggered legislative changes in electoral law.The Knesset then adopted the aforesaid Parties Law, allowing the outlawing of political parties. 190n 2020, the Central Elections Committee disqualified the Balad-the Democratic National Assembly Party-subsequently to the declarations made by Balad's founder, Bishara, in Syria, supporting the terrorist organization Hezbollah. 191The Supreme Court overturned the decision and decided, by majority, that Bishara could participate in the Knesset elections. 192It follows that the dipole "speech" versus "actions" was determinant to the Court's conclusion.This confirms the assertion that democracies will be more hesitant to proscribe parties for their anti-democratic ideas than for any anti-democratic actions. 193he Court, while acknowledging Bishara's support for a terrorist organization, considered that what was relevant were the actions-which, in this case, did not exist-that transform "an idea into reality" and not the rhetoric goals. 194Therefore, as Navot writes, "it is preferable for undemocratic pressures to find their expression within the legitimate framework of the democracy, and not externally." 195Clearly, then, the Court consciously opted to distance itself from this highly political dilemma.To conclude, Israeli democracy "is perhaps a 'defensive democracy,' but primarily 'on paper.'" 196

III. Greece
The example of Greece is relevant, as, for historical reasons, the Greek Constitution does not allow the dissolution of political parties. 197By contrast, party banning recalled the memory of anticommunism and dictatorial oppression in Greece. 198Echoing these concerns, the Greek Constitution foresees that political parties "must serve the free functioning of the democratic form of government." 199Malkopoulou contends that while the Greek model is procedural, the Golden Dawn case provoked "a renewed interest" in consecrating militant democratic provisions. 200roscribing of political parties was not on the political agenda until 2012, when the far right and neo-Nazi party Golden Dawn won twenty-one seats in Parliament. 201Its ascension was thought to be justified by the major financial and economic crisis that tore apart the Greek's social fabric, which triggered the rise of populism.However, as Kouroutakis explains, the latest elections in central and north Europe rebutted "the theory that correlates the rise of nationalism with the austerity measures and the far right parties as a 'refuge of the poor.'" 202n most aspects, instead of a constitutional issue, the problem of extremism within political parties was addressed by a "nexus of criminal and administrative law." 203Political measures were adopted to tackle racist violence, such as creating special units within the police. 204In 2013, criminal proceedings took place to arrest Golden Dawn leadership 205 for criminal offences, and the Criminal Code was also amended to include certain forms and expressions of racism and xenophobia. 206Rules for public funding of political parties were amended to incorporate a temporary suspension of financial support in the event of criminal prosecutions and imprisonment of a party leader or one-fifth of a party's members for criminal offenses. 207Golden Dawn appealed against this legislation.However, the Council of State dismissed all claims. 208ollowing the criminal procedures started on 2013, in 2020, the Athens Criminal Court unanimously declared Golden Dawn a criminal organization operating "under the cover" of a political party and ruled jail sentences for all the defendants. 209

IV. Spain
After nearly forty years of dictatorship from 1939-1975, during which political parties were illegal, Spain distanced itself from the German model of militant democracy. 210Provided that they did not use violence and obey the rule of law, all parties were welcomed in the new established democracy.This constitutional design was distinct from the neighbor state Portugal that, after similarly having endured one of Europe's longest dictatorships from 1933−1974, opted for a clearcut militant constitutional design. 211f the Spanish Constitutional Court (SCC) was more cautious after the transition to democracy precisely to encourage political participation through parties, after the democratic consolidation such restraint disappeared. 212As the SCC ruled, the political character of parties does not convert them into "state's organs." 213Instead, they are social organizations with "constitutional relevance." 214rticle 6 of the Spanish Constitution very clearly states that: Political parties are the expression of political pluralism; they contribute to the formation and expression of the will of the people and are a fundamental instrument for political participation.Their creation and the exercise of their activities are free in so far as they respect the Constitution and the law.Their internal structure and operation must be democratic.Furthermore, the SCC has determined that, notwithstanding Article 22 of the Constitution mentioning only to the dissolution of "associations," political parties are also to be considered associations for the purposes of that provision. 215evertheless, and distinctively from the German experience, it is relevant to mention that the possibility of the dissolution of parties that do not respect internal democracy is not consecrated on the Constitution itself. 216In 2002, Spain's Congress approved by a vast majority the Law of Political Parties, which introduced new procedures to dissolve political parties threatening to undermine the democratic system, including promoting discrimination, legitimizing violence, and supporting a terrorist organization. 217he law followed an agreement-the Agreement for Liberties and Against Terrorismbetween the governing party, the right-wing Popular Party (PP), and the main opposition party, the Socialist (PSOE). 218The focus of the proscription was not the ideology, but the concrete actions. 219The Law of Political Parties did not intend to metamorphose Spain into a militant democracy.As Ferreres Comella ironically wrote, it was a statute aimed to target parties of the "Batasuna-type" and could very well have been baptized as "Statute to Outlaw Batasuna." 220n August 2002, Judge Baltasar Garzón of the court of first instance suspended the activities of Batasuna amid a criminal investigation against eleven individuals suspected of terrorism. 221In 2003, the Supreme Court unanimously dissolved the Basque party, Batasuna, 222 on the ground that it was supporting terrorism. 223This decision was confirmed by the SCC on January 16, 2004.Simply put, Batasuna stands for the independence of the Basque Country and has close links with the Euzkadi ta Askatasuma (ETA)-the Basque Fatherland and Liberty terrorist group. 224his case, as well as similar cases related to the disqualification of electoral lists, was referred to the European Court of Human Rights (ECtHR).As predicted by several scholars, the ECtHR found that Spain's ban on Batasuna served a legitimate aim. 225So far, in all the cases appealed to the ECtHR regarding the disqualification of lists and parties, 226 the ECtHR upheld the rulings of the Spanish courts. 227On the Batasuna case, and anchoring on Articles 10(2) and 11(2) of the European Convention on Human Rights, the ECtHR considered that the close link between Batasuna and the terrorist organization ETA could be considered as an "objective threat to democracy." 228 Turkey Inspired by the German example, the 1982 Turkish Constitution recognized a "militant" type of democracy.229 The Turkish Constitution states that the Republic of Turkey is a "democratic, secular, and social State based on the rule of law, respectful of human rights in a spirit of social peace."230 In the first decade, in most respects, the judiciary banned several parties for formal reasons, such as names or symbols.After that, laws on political parties were focused on ideology-the United Communist Party of Turkey-Kurdish nationalism, or faith-Islamic parties.
In 1998, the Turkish Constitutional Court (TCC) dissolved the largest party in the country, the Refah Party. 231The banning was very interesting as, contrary to all the cases that we have previously revisited, the outlawed party was not irrelevant or a minority party, but it was the main political party, whose leader, Necbettin Erbakan, was the Prime Minister of Turkey.One might query: Should the Turkish example be considered as super-militancy? 232ll things considered, Refah was the ruling party by then in a coalition government.It sustained a so-called plural system, whereby citizens of different faiths were allowed to select the legislation, for example, Sharia, regulating certain aspects of their life, such as family and inheritance law. 233The TCC ruled that this violated the constitutional principles of secularism and the prohibition of non-discrimination. 234s far as the ECtHR is concerned, its jurisprudence recognizes the right to associate in political parties as falling within the scope of the freedom of association. 235On the Refah case, the ECtHR upheld the ban, as the party's support of violence and religious discrimination contradicts the freedoms guaranteed by the European Convention on Human Rights. 236To conclude, "implicitly" 237 to some, or "explicitly" 238 to others, the ECtHR uplifted militant democracy into a constitutional value at the European level.239

F. Conclusions
This Article has endeavored to unveil a symbiotic relation between democracy and parties.Drawing out exactly what this means is trickier than it might seem.Democracy is needed in order to maintain parties, and parties are needed in order to maintain democracy.Therefore, and as Flores Giménez wrote, "if political parties exist within evolving societies, it means that the parties themselves can evolve." 240et not all is well in the land of political parties' regulation.At this juncture, liberal democracies address regulation of political parties in distinct ways.Our positive theory suggests that these differences reflect concerns about false positives and false negatives.The empirical findings document that variance in regulation of political parties is associated with broader concerns about freedom.Thus, diverse approaches to regulating political parties are embedded in social preferences and priorities that vary across jurisdictions.
Normative disagreement about the proper regulating of anti-democratic parties and distinct concerns about false positives and false negatives converge in explaining why different countries have taken different routes.From a comparative constitutional studies perspective, one size does not fit all.For example, the empirical results suggest that perceptions about freedom are more relevant than the specifics of election law in explaining variance.
In sum, "illimited relativism," 241 or as we prefer to call it-the absolutism of relativism-can challenge the enforcement of human rights and the quality of democracy.Social inequalities, corruption, and populism are so very much a threat, or a reality, even in the so-called consolidated democracies.Hence, the "democratic project" 242 is perhaps never entirely fulfilled.
Remarkably, almost four decades ago, Otto Pardo warned about what he called "the illusion of the legal scholars," 243 when believing that the mere consecration of norms in a constitution would suffice to magically prevent some political forces from striking down democracy through the democratic process itself.
To be perfectly clear: The world has changed.Not all revolutions are bloody revolutions in the streets.Several illiberal states maintain a façade of a minimum of democratic traits.In so doing, extremism works by manipulating the existing legal arenas.Simply put, as extremists "operate in a world of legalism," 244 procedural democracy will not be able to address them.
Again, militant democracy and party banning might not be the most intellectually sophisticated and normatively dense strategies to restrain illiberal democracy, but at least it offers a solution, even if it is a fragile one.Thus, the idea would be not a maximum of freedom, but an optimum of freedom-that responds to difficult trade-offs such as false positives and false negatives. 245As a result, merely striking down these responses without offering alternatives conspicuously fails to consider the bigger picture and surrenders to "political quietism." 246For one thing, strict procedural democracy "is an unsatisfying response." 247To conclude, and as Wilkinson astutely outlined, at least regulation on the proscription of political parties reassures the public "that something is being done." 248pplementary material.To view supplementary material for this article, please visit https://doi.org/10.1017/glj.2023.117

80
See Richard S. Katz & Peter Mair, Changing Models of Party Organization and Party Democracy: The Emergence of the Cartel Party, 1 PARTY POL. 5, (1995).

Table 2 .
Survey of 37 countries

Table 3 .
Clustering by average linkage SEATS AND VOTES: THE EFFECTS AND DETERMINANTS OF ELECTORAL SYSTEMS 1 (1989) (explaining that the index is based on the authors' work and the statistics calculated by Rein Taagepera and expanded by The Politics of Electoral Systems).See MICHAEL GALLAGHER & PAUL MITCHELL, THE POLITICS OF ELECTORAL SYSTEMS 3 (2008).

Table 5 .
Regression analysis of dendrogram order by ordinary-least squares

Table 6 .
Regression analysis of dendrogram order by ordinary-least squares