O Routledge Taylor &. Francis Group West European Politics ISSN: 0140-2382 (Print) 1743-9655 (Online) Journal homepage: http://www.tandfonline.com/loi/fwep20 Political parties or party systems? Assessing the 'myth' of institutionalisation and democracy Fernando Casal Bertoa To cite this article: Fernando Casal Bertoa (2017) Political parties or party systems? Assessing the 'myth' of institutionalisation and democracy, West European Politics, 40:2, 402-429, DOI: 10.1080/01402382.2016.1216921 To link to this article: http://dx.doi.Org/10.1080/01402382.2016.1216921 ^ Published online: 18 Aug 2016. Submit your article to this journal G? ■Ill Article views: 821 pjl View related articles G? View Crossmark data G? Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journal lnformation?journalCode=fwep20 Download by: [Masarykova Univerzita v Brne] Date: 06 October 201 7, At: 06:10 WEST EUROPEAN POLITICS, 2017 VOL. 40, NO. 2,402-429 http://dx.doi.org/10.1080/01402382.2016.1216921 § ^ Taylors Francis Group iRoutledge Taylor & Francis GroL Political parties or party systems? Assessing the Vnyth'of institutionalisation and democracy Fernando Casal Bertoa School of Politics and International Relations, University of Nottingham, UK ABSTRACT No matter the region of the world under study, party (system) institutionalisation has been traditionally considered to be a necessary, but not sufficient, condition for the survival of democracy. Despite being one of the most quoted statements in the democratisation literature, the few studies looking at the relationship between institutionalisation and democratic endurance have found no evidence of the 'almost magical'powers of the former. This article revisits the abovementioned research question by making use of an original dataset covering all European democracies between 1848 and 2014. The main findings are threefold: (1) it is not the institutionalisation of political parties but the institutionalisation of party systems as a whole that has fostered the prospects for democratic survival in Europe; (2) there is a threshold of systemic institutionalisation which, once reached, will avoid democratic collapse; and (3) systemic over-institutionalisation does not seem to be so perilous for the survival of democracy. KEYWORDS Party system institutionalisation; party institutionalisation; democratic survival; Europe; 1848-2015 Reflecting Huntington's preoccupation with weak institutions in the second half of the 1960s, party (system) institutionalisation [P(S)I] has remained a recurrent theme in the (usually pessimistic) democratic consolidation literature (Pridham 1990). Among scholars there seems to be a widespread agreement that whether in Africa (Lindberg 2007; Weghorst and Bernhard 2014), Asia (Johnson 2002; Hicken 2006), Europe (Lewis 1994; Morlino 1998) or Latin America (Dix 1992; Mainwaring and Scully 1995), few institutional developments are more critical to the endurance and healthy functioning of democracy than the institutionalisation of both political parties and party systems (Abeje 2013; Diamond and Linz 1989; Kuenzi and Lambright 2005; Tavits 2005). As a result, P(S)I has traditionally been considered to be a necessary, but not sufficient, condition for the survival of democracy (Arter and Kestila-Kekkonen 2014; Mainwaring 1999). CONTACT Fernando Casal Bertoa @ Fernando.Casal.Bertoa@nottingham.ac.uk © 2016 Informa UK Limited, trading as Taylor & Francis Group WEST EUROPEAN POLITICS @ 403 However, despite having become one of the most repeated statements in the democratisation literature, such a close relationship between institution-alisation and democracy has not yet been sufficiently proved beyond a myriad of countries (Johnson 2002; Mainwaring 1999; Markowski 2001) and/or regional (Kneuer 2011; Lewis 2006) studies. In fact, when tested in a larger number of cases, 'the correlation [has proved to be] not as impressive as theory predicts' (Basedau 2007: 125), not to say negative (Stockton 2001) or even non-existent (Thames and Robbins 2007), putting into question Huntington's and Mainwaring's seminal concerns about the benefits of both party institu-tionalisation (PI) and party system institutionalisation (PSI) for the durability of democracy (Hicken and Kuchonta 2014).1 Notwithstanding what has been said, and taking into consideration that the relationship between institutionalisation and democracy may not be as unidirectional or linear as expected (Schedler 1995; Wallis 2003), the present article revisits the abovementioned relationship but differs from previous studies in the following manner. First of all, it distinguishes between PSI and PI. Secondly, it tries to improve the way in which both phenomena have been operational-ised, by measuring the whole process of institutionalisation, rather than at one point in time.2 Thirdly, it employs an original dataset comprising all European democratic political regimes since 1848, allowing for both geographical and chronological comparisons. Finally, it makes use of a different method enabling identification of to what extent PSI and/or PI are to be considered necessary and/or sufficient conditions for the endurance of democracy. In order to avoid confusion, it is important to note here that the article does not aim to either explain variation or measure the impact PSI and/or PI have on democratic survival and/or collapse or how the former increase or decrease the probability of the latter.3 Neither the research design nor the method employed here are, respectively, designed to do so. In turn, the main ambition of the article is to determine if PSI and/or PI (or their lack) have (historically) been necessary and/or sufficient for the survival and/or collapse of democracy in Europe.4 Trying to fulfil all the above-cited goals, the article is structured in four different sections. The first one reconsiders the conceptualisation debate regarding PI/PSI. Section two identifies satisfactory indicators for the two notions examined, trying to adequately capture their different dimensions. The third section presents the dataset as well as explaining the methodology employed. After briefly reviewing the literature discussing the different mechanisms linking institutionalisation and democracy, the final section examines the veracity of the necessary, but not sufficient, relationship between P(S)I and democratic survival. In the conclusion, the implications of the main findings of the article are addressed. Party and party system institutionalisation Any scholar studying the institutionalisation of party systems faces the problem of the unit of analysis: political parties, party systems, or both? Surprisingly 404 (g) F. CASAL BERTOA enough, notwithstanding an ever growing number of systematic comparative works and countless case studies, most scholars (e.g. Kreuzer and Pettai 2003; Lewis 2006; Mainwaring and Scully 1995; Toka 1997) still approach the insti-tutionalisation of individual parties and party systems as two interchangeable and synonymous concepts, 'the implication being that [since individual political parties constitute integral parts of the whole party system] the institutionalization of the party system directly depends on that of individual parties' (Meleshevich 2007: 16). The relationship between these two notions is, however, not nearly so 'simple and deterministic' (Markowski 2001: 56): while individual political parties may be institutionalised, their operation in a party system may not be. In this sense, Randall and Svasand argue that, although closely related, 'individual PI and the institutionalization of the party system are neither the same thing nor necessarily and always mutually compatible' (2002:6). Moreover, they 'could be at odds' (2002: 8), particularly in the case of young democracies. As a result of this lack of conceptual clarity or absence of consistent analytical frameworks, research on PSI and PI has thus far led to inconclusive, in many cases even contradictory, assessments on the relationship between institutionalisation and democracy. It is for this reason that I will turn to the distinction between these two phenomena first, trying to put some flesh on the bones of both concepts. Party system institutionalisation Although it may be difficult to believe given its central importance, the concept of PSI has no established definition. The concept was first introduced by Mainwaring and Scully in their classic Building Democratic Institutions: Party Systems in Latin America. There, the authors denned the institutionalisation of a party system as: [the] process by which a practice or organization becomes well established and widely known, if not universally accepted. Actors develop expectations, orientations, and behaviour based on the premise that this practice or organization will prevail into the foreseeable future. (Mainwaring and Scully 1995: 4) According to these two authors, institutionalised party systems are characterised by four different dimensions: regular patterns of inter-party competition, strong party roots in society, electoral and partisan legitimacy, and solid party organisations. While their discussion of the four dimensions is certainly insightful, Mainwaring and Scully failed to provide objective measures for the last two dimensions (i.e. legitimacy and party organisation).5 Most authors follow Mainwaring and Scully's pattern of proposing a series of'dimensions' of PSI. Morlino (1998) claims that 'structured' party systems must be stable in terms of electoral behaviour, partisan competition and political class; Bielasiak (2001), who is interested in the institutionalisation of party systems in Eastern Europe and post-Soviet states, distinguishes three dimensions of stability: electoral democracy, political contestation and political WEST EUROPEAN POLITICS @ 405 representation; Kreuzer and Pettai assert, from a different perspective, that PSI 'is ultimately shaped by the interaction of both politicians' organizational affiliations and voters' electoral choices' (2003: 81); and, even more explicitly, Meleshevich (2007) conceives of it as involving both (external) autonomy and (internal) stability. More recently, Lindberg (2007) simply puts institutionali-sation on a level with stabilisation. Randall and Svasand (2002) offer the only exception to this principle of simply enumerating dimensions but, notwithstanding its originality, their framework does not provide us with any means of operationalisation. It follows from this brief review, then, that political scientists have conceptualised PSI in numerous ways. Most agree on some dimensions of the notion but not many arrive at the same final combination (Table 1). Criticism of the conceptual and operational approaches of these studies appears elsewhere (Casal Bertoa 2016), but the fundamental problem running through all these works is that, more preoccupied with an empirical assessment of institutionalisation, they tend to pay very little attention to conceptualisation per se. However, as we know from the literature, for an empirical analysis to be valid it is essential first to establish a sound conceptual base (Delia Porta and Keating 2008). Only then can scholars take care of matching such conceptual framework with the most appropriate measures (Adcock and Collier 2001). How, then, can we define PSI? Strictly speaking, we can only speak of institutionalisation when we are able to define what it is that has been institutionalised. Our first task then is, perhaps, insurmountable: to specify the 'essence' of what constitutes a given party system. Sartori (1976: 44) offers the clearest definition of a party system as 'the system of interactions resulting from inter-party competition. This definition has three main different implications. First, a party system must consist of more than a single party (otherwise there is no inter-party competition). Second, a party system clearly involves something more than the sum of its component parts (i.e. political parties). This way it incorporates some element of understanding of the mode of interaction between the latter. Third, the notion of 'system' implies some degree of regularity, suggesting some continuity of inter-party interactions between elections (Sartori 1976: 43). Once the nature of what constitutes a party system has been established, it becomes possible to define PSI and, hence, to specify the dimensions determining whether any given system is already institutionalised or remains under-institutionalised. As is clear from what has been said, all meanings of the notion of institutionalisation contain the idea of stability and persistence (Riker and Ordeshook 1973; see also Table 1). In fact, following Mair's (2006) idea that the core of a party system is to be found in the patterns of interaction among its units, that is, political parties, it follows that the most important and necessary attribute of PSI is stability in the rules and nature of inter-party competition (Lindberg 2007). Indeed, as Mainwaring and Scully (1995: 4-5) stated in their seminal analysis of Latin American party systems, 'where such stability does not exist, Downloaded by [Masarykova Univerzita v Brne] at 06:10 06 October 2017 Table 1. Dimensions of party system institutionalisation proposed by selected scholars. Dimensions of party system institutionalisation Author(s)/subject of study Complexity Autonomy/ Affiliation Scope/stable and coherence Adaptability boundary Stability patterns roots Legitimacy Party organisation Mutual acceptance Mainwaring and Scully (1995) Latin American party systems Morlino(1998) Southern European party systems Bielasiak (2001) Post-communist party regimes Randall and Svasand (2002) Third World party systems Kreuzerand Pettai (2003) Baltic States party systems Meleshevich (2007) Post-Soviet party systems Lindberg (2007) African party systems X X X X WEST EUROPEAN POLITICS @ 407 institutionalization is limited'. Therefore, the more stable the system, the more institutionalised it becomes (Mair 2001: 35). Bearing in mind all that has been said, and drawing on Huntington's (1968: 12) original definition of institutionalisation as the 'process by which organizations and procedures acquire value and stability', I define PSI as the process by which the patterns of interaction among political parties become routine, predictable and stable over time.6 In other words, a system of parties can be said to be institutionalised when political parties cooperate, collaborate and colligate in a standardised and structured way - a way that is independent of the relevant issues in each moment and which random shocks cannot alter (Mainwaring 1998). On the contrary, in under-institutionalised party systems political parties are incapable of interacting in any patterned manner, failing to present voters with clearly stable political alliances and, therefore, predictable governmental alternatives (Mair 2001: 39). Party institutionalisation Although widely employed in the literature, the concept of PI has more often than not been poorly and/or ambiguously denned; while some scholars have used the term without further clarification, others - as we have previously seen - have tended to simply equate it with that of PSI. The result has been a lingering uncertainty about its 'real' meaning. Although the notion of institutionalisation had been previously employed in relation to political organisations, Huntington (1968) was the first scholar to apply it to the analysis of political parties. However, he did not dedicate much time to its definition (just one sentence), and preferred to focus on its (four) dimensions: namely, adaptability, autonomy, complexity, coherence. Most scholars have preferred to follow Huntington's seminal approach of merely suggesting a series of dimensions of institutionalisation, hastening to operationalise them. For some, the notion of PI is uni-dimensional. But while for Janda (1980: 19) a party is institutionalised exclusively when it is 'reified in the public mind', for Rose and Mackie (1988) electoral continuity (i.e. more than three national elections) is the only dimension. For the majority, PI needs to be treated as a multi-dimensional phenomenon. However, while for Panebianco (1988) PI has only two dimensions (i.e. autonomy and systemness), for Harmel and colleagues (Harmel and Svasand 1993; Harmel et al. forthcoming), or McGuire (1997) the concept has a tri-dimensional character: namely, routi-nisation, survival/rootedness and reification. Jin (1995) also agrees with the first two, adding a different 'party efficacy in the legislative process' as a third dimension. Other scholars, like Dix (1992), simply adopt Huntington's conceptual framework but incorporate a completely new set of operational indicators. As with the concept of PSI examined above, the fundamental problem running through most of these works is that they tend to pay little attention 408 (g) F. CASAL BERTOA to conceptualisation per se as they are almost exclusively oriented toward an empirical assessment of institutionalisation. However, it should not be forgotten that any valid empirical analysis needs first a sound conceptual base. In this context, two studies are remarkable, namely Levitsky's (1998) analysis of the transformation of the Justicialist Party in Argentina, and Randall and Svasand s (2002) analysis of the institutionalisation of political parties in the 'Third World'. Well aware of the disjuncture between the initial conception of institutionalisation and the way it has been elaborated and related to specific criteria by the majority of scholars, Levitsky (1998: 88) suggests that 'the concept of [party] institutionalization be unpacked'. Thus, bearing in mind the predominant conceptualisations of institutionalisation used in the literature of political parties, Levitsky (1998: 79) distinguishes two different elements: (1) value infusion, encompassing rootedness, and denoting a 'shift from the pursuit of particular objectives through an organization to the goal of perpetuating the organization per se; and (2) behavioural routinisation, which covers stable patterns of organisation, pointing to entrenched forms of intra-organisational interaction. Clearly influenced by Levitsky's work, Randall and Svasand (2002: 12) distinguished four different dimensions of PI, understood as 'the process by which [a] party becomes established in terms both of integrated patterns of behaviour and of attitudes, or culture'. On the one hand, within the internal sphere, both authors distinguish between systemness (i.e. the increasing scope, density and regularity of the interactions that constitute the party as a structure) and value infusion which refers to the extent to which party actors and supporters acquire an identification with and commitment to a party. On the other hand, the external dimension includes autonomy (i.e. the degree of differentiation from other social groups and methods of behaviour) and reification which, capturing Janda's (1980) notion, requires the party's existence to be established in the public imagination. However, as in the case of PSI, they failed to put their own concept to the test. More recently, in perhaps the more complete study to date of the level of institutionalisation of a political party (i.e. True Finns), Arter and Kestila-Kekkonen added a fifth dimension: namely cohesion. Thus, together with a 'body of "core supporters" (social rootedness),... a core membership, an effective candidate supply, a de facto dispersal of roles and authority (autonomy and systemness), and ... the ability of survive challenges (adaptability)', an institutionalised political party needs to display a certain level of legislative (voting and policy) unity (Arter and Kestila-Kekkonen 2014: 937). From this discussion, it becomes clear that the concept of PI is 'multifaceted, difficult to operationalize, and sometimes conductive to tautological argument' (Gunther and Hopkin 2002:193). However, despite the fact that no two scholars have arrived at the same set of dimensions (Table 2), two broad areas of consensus seem to emerge: PI involves a combination of both stable roots in society as well as firmly established/properly routinised organisational structures. And if Downloaded by [Masarykova Univerzita v Brne] at 06:10 06 October 2017 Table 2. Dimensions of party institutionalisation according to selected scholars. Dimensions of party institutionalisation Author(s)/subject of study Adaptability Complexity Autonomy Coherence/ Systemness/Routini- Reifica- Rootedness/ cohesion sation/organisation tion valueinfusion Continuity/ survival Party efficacy Huntington (1968) Political institutions Janda(1980) Political parties of the world Rose and Mackie (1988) Western political parties Panebianco (1988) Western European political parties Harmel and Svasand (1993) North European political parties Jin (1995) South Korean political parties McGuire(1997) Argentinean political party Levitsky (1998) Argentinean political party Randall and Svasand (2002) Third World political parties Lewis (2006) Post-communist political parties Arterand Kestila- Kekkonen (2014) Finnish political party X X X X 410 (g) F. CASAL BERTOA we accept, as the majority of scholars does, that institutionalisation is characterised by rootedness and systemness (Webb and White 2007:11), then it becomes possible to establish a common definition of PI which is then understood as the process by which parties reproduce consistent patterns of mass mobilisation and internal organisation. In other words, institutionalised political parties are expected to remain stable both at the level of popular support (especially at the moment of elections) as well as in terms of their organisational structure. How to measure institutionalisation Trying to capture what has been identified as the different dimensions of PSI (i.e. stability) and PI (i.e. rootedness and systemness), I will introduce here two indicators which, for the reasons explained below, are more suitable than other 'more traditional' ones for measuring these two notions. Still, for the sake of robustness, the analysis in section four will include all indicators (i.e. both 'traditional' and not). Party system institutionalisation PSI has been traditionally measured using two different indicators: the number of parties and/or electoral volatility (e.g. Bielasiak 2001; Booth and Robbins 2010; Morlino 1998). However, neither of these two classical indicators really captures the notion of PSI. On the one hand, because they are measured only at the time of elections, both indicators do not capture what is also a process (i.e. overtime), not just a property (i.e. at one point in time). On the other hand, although the number of parties gives us important information about the so-called 'streams of interaction (Sartori 1976), it does not address how parties cooperate/compete. Indeed, a country with eight parties, but with a two-bloc competition, may certainly be more institutionalised than another with just four parties. Moreover, even having in a country the same number in two consecutive elections would not mean systemic stability, as the identity of the parties may be totally different, giving rise to different (less predictable) interactions. On the other hand, notwithstanding the problem of 'ecological fallacy' (Luna and Altman 2011: 4), Pedersen's (1979) index of electoral volatility was specifically designed to capture the (in) stability of voters' preferences rather than to address the process of how parties cooperate and/or compete (Mair 1997). Finally, because electoral volatility has always both a supply- and a demand-side (Rose and Munro 2009), Pedersen's index is inadequate to clearly distinguish between party (supply-side) and systemic (demand-side) institutionalisation (Birch 2003; Luna 2014: 412; Powell and Tucker 2014). For all these reasons, I prefer to rely on Mair's (1997) notion of'party system closure' and, more particularly, on Casal Bertoa and Enyedi's (2016) operation-alisation of it. There are five main reasons for this, namely: WEST EUROPEAN POLITICS @ 411 (1) Considering the structure of inter-party competition for government 'the most important aspect of party systems' (Mair 1997: 206), makes it possible to focus on the fundamental 'core' of any party system: namely the process of partisan interactions (Rokkan 1970; Smith 1989). (2) It operationalises institutionalisation at the systemic level independently of the static parameters of its units, making a clear distinction between PSI and PI. (3) It is eminently suited for 'large-scale geographic and inter-temporal comparisons' (Miiller and Fallend 2004: 804). (4) Perhaps more importantly, it allows for an evaluation of the process of institutionalisation on a yearly basis, not only at the time of elections (e.g. fragmentation or volatility measures).7 Building on Mair's (1997) considerations that institutionalised party systems are characterised by (1) wholesale (i.e. total or none) alternations of governments, (2) familiar governing formulae and (3) closed (i.e. to a limited number of parties) governmental access, Casal Bertoa and Enyedi (2016) have created a composite index of PSI which, combining all these three factors, also takes time into consideration. First of all, the degree to which governing alternations of political parties are wholesale is captured by the so-called Pedersen Index of Ministerial Volatility (MV), which adds the net change in percentage of ministers (including the prime minister) gained and lost by each party from one government to the next, and then dividing by two. However, because wholesale alternation (both total and none) can be reflected by scores at both extremes of the MV scale (both 100 and 0, respectively), if the initial MV score obtained according to the formula described above is lower than 50 (i.e. perfect partial alternation), the former figure will be subtracted from 100. If MV is higher than 50, the Index of Government Alternation (IGA) will be equal to the initial MV score. The second and third criteria are calculated by the percentage of ministers belonging, respectively, to familiar combination of parties and old governing parties, with the caveats presented in Table 2. The time component - so important in any measurement of institutionalisation as ^process - is captured by taking into consideration all the years a particular partisan interaction has endured. Finally, in order to avoid measuring incompatible scores, the standardised Z-scores of the three indicators are combined into one unique measurement (i.e. iPSI), paying due attention to stability as the sole dimension of PSI (Casal Bertoa and Enyedi 2016: 268-70).8 Party institutionalisation The discussion in section one provides clear justification for my choice of the dimensions of the concept of PI. Hence, not only should one be interested 412 (g) F. CASAL BERTOA in examining the overall rootedness of political parties, one should also be interested in their organisational systemness, employing Panebianco's (1988) terminology. The degree of PI has been traditionally operationalised in several ways and using multiple indicators: party identification (Dalton and Weldon 2007); levels of professionalisation (Johnson 2002) and personalism (Mainwaring and Torcal 2006), the capacity of parties to sponsor parties cross-nationally (Rose and Mackie 1988), the percentage of independent candidates (Moser 1999), to name only a few. However, notwithstanding their validity, none of these indicators is able to measure at the same time both rootedness and systemness, not even the most widely used indicator of PI - that is, the average age of 'relevant'9 parties (Huntington 1968; Jin 1995; Tavits 2005). This indicator may not only exclude very well rooted and routinised party organisations, even if enjoying limited electoral support, but also fail to capture the process of party development by measuring the level of institutionalisation at just one point in time. Moreover, this indicator clearly discriminates against new democracies, favouring countries with previous democratic experience. Last but not least, the average age may well hide the fact that only one or two parties are well institutionalised, while the others are not.10 For all these reasons, I turn here to Lewis' (2006) Index of Party Stabilisation (IPS), which 'weights and "rewards" the electoral achievement [rootedness] of parties in a sequence of elections' through the progressive enhancement (systemness) of the proportion of the total vote for political parties in a given election over time - by 20 per cent for a party's second appearance in parliament, 40 per cent for the third, and so forth (Lewis 2006: 574-5). Therefore, if a parliamentary party gets 35 per cent of the votes in the first election, 40 per cent in the second, and 10 per cent in the third; a second parliamentary party gets 65 per cent in the first election, 50 per cent in the second, and 80 per cent in the third; and a third party gets into parliament only after the second election with 10 per cent of the vote; the IPS is calculated in the following way: for the first election 35+65-100, for the second [48 (40 + 40*0.2) + 60 (50 + 50*0.2) + 10] - 118, and for the third [14(10 +10*0.4) +112 (80 + 80*0.4) + 12(10+ 10*0.2)] - 138. Then, the sum of the enhanced representation score in the three elections is divided by a notional total score of 360 (i.e. 100 per cent for the first election, 120 per cent for the second, and 140 per cent for the third) and multiplied by 100. The logic is that, taking notice of both voter stability in voters' electoral preferences (rootedness) and the age of a party organisation (systemness),11 the IPS measures the two dimensions of PI together, providing us with a final measure of the institutionalisation of political parties in a country at the state level.12 WEST EUROPEAN POLITICS @ 413 Table 3. Necessary, but not sufficient, conditions. Cause absent Cause present Outcome present 1. No cases here 2. Cases here Outcome absent 3. Not directly relevant 4. Not directly relevant Source: Ragin (2003:182). Method and data The main goal of this article is to empirically test if either PSI or PI (or both) should be considered necessary and/or sufficient conditions for the survival of democracy as the bulk of the scholarly literature maintains. Traditional statistical analysis, based on correlational and linear-additive models, is unable to provide a proper answer to this question,13 but Boolean algebra certainly does (Caramani 2009; Goertz 2003; Mello 2013; Ragin 2008). Although constantly referred to in the literature, it has not been until recently that scholars have started to emphasise the relevance of necessary conditions (Goertz 2006; Goertz and Starr 2003). Because this type of conditions has to be fulfilled every time an outcome is present (see Table 3), necessary conditions are essential to 'the process of coaxing generalizations from empirical evidence' (Ragin 2003: 179) and, therefore, should be examined appropriately. Traditionally, scholars would simply try to cluster the different cases available into the two-by-two arrangement shown in Table 3. So if some cases would fall in sector 2 while sector 1 remained empty, the specific condition was considered to be necessary for the outcome. If this was not the case, the necessary character of the condition would be denied (Caramani 2009). Currently, thanks to Ragin's (2008) Qualitative Comparative Analysis (QCA) method, it is possible not only to establish whether a condition (i.e. factor) is necessary for the outcome, but also to qualify such statement (e.g. 'almost always'). For this Ragin's measurements of consistency and coverage of a condition are essential. Consistency indicates to what extent instances of the outcome constitute a subset of the condition, and is calculated by dividing the number of cases where both the condition and the outcome are present by the number of cases with the outcome. In turn, coverage gauges the importance of a consistent subset, and is obtained by dividing the number of cases where both the condition and the outcome are present by the number of cases with the condition (Ragin 2008; Schneider and Wagemann 2012). These same measurements also help to assess at the same time if a certain condition is (or is not) sufficient for the outcome, with the latter taking place every time the former is present. And this is so because coverage scores for necessity also work as consistency scores for sufficiency, while necessity consistency figures display the coverage of sufficient conditions (Ragin 2008). All in all, as it is now well established in the literature (Ragin 2003, 2008; Schneider and 414 (g) F. CASAL BERTOA Wagemann 2012), the consistency/coverage threshold for necessary/sufficient conditions is 0.9, while for 'almost necessary/consistent' conditions is just 0.8. In order to test the relationship between P(S)I and democracy, I have built a new dataset comprising 64 democratic European political regimes between 1848 and 2015.14 Because I am interested in the impact party competition has on the survival of democracy, a country is considered to be democratic only when (1) it displays at least a score of 6 in the Polity IV index, (2) universal (male) suffrage elections have been held at least once, and (3) governments are formed (and rely) on a parliamentary majority, rather than on the exclusive will of the head of state.15 Moreover, because time has proved to be as important for both PI (Dix 1992) and PSI (Mair 1997) as for democratic survival (Huntington 1991), I will only analyse here the first 25 years after the (re-)inauguration of democracy, as denned above. This will allow me to evaluate analogous periods and avoid faulty comparisons (Casal Bertoa and Mair 2012: 105).16 Institutionalisation and democracy: reality or myth? Paraphrasing Schattschneider (1942), few scholars would currently question that modern democracy is unthinkable save in terms of stable parties and party systems. The idea is that institutionalized party systems ... increase democratic governability and legitimacy by facilitating legislative support for government policies; by channelling demands and conflicts through established procedures; by reducing the scope for populist demagogues to win power; and by making the democratic process more inclusive, accessible, representative, and effective. (Diamond 1997: xxiii) Nevertheless, the institutionalisation of political parties themselves is considered to have positive implications not only for democratic accountability and responsiveness, but also in terms of the linkage between citizens and the state (Mainwaring et at. 1992; Zielinski et at. 2005). Moreover, as Mainwaring has constantly repeated, the problem is that when institutionalisation does not take place, either at the supra- or at the infra-level, citizens may become increasingly frustrated with the (democratic) system, leading not only to high levels of social dissatisfaction (e.g. mass demonstrations) and political disengagement (e.g. low turnout), but also to the appearance (and electoral success) of populist parties and demagogic leaders threatening the survival of the regime (Mainwaring 1998,1999; see also Innes 2002; McGuire 1997). Indeed, as Diamond and Linz (1989:21) already stated almost 30 years ago 'the historical evidence ... suggests that the crucial consideration for democracy is ... the degree of party [system] institutionalisation'.17 It is for all these reasons that when dealing with the question of democratic survival and collapse both types of institutionalisation need to be approached complementarily. Based on the indicators explained in the previous sections, Figures 1 and 218 summarise the levels of PSI and PI in Europe since 1848, with countries where WEST EUROPEAN POLITICS @ 415 6.0 4.0 -8.0 ■ -10.0 --12.0 ■ ■ Party System Institutionalization [iPSI] Figure 1. PSI in Europe {N = 64). Source: Own calculations. Note: Cases of democratic collapse are displayed in capital letters. All countries where iPSI >0 are considered to be institutionalised. 30 ■ 20 -I tyi.L'''mco.Lvo_'J co m t o er» cr> ( O * IM JHrtHHHHHNrlHHHNNHHHHHH^HH IS- t to co re I ; "b o> > I tfl £ 2 w ffl tn< 35 ■S Si oä ■ Party Institutionalization (IPS) Figure 2. PI in Europe {N = 57). Source: Own calculations. Note: Cases of democratic collapse are displayed in capital letters. All countries where IPS > 84 are considered to have institutionalised political parties. democracy collapsed (Polity IV < 6) in capital letters. A first look at both figures reveals that both political parties and party systems in Europe have institutionalised at different rates and in different ways. Secondly as explained elsewhere (Casal Bertoa 2013), there seems to be a geographical difference in terms of institutionalisation, as both political parties and party systems seem to have institutionalised during the same period (i.e. 25 years) at a faster pace in the western than in the eastern part of the continent. Thirdly, a definitively more important observation, notwithstanding the expected (positive) correlation (r = 0.63, sig. at 0.01) between PSI and PI (Casal Bertoa 2016), the two do not 416 (g) F. CASAL BERTOA exactly correspond, therefore justifying the need for a close examination of these two processes independently from each other. PSI and democracy One of the first interesting findings to follow from Figure 1 is that, even when compared in their infancy, post-communist party systems stand as extremely inchoate in comparison with their Western and Southern European counterparts. Indeed, not only the majority of under-institutionalised (i.e. iPSI < 0)19 European party systems are in post-communist Europe, but also none of these post-communist party systems can be considered strongly (i.e. iPSI > 1.5) institutionalised. The only exception is Montenegro which, together with three 'classic cases' in the list (i.e. Switzerland, Austria and Malta), is to be classified as 'over-institutionalised' (i.e. iPSI > 3). Another important discovery is that among all those democracies that did not collapse, the earlier a polity democratised the better: namely, PSI tends to be stronger among the earliest democratised polities. Thus, it is not only that the first democracy in Europe has the most institutionalised party system, but also most post-World War I party systems tend to be more institutionalised than post-World War II, post-fascist and post-communist democracies, in this order. This seems to confirm previous findings that the earlier the 'time of transition the higher the PSI (Casal Bertoa and Mair 2012; Mainwaring et at. 2016). In terms of the relationship between PSI and the survival of democracy in Europe, a quick look at Figure 1 already disconfirms the often-repeated statement that PSI is a necessary, but not sufficient, condition for the survival of democracy. In fact, up to 10 post-communist countries as well as post-1983 Turkey and Andorra did not experience the collapse of their democracies despite having more or less under-institutionalised party systems. Moreover, PSI does not even come close to being an 'almost necessary' condition for the survival of democracy in Europe. Indeed, the necessity test reveals a consistency score of 0.71, far from the 0.8 threshold recommended in the literature. What the previous analysis certainly uncovers, though, is that PSI should be considered as a 'sufficient'20 condition for the survival of democracy as democracy never collapsed in countries where the structure of partisan interactions had achieved a certain 'minimum' degree of stabilisation. Indeed, contrary to what Stockton (2001: 112, 117) maintained, Figure 1 seems to suggest the existence of a threshold of PSI (iPSI > 0.5) that when surpassed will certainly guarantee the survival of a nation's democratic regime. In fact, the only country where democracy collapsed despite having an institutionalised party system is inter-war Austria. Here external factors like the post-Versailles crisis and the Great Depression, together with the political pressures from Fascist Italy and Nazi Germany, mostly contributed to Dollfuss' 'self-coup' in March 1933 (Gerlich and Campbell 2000: 53-6; Berg-Schlosser 2002: WEST EUROPEAN POLITICS @ 417 Table 4. Cross-tabulations of PSI and democratic collapses/survivals in different 'democratisation' periods. Time period Collapses Survives 1897-1914 3(0) 5(100) 1917-1940 10(10) 7 (85.7) 1945-1973 4(0) 8(100) 1974-1988 0 5(80) 1989-2015 2(0) 20 (45) Total 19(5.3) 45 (71.1) Note: The percentage of institutionalised party systems is shown in parentheses. 314). But even in this instance it is possible to observe, especially after August 1930, an increase in the frequency of partial (and between elections) alternations as well as the appearance of previously unseen patterns of competition/ collaboration.21 For the sake of robustness, I have repeated this analysis using other indicators of PSI instead. The results do not change. Thus, employing Pedersen's index of electoral volatility (TEV) and setting 15 per cent - traditionally considered to be the threshold to identify 'earthquake' elections - as a cutting point also renders PSI as a sufficient (0.92) rather than necessary (0.55) condition for democratic survival. Laakso and Taagepera's (1979) effective number of parliamentary parties' (ENPP)22 also points to the former as an 'almost sufficient' (0.87), but in no way necessary (0.75) condition. However, it could well be that previous scholars had put so much weight in the process of institutionalisation due to the impact low PSI has had on the collapse of democracy in Latin America, Africa or even inter-war Europe. Still, we should remember that causation is not essentially symmetric, meaning that party system under-institutionalisation could be a necessary, but not sufficient, condition for democratic collapse, even if its opposite is not. It clearly follows from Figure 1 that this has been the case since the dawn of modern democracy in Europe in 1848. Thus, with the only exception mentioned above, all other instances of regime collapse display low levels of PSI. Moreover, a detailed analysis of the process of party system development in those 18 cases clearly reveals in almost all instances (post-World War II Turkey and Russia being the only exceptions) an increasing progression towards unpredictability, meaning higher frequency of partial alternations, innovative formulae and newly formed parties in government in the years preceding the collapse of the democratic regime. In consonance with what has been said, the necessity tests reveal a consistency score of 0.92, well above the 0.9 threshold.23 Conversely, the 'sufficiency hypothesis' is totally rejected with independence of the PSI indicator employed. A confirmation of the previous results is also to be found in Table 4, which cross-tabulates the percentage of countries with institutionalised party systems and democratic survival or collapse according to the different 'democratisation periods.24 What Table 4 reveals, however, is that while still relevant, the positive 418 (g) F. CASAL BERTOA impact of PSI for democratic endurance has decreased over time. Thus, we have passed from a period (before World War II) where it was both a necessary and sufficient condition, to a period (during the Southern European wave) where it was a sufficient and 'almost necessary' condition, to a final (i.e. current) period when it is a sufficient, but not necessary, condition for the survival of democracy. This coincides with one of the initial findings that earlier democracies tend to have higher institutionalised party systems. However, if there is one thing that has remained constant since 1848 it is the necessary relationship between the absence of PSI and democratic collapse. Another, perhaps even more relevant finding, follows from a comparison of Table 4 with Table Al in the Appendix. The latter cross-tabulates the percentage of countries with electorally stable (columns 2 and 3) and legislative concentrated (columns 4 and 5) party systems and democratic survival or collapse according to the different 'democratisation periods. Such comparison reveals that iPSI as an indicator of PSI clearly outperforms TEV and ENPP as a sufficient and necessary condition for, respectively, the survival and collapse of democracy in Europe. Thus a simple look at the last rows in Tables 4 and Al clearly shows that while more democracies with stable structures of inter-party competition than with stable electorates or concentrated legislatures survived, the percentage of democracies presenting inchoate competitive patterns that collapsed was also higher than those characterised by high volatility and fragmentation. And this can also be observed at the period-level where, in some cases, the identification of stable electorates (e.g. 1989-2015) or concentrate legislatures (e.g. 1897-1914 and 1945-1973) when predicting democratic survival is not very useful as almost as many democracies of the same systemic characteristics (i.e. low volatility and fragmentation) collapsed. All in all, this seems to confirm the higher suitability of using iPSI, rather than the two other more traditional proxies (i.e. TEV and ENPP), as an indicator when looking at the relationship between PSI and democracy. PI and democracy Similarly to what has been previously observed, there seems also to be a clear geographical cleavage between East and West in terms of PI. In fact, as follows from Figure 2, only five out of 19 Eastern European democracies are considered to have institutionalised (i.e. IPS > 84)25 political parties. Moreover, even Montenegro, which is the post-communist country with the strongest (i.e. institutionalised) political parties, occupies a discrete twenty-fourth position, very far from those Western (e.g. Austria, Italy, Finland, etc.) and Southern (e.g. Cyprus, Spain) European nations where political parties could be considered to be over-institutionalised (i.e. IPS > 95). However, although political parties tend to be more institutionalised in earlier democratic periods, the contrast is not as straightforward as in the case of WEST EUROPEAN POLITICS @ 419 Table 5. Cross-tabulations of PI and democratic collapses/survivals in different 'democratisation' periods. Time period Collapses Survives 1897-1914 n/a 4(75) 1917-1940 8 (62.5) 7(71.4) 1945-1973 3 (66.7) 8 (87.5) 1974-1988 0 5(80) 1989-2015 2(0) 20 (30) Total 13(53.8) 44 (56.8) Note: The percentage of regimes with institutionalised political parties is shown in parentheses. PSI. Thus, even if it is true that most countries democratised between 1945 and 1989 display very high levels of PI, there are instances of the contrary: e.g. the French Fifth Republic or post-1983 Turkey. Moreover, political parties did not manage to institutionalise, at least during the first 25 years of democracy, in inter-war Luxembourg and Ireland as well as during the French Third Republic. And what is also very revealing is that the majority of Southern European political parties clearly outperform most of their post-war (both World War I and World War II) counterparts. Do these different findings between PSI and PI extend to the relationship between institutionalisation and democracy? The answer is yes. Indeed, a quick look at Figure 2 suggests that PI is neither a necessary nor a sufficient condition for the survival of democracy. Thus, not only have there been countries where democracy collapsed despite high levels of PI (e.g. pre-1983 Turkey, inter-war Finland as well as the Hellenic and First Austrian Republics), in most post-communist countries democracy has survived despite the presence of very weakly institutionalised political parties (mainly in Georgia, Latvia, Poland and Slovakia). Necessity (consistency = 0.57) tests confirm the latter. And while in terms of sufficiency PI comes close (0.78), it fails to reach the 0.8 threshold. Similarly, using the 'average party age'26 as a proxy for PI and taking into consideration all 64 democratic party systems instead does not change any of the previous statements and confirms the robustness of these results. In a similar vein, there seems to be no relationship at all, either of necessity or sufficiency (consistency = 0.46 and 0.24, respectively), between democratic collapse and party under-institutionalisation. Indeed, the fact that most regimes with under-institutionalised political parties have survived (see Figure 2) seems to suggest the almost trivial, if not non-existent, relationship between PI and democracy. This is something that can be observed in the last row of Table 5, which shows a very similar percentage for both consolidated and failed democracies with institutionalised political parties, but not only. Thus, even if after World War I regimes with institutionalised political parties tended to survive to a greater extent than those with under-institutionalised political parties, the differences are not very significant (roughly 30 per cent at the most). 420 (g) F. CASAL BERTOA For the sake of robustness, it is important to note here that, as it immediately follows also from Tables A2 and A3 in the Appendix, all of the findings mentioned in the previous two sub-sections remain the same independently of the indicators employed or the countries included in the analysis. Thus, the exclusion of non-multiparty parliamentary democracies (i.e. Malta and Cyprus), microstates27 (e.g. San Marino, Andorra, Iceland, Montenegro) or those countries whose 'democratic credentials' are not so obvious (e.g. Ukraine, Moldova, Kosovo, Albania, Georgia) has no impact at all on the above-cited results. The same can be said even if one was to consider that democracy survived in certain 'controversial' cases - like inter-war Finland or the French Fourth Republic - marked by Polity IV as instances of collapse. Conclusions Summarising an almost unanimous belief within the democratisation literature, Mainwaring (1999:6) stated 15 years ago that 'democracy is likely to have shortcomings if a moderately institutionalized party system does not emerge after democratic government has been in place for some time'. Although never satisfactorily tested, the assertion that PSI was a necessary, but not sufficient, condition for the endurance of democracy became one of the most repeated among both party politics and regime transition scholars. To the point that, driven by the conceptual assimilation embedded in Mainwaring's theoretical framework (Mainwaring 1999; Mainwaring and Scully 1995; Mainwaring and Torcal 2006), such an essential role for democratic survival was extended also to the institutionalisation of political parties themselves. Trying to clarify the extent to which - as repeatedly maintained - both PSI and PI have a positive impact on the endurance of democracy, this article has assessed the above-cited relationship by employing a different methodology on an original dataset, large enough to allow for both cross-national/regional as well as cross-temporal comparisons. These are the main findings. First of all, following Randall and Svasand's (2002) steps, it seems reasonable that PSI and PI are two different, if related, concepts which should not be conflated. As a result, the former should be operationalised with an eye to excluding indicators that measure aspects at the party, rather than the systemic level. Secondly, it has not been the institutionalisation of political parties but the institutionalisation of party systems as a whole that has had a positive effect on the prospects for democratic survival in Europe. However, thirdly, such impact has taken place in a different manner than what most scholars had predicted, as PSI has not been a necessary, but a sufficient condition for the survival of European democracies. In fact, democracy has survived in many post-communist countries despite, sometimes even in spite of, extremely inchoate party systems (e.g. Latvia, Bulgaria, Lithuania, Poland, etc.). What scholars, and politicians, should bear in mind is that continuous lack of PSI WEST EUROPEAN POLITICS @ 421 'necessarily' puts democracy in peril, as most pre-World War II European cases demonstrate. Fourthly, with apologies once again to Schattschneider (1942:1), European democracies do seem to be unthinkable save in terms of moderately institutionalised party systems. In other words, democracy will never collapse as soon as a certain degree of PSI is reached. At least this is what our historical analysis of 64 different European political regimes reveals. Fifthly, when trying to measure PSI or looking for the ways to avoid democratic collapse, scholars might want to consider putting their focus on the stability of partisan interactions rather than on the volatility of electoral attachments or the number of political parties. Finally, party system over-institutionalisation (i.e. iPSI > 3) has not proved as dangerous for the survival of democracy in Europe as some may have predicted (see Coppedge 2004; Schedler 1995). This is not to deny that excessive levels of institutionalisation could harm the quality of a nation's democracy. But this is certainly a topic for future research. All in all, as Sontheimer (1987: 10) noted almost 30 years ago, 'the stability of the party system [rather than the parties] is the really decisive factor for the stability of the whole system in all democratic systems'. Notwithstanding the fact that any generalisation of the results of this article outside the European continent should be treated with caution, its findings have important implications in terms of how democracy should be promoted as - needless to say - the whole question of democratic survival should be approached with a preferential eye on party systems rather than merely on parties, as has usually been the case (e.g. Burnell and Gerrits 2012). In other words, paraphrasing Pridham (1990:2), focusing on party systems must remain a basic if not the central theme for examining the survival of liberal democracy. Notes 1. For a criticism of the positive relationship between institutionalisation and democratic consolidation, see Tóka (1997), Chabal and Daloz (1999) or, more recently, Enyedi (2016). 2. Still, for the sake of robustness, the article also uses other 'traditional indicators' of PSI and PI. 3. Similarly, the article does not deny the eventual contribution of other factors to the survival (e.g. economic development, EU membership, etc.) or collapse (e.g. economic crisis, historical legacies, etc.) of (a particular) democracy. 4. Something which, otherwise, has become one of the most popular (but never confirmed) assumptions in political science. 5. For an in-depth analysis of the problems with Mainwaring and Scullys (1995) framework see Luna's recent work, either with Altman (2011) or alone (2014). 6. For a similar understanding, please see Bakke and Sitter (2005) or Mair (2001). 7. Also because, as Casal Bértoa and Enyedi (2016: 265) point out, it is 'conceptually and empirically superior to the ones suggested by scholars so far'. 8. A step-by-step explanation (examples included) of how iPSI is calculated can be found at: http://whogoverns.eu/wp-content/uploads/2015/02/Rules-for-calculating-iPSI.pdf. (g) F. CASAL BERTOA 9. Usually those with more than 10 per cent of the vote at a given election (Mainwaring and Scully 1995). Conversely, Lewis' IPS considers all political parties with parliamentary representation (at least twice). 10. Randall and Svasand (2002: 9) point here to the so-called unevenness of PI. 11. Including the presence or absence of certain organisational changes (i.e. splits and mergers). 12. Contrary to the title of Lewis' article, IPS looks more at the continuity of party representation-cum-organisation (Lewis 2006: 562, 566) and 'their relative success in elections' (574). I am particularly grateful to one of the reviewers for this point. 13. Logit regression or survival analysis would be perfectly suitable had I wanted to, respectively, explain variation in democratic endurance, measuring the impact PSI and/or PI have had on democratic survival, or analyse how do any of the former factors increase or decrease the probability of the latter phenomenon. However, as mentioned above, this is not the aim of this article. 14. Available at http://whogoverns.eu/cabinets/#. 15. My case selection not only coincides with most studies in the field (Mainwaring et al. 2016), but also mostly overlaps with other similar datasets (Boix et al. 2013; Coppedge et al. 2016). 16. The idea is to avoid situations in which time constitutes the main explanatory variable. Thus, it would be unfair to compare Hungary or any post-communist democracy in 2015 with the UK or any other traditional Western European democracy in the same year as in the latter political parties had four times as long as the former to interact and, therefore, create a cumulative experience helping them to routinise their behaviour making it more predictable and stable. 17. For an in-depth study of why P(S)I should positively affect the level of democracy see also Thames and Robbins (2007). 18. While the Kingdom of Greece (1875-1900), Restoration Spain (1900-1923) and inter-war Portugal (1911-1925) have not been included due to unavailability of the data (Nohlen and Stover 2010: 1539, 1815), the lack of at least a pair of elections in the French Second Republic (1848-1851), the Kingdom of Serbs, Croats and Slovenes (1921), inter-war Poland (1922-1926) and post-World War II Greece (1946-1948) impeded the calculation of the IPS for these four democratic regimes. 19. This distinction between institutionalised (iPSI>0) and under-institutionalised (iPSI < 0) party systems is consistent with other previous large-N (Casal Bertoa 2013; Tavits 2005) and small-N (Berg-Schlosser and Mitchell 2000) comparative studies. Changes in the threshold (e.g. iPSI > 0) do not alter the results (available from the author upon request). 20. Coverage equals 0.97. 21. Thus, in an unprecedented move, representatives of the extreme right-wing 'Pro Patria bloc were granted access to government in September 1930. Less than three years later (i.e. in January 1932), the nationalist Greater German People's Party (GDVP) would be excluded from a coalition government between the Christian Social Party (CS) and the Rural Party (LB) for the first time. 22. Following Mainwaring and Scully (1995: 31-2), the cutting point here is an ENPP of 4, which distinguishes extreme pluralist' (ENPP > 4) party systems from the rest: namely, limited pluralist' (ENPP between 3 and 3.9), two-and-a-half (ENPP between 2.5 and 2.9) and bi-party (ENPP < 2.5) systems. 23. TEV, although not ENPP (0.62), also points to the lack of PSI as an almost necessary' (0.85) condition for the collapse of a democracy. WEST EUROPEAN POLITICS @ 423 24. These are clearly distinguishable periods, separated by a major event or structural juncture' (e.g. world war, end of the Cold War, etc.), in which a clearly identifiable group of countries (e.g. Southern European, post-communist, etc.) were (re-)incorporated into the family of democratic nations. 25. This cut-off point has been chosen according to the following criteria: (1) it displays by far the higher gap in the level of PI between two (consecutive) countries, (2) it perfectly coincides with the average European level of PI in the period under study, (3) it divides the sample in two almost equal clusters, and (4) it is consistent with the literature (e.g. Janda 1980). 26. Significantly (at 0.01 level) correlated, in any case, with IPS (r = 0.63). 27. Those with less than 1 million inhabitants. Acknowledgements I would like to thank the participants of a panel on 'The Struggle for Democracy: Historical Accounts' at the EPSA Conference in Edinburgh (2014) as well as the participants of the Seminar Series held at the Department of Comparative Politics of the University of Bergen, and especially Lars Svasand, Michael Alvarez and Yvette Peters, together with Kevin Keegan-Krause and Zsolt Enyedi, for their value comments and suggestions on previous versions of this article. Disclosure statement No potential conflict of interest was reported by the author. Notes on contributor Fernando CasalBertoa is a Nottingham Research Fellow of the School of Politics and International Relations at the University of Nottingham. He is also Co-chair of the Council for European Studies' Research Network on 'Political Parties, Party Systems and Elections' as well as Research Theme director at the Nottingham Interdisciplinary Centre for Economic and Political Research. His work has been published in Sociological Methods and Research, Party Politics, Democratization, Government and Opposition, International Political Science Review, South European Society and Politics and East European Politics. His most recent (co-edited) book is Party Politics and Democracy in Europe: Essays in Honour of Peter Mair (Routledge, 2016). [fernando.casal.bertoa@ nottingham.ac.uk] References Abeje, Alefe (2013). 'Party System Institutionalization and the Prospect of Democratic Consolidation in Ethiopia, European Science Journal, 9:23, 315-328. Adcock, Robert, and David Collier (2001). 'Measurement Validity: A Shared Standard for Qualitative and Quantitative Research', American Political Science Review, 95:03, 529-546. 424 (g) F. CASAL BÉRTOA Arter, David, and Elina Kestilä-Kekkonen (2014). 'Measuring the Extent of Party Institutionalization: The Case of a Populist Entrepreneur Party', West European Politics, 37:5, 932-956. Bakke, Elisabeth, and Nick Sitter (2005). 'Patterns of Stability. Party Competition and Strategy in Central Europe since 1989', Party Politics, 11: 2, 243-263. Basedau, Matthias (2007). 'Do Party Systems Matter for Democracy?', in Matthias Basedau, Erdmann Gero and Andreas Mehler (eds.), Votes, Money and Violence. Political Parties and Elections in Sub-Saharan Africa. Sweden and South Africa: Nordiska Afrikainstitutet and University of Kwazuiu-Natai Press, 105-143. Berg-Schlosser, Dirk (2002). 'Cross-sectional and Longitudinal Analysis', in Dirk Berg- Schlosser and Jeremy Mitchell (eds.), Authoritarianism and Democracy in Europe, 1919-1939. Comparative Analyses. London: Palgrave, 285-315. Berg-Schlosser, Dirk, and Jeremy Mitchell (2000). Conditions of Democracy in Europe, 1919-1939. London: Macmillan. Bielasiak, Jack (2001). On the Institutionalization of Party Regimes in Emerging Democracies, Working Paper, No. 351. Glasgow: University of Strathclyde. Birch, Sarah (2003). Electoral System and Political Transformation in Post-Communist Europe. New York, NY: Palgrave Macmillan. Boix, Carles, Michael Miller, and Sebastian Rosato (2013). A Complete Data Set of Political Regimes, 1800-2007', Comparative Political Studies, 46:12, 1523-1554. Booth, Eric, and Joseph Robbins (2010). Assessing the Impact of Campaign Finance on Party System Institutionalization', Party Politics, 16:5, 629-650. Burnell, Peter, and André Gerrits (2012). Promoting Party Politics in Emerging Democracies. London: Routledge. Caramani, Daniele (2009). Introduction to the Comparative Method with Boolean Algebra. Los Angeles, CA: Sage. Casal Bértoa, Fernando (2013). 'Post-Communist Politics: On the Divergence (and/ or Convergence) of East and West', Government and Opposition, 48:03, 398-433. Casal Bértoa, Fernando (2016). 'Party System Institutionalisation: A Travelling Concept?', in Ferdinand Müller-Rommel and Fernando Casal Bértoa (eds.), Party Politics and Democracy in Europe. Abingdon: Routledge, 117-130. Casal Bértoa, Fernando, and Zsolt Enyedi (2016). 'Party System Closure and Openness: Conceptualization, Operationalization, and Validation, Party Politics, 22:3,265-277. Casal Bértoa, Fernando, and Peter Mair (2012). 'Party System Institutionalization across Time in Post-Communist Europe', in Ferdinand Müller-Rommel and Hans Keman (eds.), Party Government in the New Europe. London: Routledge, 85-112. Chabal, Pascal, and Jean-Pascal Daloz (1999). Africa Works. Disorder as Political Instrument. Oxford: James Currey Publishers. Coppedge, Michael (2004). Strong Parties and Lame Ducks: Presidential Partyarchy and Factionalism in Venezuela. Stanford: Stanford University Press. Coppedge, Michael, John Gerring, Staffan Lindberg, Sven-Erik Skaaning, and Jan Teorell (2016). 'V-Dem Codebook v6', Varieties of Democracy (V-Dem) Project. Dalton, Russel, and Steven Weldon (2007). 'Partisanship and Party System Institutionalization, Party Politics, 13:2, 179-196. Diamond, Larry (1997). 'Introduction: In search of consolidation, in Larry Diamond, Marc F. Plattner, Yun-han Chu, and Hung-Mao Tien (eds.), Consolidating the Third Wave Democracies. Baltimore, MD: John Hopkins University Press, xiii-xl. Diamond, Larry, and Juan Linz (1989). 'Introduction: Politics, Society, and Democracy in Latin America', in Larry Diamond, Juan Linz and Seymour Martin Lipset (eds.), Democracy in Developing Countries: Latin America. Boulder: Lynne Reinner, 1-58. WEST EUROPEAN POLITICS @ 425 Dix, Robert (1992). 'Democratization and the Institutionalization of Latin American Political Parties', Comparative Political Studies, 24:4, 488-511. Enyedi, Zsolt (2016). 'Populist Polarization and Party System Institutionalization: The Role of Party Politics in De-Democratization', Problems of Post-Communism, 63:4, 210-220. Gerlich, Peter, and David Campbell (2000). 'Austria: From Compromise to Authoritarianism', in Dirk Berg-Schlosser and Jeremy Mitchell (eds.), Conditions of Democracy in Europe, 1919-1939. London: Macmillan, 40-58. Goertz, Gary (2003). 'Cause, Correlation, and Necessary Conditions', in Gary Goertz and Harvey Starr (eds.), Necessary Conditions: Theory, Methodology and Applications. Rowman: Lanham, 47-65. Goertz, Gary (2006). Assessing the Trivialness, Relevance, and Relative Importance of Necessary or Sufficient Conditions in Social Science', Studies in Comparative International Development, 41:2, 88-109. Goertz, Gary, and Harvey Starr (2003). Necessary Conditions: Theory, Methodology and Applications. Lanham: Rowman and Littlefield. Gunther, Richard, and Jonathan Hopkin (2002). A Crisis of Institutionalization: The Collapse of the UCD in Spain, in Richard Gunther, Jose Ramon Montero and Juan Linz (eds.), Political Parties: Old Concepts and New Challenges. Oxford: Oxford University Press, 191-229. Harmel, Robert, and Lars Svasand (1993). 'Party Leadership and Party Institutionalization: Three Phases of Development', West European Politics, 16:2, 67-88. Harmel, Robert, Lars Svasand, and Hilmar Mjelde (forthcoming). Institutionalization (and Deinstitutionalization) of Right-wing Protest Parties. Colchester: ECPR Press Hicken, Allen (2006). 'Stuck in the Mud. Parties and Party Systems in Democratic Southeast Asia, Taiwan Journal of Democracy, 2:2, 23-46. Hicken, Allen, and Erik Kuchonta (2014). Party System Institutionalization in Asia. Cambridge: Cambridge University Press. Huntington, Samuel (1968). Political Order in Changing Societies. New Haven, CT: Yale University Press. Huntington, Samuel (1991). The Third Wave: Democratization in the Late Twentieth Century. Norman, OK: University of Oklahoma Press. Innes, Abby (2002). 'Party Competition in Post-Communist Europe', Comparative Politics, 35:1,85-104. Janda, Kenneth (1980). Political Parties: A Cross National Survey. New York, NY: Free Press. Jin, Youngjae (1995). 'Testing Political Party Institutionalization: A Theory and Practice', Journal of Political and Military Sociology, 23:1, 43-63. Johnson, Elaine (2002). Streams of Least Resistance: The Institutionalization of Political Parties and Democracy in Indonesia. Ph.D. Dissertation. Charlottesville: University of Virginia. Kneuer, Marianne (2011). 'Deficits in Democratic Quality? The Effects of Party System Institutionalization on the Quality of Democracy in Central Eastern Europe', Zeitschrift fur vergleichende Politikwissenschaft, 1,133-172. Kreuzer, Markus, and Velio Pettai (2003). 'Patterns of Political Instability: Affiliation Patterns of Politicians and Voters in Post-Communist Estonia, Latvia, and Lithuania', Studies in Comparative International Development, 38:2, 76-98. Kuenzi, M., and G. Lambright (2005). 'Party Systems and Democratic Consolidation in Africa's Electoral Regimes', Party Politics, 11:4, 423-446. Laakso, Markku, and Rein Taagepera (1979). ""Effective" Number of Parties', A Measure with Application to West Europe, Comparative Political Studies, 12:1, 3-27. 426 (g) F. CASAL BERTOA Levitsky, Steven (1998). 'Institutionalization and Peronism: The Concept, the Case and the Case for Unpacking the Concept', Party Politics, 4:1, 77-92. Lewis, Paul (1994). 'Democratization and Party Development in Eastern Europe', Democratization, 1:2, 391-405. Lewis, Paul (2006). 'Party Systems in Post-Communist Central Europe: Patterns of Stability and Consolidation', Democratization, 13:4, 562-583. Lindberg, Staffan (2007). 'Institutionalization of Party Systems? Stability and Fluidity among Legislative Parties in Africa's Democracies', Government and Opposition, 42:2, 215-241. Luna, Juan Pablo (2014). 'Party System Institutionalization: Do We Need a New Concept?', Studies in Comparative International Development, 49:4, 403-425. Luna, Juan Pablo, and David Altman (2011). 'Uprooted but Stable: Chilean Parties and the Concept of Party System Institutionalization, Latin American Politics and Society, 53:2, 1-28. Mainwaring, Scott (1998). 'Party Systems in the Third Wave', Journal of Democracy, 9:3, 67-81. Mainwaring, Scott (1999). Rethinking Party Systems in the Third Wave of Democratization: The Case of Brazil. Stanford: Stanford University Press. Mainwaring, Scott, and Timothy Scully (1995). Building Democratic Institutions: Party Systems in Latin America. Stanford: Stanford University Press. Mainwaring, Scott, and Mariano Torcal (2006). 'Party System Institutionalization and Party System Theory after the Third Wave of Democratization, in Richard Katz and William Crotty (eds.), Handbook of Party Politics. London: Sage, 204-227. Mainwaring, Scott, Guillermo O'Donnell, and Samuel Valenzuela (1992). 'Introduction, in Scott Mainwaring, Guillermo O'Donnell and Samuel Valenzuela (eds.), Issues in Democratic Consolidation: The New South American Democracies in Comparative Perspective. Notre Dame: University of Notre Dame Press, 1-16. Mainwaring, Scott, Carlos Gervasoni, and Annabella Espana (2016). 'Extra- and Within-system Electoral Volatility', Party Politics, doi:10.1177/1354068815625229. Mair, Peter (1997). Party System Change. Approaches and Interpretations. Oxford: Oxford University Press. Mair, Peter (2001). 'The Freezing Hypothesis: An Evaluation', in Lauri Karvonen and Stein Kuhnle (eds.), Party Systems and Voter Alignments Revisited. London: Routledge, 24-41. Mair, Peter (2006). 'Party System Change', in Richard Katz and William Crotty (eds.), Handbook of Party Politics. London: Sage, 63-73. Markowski, Radoslaw (2001). 'Party System Institutionalization in New Democracies: Poland - A Trend-setter with no Followers', in Paul Lewis (ed.), Party Development and Democratic Change in Post-communist Europe: The First Decade. London: Frank Cass, 55-77. McGuire, James (1997). Peronism without Peron: Unions, Parties, and Democracy in Argentina. Stanford: Stanford University Press. Meleshevich, Andriy (2007). Party Systems in Post-Soviet Countries: A Comparative Study of Political Institutionalization in the Baltic States, Russia, and Ukraine. New York, NY: Palgrave. Mello, Patrick (2013). 'From Prospect to Practice: A Critical Review of Applications in Fuzzy-set Qualitative Comparative Analysis', Paper presented at the 8th Pan-European Conference on International Relations in Warsaw, 18-21 September. Morlino, Leonardo (1998). Democracy between Consolidation and Crisis: Parties, Groups, and Citizens in Southern Europe. Oxford: Oxford University Press. WEST EUROPEAN POLITICS @ 427 Moser, Robert (1999). 'Electoral Systems and the Number of Parties in Post-Communist States', World Politics, 51:03, 359-384. Müller, Wolfgang C, and Franz Fallend (2004). 'Changing Patterns of Party Competition in Austria: From Multipolar to Bipolar System, West European Politics, 27:5,801-835. Nohlen, Dieter, and Philip Stöver (2010). Elections in Europe: A Data Handbook. Baden-Baden: Nomos. Panebianco, Angelo (1988). Political Parties: Organization and Power. Cambridge: Cambridge University Press. Pedersen, Mogens (1979). "The Dynamics of European Party Systems: Changing Patterns of Electoral Volatility', European Journal of Political Research, 7:1,1-26. della Porta, Donatella, and Michael Keating, eds. (2008). Approaches and Methodologies in the Social Sciences. Cambridge: Cambridge University Press. Powell, Eleanor, and Joshua Tucker (2014). 'Revisiting Electoral Volatility in Post-Communist Countries: New Data, New Results and New Approaches', British Journal of Political Science, 44:01, 123-147. Pridham, Geoffrey (1990). Securing Democracy: Political Parties and Democratic Consolidation in Southern Europe. New York, NY: Routledge. Ragin, Charles (2003). 'Fuzzy-Set Analysis of Necessary Conditions', in Gary Goertz and Harvey Starr (eds.), Necessary Conditions: Theory, Methodology and Applications. Lanham: Rowman and Littlefield, 179-196. Ragin, Charles (2008). Redesigning Social Inquiry: Fuzzy Sets and Beyond. Chicago, IL: University of Chicago Press. Randall, Vicky, and Lars Svasand (2002). 'Party Institutionalization in New Democracies', Party Politics, 8:1, 5-29. Riker, William, and Peter Ordeshook (1973). An Introduction to Positive Political Theory. Englewood Cliffs, NJ: Prentice Hall. Rokkan, Stein (1970). Citizens, Elections, Parties: Approaches to the Comparative Study of the Processes of Development. Oslo: Universitetsforlaget. Rose, Richard, and Thomas Mackie (1988). 'Do Parties Persist or Fail? The Big Tradeoff Facing Organizations', in Kay Lawson and Peter Merkl (eds.), When Parties Fail: Emerging Alternative Organizations. Princeton, NJ: Princeton University Press, 533-560. Rose, Richard, and Neil Munro (2009). Parties and Elections in New European Democracies. Colchester: ECPR Press. Sartori, Giovanni (1976). Parties and Party Systems. A Framework for Analysis, Volume I. Cambridge: Cambridge University Press. Schattschneider, Elmer (1942). Party Government. New York, NY: Farrar and Reinhart. Schedler, Andreas (1995). Under- and Over institutionalization: Some Ideal Typical Propositions Concerning New and Old Party Systems, Working Paper, N. 213. Notre Dame: The Helen Kellogg Institute for International Studies. Schneider, Carsten, and Claudius Wagemann (2012). Set-Theoretic Methods for the Social Sciences: A Guide to Qualitative Comparative Analysis. Cambridge: Cambridge University Press. Smith, Gordon (1989). A System Perspective on Party System Change', Journal of Theoretical Politics, 13:3, 34-63. Stockton, Hans (2001). 'Political Parties, Party Systems, and Democracy in East Asia: Lessons from Latin America', Comparative Political Studies, 34:1, 94-119. Tavits, Margit (2005). 'The Development of Stable Party Support: Electoral Dynamics in Post-Communist Europe', American Journal of Political Science, 49:2, 283-298. 428 (g) F. CASAL BERTOA Thames, Frank, and Joseph Robbins (2007). 'Party System Institutionalization and the Level of Democracy', Paper presented at the Annual Meeting of the American Political Science Association, Chicago, August, 29-September, 1. Töka, Gabor (1997). Political Parties and Democratic Consolidation in East Central Europe, Working Paper No 279. Glasgow: University of Strathclyde. Wallis, Darren (2003). 'Democratizing a hegemonic regime: From institutionalized party to institutionalized party system in Mexico?', Democratization, 10:3, 15-38. Webb, Paul, and Stephen White (2007). Party Politics in New Democracies. Oxford: Oxford University Press. Weghorst, Keith, and Michael Bernhard (2014). 'From Formlessness to Structure? The Institutionalization of Competitive Party Systems in Africa, Comparative Political Studies, 47:2, 1707-1737. Zielinski, Jakub, Kazimierz M. Slomczynski, and Goldie Shabad (2005). 'Electoral Control in New Democracies: The Perverse Incentives of Fluid Party Systems', World Politics, 57:03, 365-395. Appendix Table A1. Cross-tabulations of TEV and ENPP and democratic collapses/survivals in different 'democratisation' periods. Time period Collapses Survives Collapses Survives PSI = TEV< 15% PSI = ENEP<4 1897-1914 n/a 4(100) 3 [100] 5 [100] 1917-1940 8(25) 7(71.4) 10 [30] 7 [71.4] 1945-1973 3(0) 8(100) 4 [75] 8 [87.5] 1974-1988 0 5(80) 0 5 [100] 1989-2015 2(0) 20(15) 2[0] 20 [60] Total 13 (15.4) 44 (54.5) 19 [47.4] 45 [75.6] Note: The percentage of volatile/fragmented party systems is shown in parentheses. Table A2. Party system (under-)institutionalisation and democratic survival/collapse in Europe (1848-2015). Survival Collapse Partysysteminsti- Switzerland, Austria II, Liechtenstein, tutionalisa- Montenegro, Malta, San Marino II, UK, tionfiPSI > 0) The Netherlands, Greece IV, Denmark, Ireland, Belgium, France IV, Italy, Spain III, Portugal II, Greece I, Iceland, West Germany, Norway, Georgia, Czechoslovakia, Romania, Hungary, Finland II, Albania, France II, Luxembourg, Macedonia, Cyprus, Croatia, Kosovo Partysystemu- Latvia II, Bulgaria, Lithuania, Poland II, nder-institu- Turkey III, Andorra, Moldova, Serbia, tionalisation Czech Republic, Estonia II, Slovenia, (iPSI < 0) Slovakia, Sweden Austria I France I, Poland I, Greece III, Portugal I, San Marino I, Russia, Ukraine, Finland I, Estonia I, Spain II, Weimar Republic, Turkey II, Latvia I, Spain I, France III, KSHS, Turkey II, Greece II WEST EUROPEAN POLITICS @ 429 O 84) PartyUnder-insti-tutionalisation (IPS < 84) Austria II, UK, Liechtenstein, Italy, Finland II, Switzerland, Sweden, San Marino II, Cyprus, Spain III, Denmark, Malta, Greece IV, Belgium, Portugal II, West Germany, Norway, Iceland,The Netherlands, Montenegro, Czechoslovakia, Hungary, Albania, Romania, Croatia France II, Serbia, Luxembourg, Macedonia, France IV, Czech Republic, Ireland, Estonia II, Lithuania, Andorra, Slovenia, Kosovo, Bulgaria, Moldova,Turkey III, Slovakia, Poland II, Latvia III, Georgia Turkey II, Greece II, Austria I, Turkey II, Finland I, Weimar Republic, Spain II Latvia I, France III, Estonia I, Russia, Ukraine, San Marino I > -t—» B > > o M >> aj O o Q