0

Inter University Attraction Pole

Participation and Representation (PartiRep)

Brussel/Bruxelles, Antwerp, Leuven, Leiden

Party cartelization and party membership in Belgium and Italy

Giulia Sandri, gsandri@ulb.ac.be and Teun Pauwels, tepauwel@ulb.ac.be CEVIPOL, Université libre de Bruxelles

Paper prepared for presentation at the PARTIREP International Conference

‘Party membership in Europe. Explorations into the anthills of party politics’

Université Libre de Bruxelles, Brussels, Belgium, 30-31 October 2008

FIRST DRAFT, PLEASE DO NOT QUOTE

1

1. Introduction

Party change in organizational terms has been conceptualized by different party models in the

literature: cadre or individual integration parties (Ostrogorski, 1903; Neumann, 1956), mass parties or parties of social integration (Duverger, 1954; Michels, 1966; Seiler, 1993; Wolinetz, 2002), catch-all parties (Kirchheimer, 1966), professional-electoral parties (Panebianco, 1988), modern cadre parties (Koole, 1996), business-firm parties (Hopkin and Paolucci, 1999), parties as franchise organizations (Carty, 2004), parties of political entrepreneurs (Krouwel, 2006) and parties of professional politicians (von Beyme, 1996). The cartel party model, whilst sometimes considered as controversial (Koole, 1996), represents the current analytical endpoint of party politics research (Bolleyer, 2007). Therefore, it would be interesting to contribute to the further assessment and theoretical development of the model.

Even though the cartel party theory has triggered a flamboyant literature on the empirical assessment of the model, there are few quantitative and cross-national attempts aimed at verifying the explanatory potential of the model (Aucante and Dézé, 2008; Pelizzo, 2005, 2007 and 2008). Even less numerous are the studies that try to apply the model at the organizational-individual party level and not only at the party system level (Detterbeck, 2005; Bolleyer, 2007 and 2008). Therefore, in this paper we will set out a research outline aimed at assessing the cartel party model on empirical grounds by following quantitative methods. We would also like to focus on the individual party level and not on the systemic one, in order to fill the gap in the cartel party literature concerning the analysis at the organizational level.

In their seminal article on cartel party, Katz and Mair claim that political parties increasingly migrate to the state and lose their connection with society. Having become mainly office-seeking, current parties tend to reinforce inter-party cooperation and to set up barriers to the entry of new challengers in the party system, while they depend increasingly on state resources for financing their activities. These phenomena are thought to feed back internally on contemporary parties in organizational terms. The core properties of the cartel party model are, in fact, not only the professionalization of politics and the capital intensive nature of party work and party campaigning, but also the new basis of party competition, i.e. managerial skills and efficiency, and the stratarchical relations between members and party elite. The latter property, meaning that members and elites are mutually autonomous, is linked to the blurred distinction between members and supporters. The overall balance between members’ rights and obligations tend to emphasize members’ privileges, but in the end neither right nor obligations are important. Moreover, as we will explain further in the next paragraph, the qualitative shifts in the relationship between members and elites are the sole features of cartel party that transcend the intensification of the previous catch-all model (Bolleyer, 2008).

This paper aims at exploring the role of party membership and activism in contemporary parties in Belgium and Italy in order to assess their degree of party cartelization. It will be examined empirically to what extent the internal organization of the selected parties correspond to the stratarchy model of the relationships between grass-roots members and party elites (Carty, 2004) as well as whether blurred distinctions among members and non-members are equally applicable to the analysis of the selected cases. Therefore, our paper will first identify the two core dimensions of the cartel party, considered as a model of party organizational change. Then, we will try to verify whether the two mentioned characteristics of the model, the permeable organizational boundaries and the stratarchical internal organization, are linked to specific external elements, as for example the party’s ideological location and the country effect, in order to explore the main variables that might affect the degree of intra-organizational cartelization of current parties.

2. Theoretical framework and research questions

Political parties are considered, in general, as an essential element of modern democracy

(Schattschneider, 1942) and as the basic mechanism of linkage between civil society and the state (Lawson and Merkl, 1988). The latter has been the main function of the traditional mass parties as described by Duverger (1954). Nevertheless, according to the cartel party literature (Katz and Mair, 1995 and 1996; Katz, 2001; Mair, 1997; Blyth and Katz, 2005), a new trend emerged since the late 1970s: the erosion of

2

traditional social boundaries caused the weakening of formerly highly distinctive collective identities and therefore the ideological and political distinctiveness of parties became rather blurred. The linkage role of parties retrenched and new party models emerged. According to Katz and Mair (1995: 6), during the 20th century the dialectical process of party development led to the emergence of a new typology of party, the cartel party model, which is characterized not only by centrifugal dynamics of interaction between parties and civil societies, but also by tight relations between parties and the state.

The model is widely considered as fitted to capture the contemporary stage of party organizational development (Bolleyer, 2007). Starting from the assumption that the cartel party model is actually emerging and constitutes the main theoretical framework for classifying and analyzing the great majority of West European parties, we will assess the degree of cartelization of contemporary parties through the analysis of the shape and role of party membership. Party membership constitutes a particular type of political participation and it is evidently pertinent to take it into account when studying intra-organisational aspects of parties’ nature and functions. Party system features, party competition patterns and other settings at the macro-level are thought to feed back into the internal life of individual party organisation (Mair, 2005: 17).

Even considering the main theoretical and empirical critiques developed with regard to the cartel party model (Koole, 1996; Kitschelt, 2000; Pierre et al., 2000), many scholars consider this classification model as the better proxy for grasping the empirical reality and the current dominant features of West European parties (Raniolo, 2000; Ignazi, 2004). Several scholars have pointed out that the cartel party model focuses mainly on describing systemic features and less on explaining individual party characteristics (Koole, 1996; von Beyme, 1996, Seiler, 2008), but starting from the seminal work of Katz and Mair, many researchers have empirically tested and implemented the model as well as developed a large set of cartelization indicators, even though mainly at systemic level (for example, Bardi, 2006; Detterbeck, 2005 and 2008; Conti, Cotta and Tronconi, 2006 and 2008; Bolleyer, 2007 , Aucante and Dézé, 2008, Delwit, 2008 and Pelizzo, 2005, 2007 and 2008). They have therefore verified and developed the explanatory potential of the cartel party model. This model may be considered more as a set of hypotheses and suggestions on party change processes than a cohesive and structured model of party nature, but, considering the current state of the art of the scientific research on party types, it is widely accepted as a well defined analytical framework and it is generally integrated in the main literature on party models (Gunther, Montero and Linz, 2002; Krouwel, 2006).

Among all the studies concerning the relationships between party members and party organization, several different theoretical approaches have been developed in the literature: mainly, the functionalist approach, the rational choice approach, the institutional approach and the genetic one (Van Haute, forthcoming). We have considered the institutional approach, which postulates that the type of membership defines party organization, as the most appropriate in order to develop our analysis (Michels, 1966; Neumann, 1956; Kirchheimer, 1966; Panebianco, 1988). Two main hypotheses will be researched: the first one aims at verifying whether the different types of party membership allow us to draw a clear differentiation between the degrees of cartelization of intra-party structures. The second hypothesis aims at assessing whether different degrees of cartelization are associated to specific party families.

First, we will study several indicators developed by cartel party scholars for measuring the characteristics of the cartel party model concerning its organisational structure. In particular, we will analyze the share of state resources over party finances. Then, we will examine empirically to what extent a “stratarchy” model of the relationships between grass-roots members and party elites is applicable to the selected cases (Carty, 2004) as well as whether blurred distinctions among members and non-members are equally applicable to the analysis of the selected cases. The role and functions of party membership will be assessed, for example, by studying the party’s organizational reach and membership recruitment rules, the involvement of members in candidate selection procedures, the importance of members’ rights and obligations and the relationships between national and local branches concerning candidate selection.

Even though in some European countries the decline of membership figures has been less evident than in others (Mair and Van Biezen, 2001), the role of membership within party organizations has clearly evolved. We consider that the stratarchical nature of intra-party organization is the most important feature of the cartel party model (at individual party level) because it distinguishes this model from the previous ones and especially from the catch-all party model, developed by Kirchheimer, and the professional-

3

electoral party model, developed by Panebianco. The ascendancy of the party in public office and the ‘professionalization’ of politics have yet been postulated by the catch-all party model and the professional-electoral party model, as well as the capital intensive nature of electoral campaigns and the limited role (‘cheerleading role’ in Kirchheimer’s terms) of the retrenching (in terms of figures) party membership. On the contrary, the hypothesis developed by the cartel party model on the increasing division of labour between party elites, that hold strong decision-making powers, and party members, which might keep a legitimization function (Ignazi, 2004 and 2006; Bardi, Ignazi and Massari, 2008) but are endowed with an increasingly retrenching position, seem to capture the reality of current intra-party organizational features.

The relations between ordinary members and party elites and the character of membership are the sole features of the cartel party that do not consist in a simple intensification of the Kirchheimer’s and Panebianco’s models, but they represent a qualitative shift in party classification theories that allows scholars to grasp and categorize new tendencies and empirical evolutions (Bolleyer, 2008). These two dimensions can thus be analyzed as core properties of the cartel party model, endowed with significant explanatory power and empirical application potential. However, some scholars (Raniolo, 2000), have underlined the tendency towards an increasing organizational isomorphism of contemporary parties due to their ‘electoralization’, as postulated by Kirchheimer. Therefore, they suggest to classify parties mainly on the basis of their patterns of competition and their identity and to focus on inter-organizational dimensions. In this paper, on the contrary, we will not focus on the systemic perspective, but rather on the organizational-individual one. Moreover, as Detterbeck underlines (2005: 179-180 and 2008: 131), stratarchy and ascendancy of party in public office are the main features of the ideal-type cartel parties. The author has integrated in his ten dimensions analysis of party cartelization also the study of the composition of parties’ national executive committees, considering the percentage of national public office-holders in the party leading bodies as a pertinent indicator for measuring the ascendancy of the party in public office. Nevertheless, the role of party in public office constitutes a feature present also in previous party models. We need then to integrate more specific indicators in order to attain valid conclusions on party cartelization.

A large set of empirical indicators of the degree of cartelization of parties and party systems has been described in the literature. On these bases, whether consisting on theoretical developments, critiques or empirical analysis both at systemic or individual level, we may identify four main dimensions for measuring the degree of cartelization of contemporary parties that can be applied at the meso-level, i.e. for analyzing individual parties: 1. The political role. Many scholars have studied the evolution of the linkage function of parties and of their political role, which concerns their position between society and the state (Van Biezen and Kopecky, 2007). In fact, the interpenetration of parties and the state constitutes the main theoretical feature of the cartel party model. The core variables of this dimension are the state regulation of political parties, the rent-seeking practices of parties through patronage and the dependency of parties on public funding. 2. The internal organizational structures. Several scholars have studied the evolutions of the roles and functions of the three main composing elements of parties, the Party in Public Office, the Party in Central Office and the Party on the Ground, in order to assess the ‘ascendancy of the party in public office’ hypothesis developed by the cartel party model (linked to the new representative style implied by the model, i.e. parties as agents of state). The main aspects that can be analyzed concerning this dimension are the elite-members relations and the role of membership. The latter, nevertheless, constitutes a rather understudied variable. 3. The inter-party competition patterns. Research has often focused on typologies of electoral campaigns and on the role of the different composing elements of parties during campaigns, as well as the access to governmental coalitions. The style of inter-party competition and the barriers set up against newcomers are also often analyzed with regard to this dimension. 4. Party goals and ideology. Several scholars have studied the office-seeking nature of the cartel party model and its decreasing ideological distinctiveness. Some authors have also developed other specific aspects of the cartel party model at the individual level, focusing mainly on parties’ policy positions and policy outcomes (Blyth and Katz, 2005; Pelizzo, 2005 and 2007) and on parties’ access to political resources, both in terms of media access and parliamentary recruitment (Raniolo, 2000; Conti, Cotta and Tronconi, 2008). Nevertheless, we consider that the latter variables concerning party access to political resources

4

might be incorporated within the inter-party competition dimension and that the variables concerning policy outcomes can be integrated within the dimension concerning party goals and ideology. Thus, we can conclude that the cartel party model might be apprehended on the basis of four main analytical dimensions. Table 1 summarizes the four dimensions and indicators used in the literature to analyze the cartelization phenomenon.

Table1. The core dimensions of Party Cartelization.

Level/Indicators or dimensions Systemic-relational level Organizational-individual level

Main cartelization indicators 1. Political role indicators 2. Internal organization indicators

Main cartelization indicators 3. Inter-party competition indicators

4. Party goals and ideology indicators

It is important to underline that, not only the first two dimensions are the most pertinent for

assessing the degree of cartelization of contemporary parties at individual level and the most innovative with respect to previous models, but also that the third and fourth dimension are the most easily influenced by external factors, as electoral defeats, entry of new parties, evolutions in the relations between government and opposition, etc. Therefore, we aim at expanding the research outline for applying the cartel party model to individual party level that has been developed by Bolleyer (2007a and 2007b and 2008), thus focusing our analysis on the second dimension, the one concerning intra-organizational cartelization indicators.

Moreover, the main indicator for assessing party cartelization described in the literature is the dependence of party finances over state resources (Haegel, 2008). We will try to assess whether the evaluation implemented on the basis of this indicator corresponds to the evaluation of party cartelization through the analysis of members role within party organizational structure. We would like to assess the application of the cartel party theories to the cases of main West European political parties in order to study its spreading. Pelizzo (2005) has assessed the impact of country factors on party membership decline, exploring the effects of public finance laws, of the degree of competitiveness of the political system and of the electoral system. Therefore we can hypothesise that a similar effect also plays on other aspects of membership beyond its general size, like its power position within party organizational structures. We will develop our analysis on the basis of the following main research questions: 1. To what extent are contemporary parties cartelized in terms of intra-organizational features? 2. Which are the main variables that determine the degree of cartelization of the selected parties?

After having assessed the explanatory power of membership role indicators and the nature of party membership in the selected cases, we will focus on the factors that may affect this variable in the party cartelization perspective. It appears evident that the country factor may play a role in the process of cartelization of current parties: the institutional settings, as said above, may influence the organizational choice of the selected parties. Political culture, social capital, and the characteristics of the political system in general might infringe on the organizational dimension of parties as well. Moreover, even though the cartel party model postulates the convergence of all contemporary parties along stratarchical organizational features disregarding their ideological location, it would be rather interesting to assess whether some party families show a more evident tendency towards cartelization than others. We can thus consider the degree of intra-organizational cartelization measured by the role of party membership as the dependent variable of our study, and, on the other hand, the country factor and the party family as the independent variables. We will also analyze the dependence on state resources and the membership size as control variables. Therefore, we will base our analysis on the following hypotheses: HP1. The role of membership is a relevant indicator for measuring the degree of intra-organisational cartelization of individual parties HP2. The country effect impacts on the degree of intra-organisational cartelization of individual parties HP3. The ideological profile of parties, measured by the belonging to specific party families, impacts on the degree of intra-organisational cartelization of individual parties HP4. The dependence on state funding and the size of the membership are correlated with the degree of intra-organisational cartelization of individual parties

5

After a brief presentation of the variables and indicators that have been selected in order to assess the four hypotheses detailed here, we will give a comprehensive account of the data gathered, the analysis performed and the empirical evidences we have collected.

3. Cases and variables selection

Several analyses developed at the systemic level (presence of cartel of parties) have classified the

European party systems according to their degree of cartelization. On the basis of these classifications, we focus our analysis on two party systems: Italy and Belgium. Both countries are generally considered as two mildly cartelized systems (Delwit, 2008; Conti, Cotta and Tronconi, 2006 and 2008; Bardi, Ignazi and Massari, 2008), which means that the parties of the two selected political systems should have equal environmental chances to undertake a cartelization process (or not). The analysis will thus be carried out under ‘coeteris paribus’ conditions concerning the party system context. The first reason to select these countries is thus that degree of cartelization of the party system may be considered as a constant and that therefore we may focus on individual party degree of cartelization. A second reason to study the cases of Belgium and Italy is the lack of research on these particular cases. Many highly cartelized party systems have been widely studied during recent years (especially France, Germany, Denmark and the Scandinavian countries), but mildly cartelized systems like Belgium and Italy have received far less attention. Finally, this study should be seen as a work in progress, and we aim at achieving a complete overview of the main West European party systems in future research (Austria, France, Germany, the Netherlands, Spain, Switzerland, and the UK).

We selected only parties endowed with national or federal parliamentary representation in 2008. For Belgium we included eleven parties: the green parties Ecolo and Groen!, the social democratic parties PS and SP.A, the Christian democratic parties CD&V and CdH, the liberal party VLD, the ethno-regionalist party N-VA and the populist right parties LDD, FN and VB. The only main Belgian party lacking is MR as this party did not responded to our survey and its statutes were too incomplete to provide the necessary information. For Italy we analyzed eight parties: the social-democrats PD and IdV, the Christian-democrats UDC, the ethno-regionalist parties UV and SVP, and the populist right parties FI and LN, and the conservative party AN.

The relationship between the individual and the party as political organization constitute the focus of our research. On one hand, we will analyze the rules through which parties control their organizational boundaries, by looking at the character of membership, and on the other hand we will observe the rules through which parties both control the relations between members and elites and assure internal party unity. As Bolleyer explains (2007:5), the character of membership is shaped by both duties and privileges and the party’s organizational reach is generally built upon specific rules for recruiting new members. In addition, the most pertinent way to analyze the formal power distribution within the party is to look at the organizational location of the candidate selection procedure (Katz, 2001). As mentioned above, the cartel party model predicts a blurred character of party membership and a stratarchical relationship between members and elites. In particular, we have gathered data on the following dimensions: the quantitative aspect of party membership (membership aggregate figures, membership as a percentage of party’s electorate, i.e. Members/Voters ratio); the share of state resources over party finances1 (share of party resources that are allocated through public funding, whether direct state financing or refunds of electoral expenses); the role of party membership and membership boundaries: importance of rights and obligations of members, relationship between national executive and local branches concerning candidate selection, party’s organizational reach and membership recruitment rules.

The first indicator we have analyzed is the degree of complexity of the recruitment procedure. If, like the cartel party model states, the relationship between members and non members has become increasingly blurred, then it should not be complex to become a party member. We estimated this complexity by exploring the party statutes to see whether there are any formal conditions to enter the party (e.g. not being a member of another party, being introduced by another party member, etc.). We also

1 We would like to analyze also the share of office-holders within national party executive and the

members’ involvement in election campaigning, but data are still being gathered.

6

explored whether it is possible to become a member through an online application, which makes it particularly easy to become a party member. Another indicator to assess complexity is the price of a party card, which can lower the threshold to become a party member (when the price is less than 25 Euro) or raise it (when the price is higher than 25 Euro).

The second indicator is the overall balance of rights over obligations that members can enjoy in political parties. In highly cartelized parties, neither rights nor obligations would be important, but there might also be a slight predominance of rights over obligations. A classical variable to measure this dimension is to investigate how the party leader is elected. The party leader can be elected by means of primaries, trough an election by all the members, or only by delegates. In some cases, the party leader is not elected at all which indicates that members do not possess many rights. Another variable to assess the blurred distinction between members and non-members (voters and supporters) is to analyse whether only members or also non members can be held as candidate on the party list during elections. The third indicator we have developed measures the degree of local autonomy in candidate selection procedures. We explored which organ is entitled to propose a list for the national (or federal) elections, whether local party organs are endowed with veto power and whether the final decision remains in the hands of central bodies. This enables us to assess the degree of stratarchy inside the party. An overview of our main variables is presented in table 2.

Table 2. The variables measuring intra-organizational cartelization.

FEATURES OF MEMBERSHIP

(boundaries and role)

RELATIONS BETWEEN MEMBERS

AND ELITES

Blurred character of membership Stratarchy and members as

individuals

Indicator Indicator

Recruitment procedure Rights and obligations of members Candidate selection procedure

Sub-indicator Sub-indicator Sub-indicator

Conditions for entry* Attending the party congress* Party selectorate definition

Formal procedure Voting right at the party congress* Proposition power of local units*

Member signature required Right to call the party congress Selection of candidates from a list determined by national bodies

Probation phase* Election of party president* Veto power of party headquarters

Level formally deciding upon membership applications Election of the party executive

Veto power of local units over headquarters propositions

Rejections and expulsions must be justified

Right to be held as candidate in elections

Online applications Obligation to follow the party program

Frequency of party recruitment campaigns

Formal procedure for expelling members

Organ launching recruitment campaigns

Non members held as candidates for elections

Inscription fees (reductions)

Therefore, concerning the party’s organizational reach (membership boundaries and membership

role) dimension of analysis, we have selected ten indicators for measuring the complexity of the recruitment procedure and eight indicators for measuring the balance between rights and obligations. With regard to the relations between members and elites, we have selected five indicators for assessing the features of candidate selection procedures. We have performed a principal component analysis on the correlations of the 23 indicators measuring intra-organizational cartelization. Three components were extracted with eigenvalues of more than one and the factors were rotated with both varimax and direct oblimin, giving similar results. The first factor seems to reflect the complexity of the recruitment procedure

7

as all the ten indicators loaded most highly on it, except for the indicator concerning the online membership application. The second factor appears to represent the balance between right and obligations of members. The only indicator that is not highly correlated with the second factor is the one assessing the expulsion procedure. Nevertheless, the indicators that had high loadings on the second factor, also loaded highly on the first factor, suggesting a narrow conceptual distance between the evaluation of the duties and privileges attached to members and the degree of complexity of the recruitment procedure. The two directly oblimin factors are in fact correlated (.24). The third factor seems to represent the relations between members and elites concerning the candidate selection procedure, as all five indicators are highly correlated with it.

For collecting data on these variables, we relied mainly on party statutes, party constitutions and party documents. As some statutes are however quite incomplete and some specific questions could not always be answered through the party statutes, we also relied on a small survey that we sent to the secretary generals of all the parties in our analysis. We have drawn on primary sources and materials, but occasionally we used additionally information obtained from survey data2. Although we rely mainly on qualitative information, we adopted a more quantitative approach by allocating values when certain characteristics were or were not present in the statutes, that way creating quantitative indexes. A quantitative approach on the intra-organizational cartelization has been neglected in most of the cartelization research (with the remarkable exception of Pelizzo, 2005 and 2008), so this study constitutes an explorative attempt at methodological level. It should be noted however, that some information is not always easy to grasp in a quantitative way. The candidate selection procedures, for instance, are often so complex that it is difficult to allocate a certain value to the party. The data in this paper should hence be interpreted and not be taken at face value.

The indicators are measured through values that in each case range between 0 and 1, sometimes classified by four categories (0, 0.3, 0.7, 1), sometimes by three (0, 0.5, 1) and most of the times by two categories (0, 1). The value 0 always corresponds to the minimum degree of cartelization while the value 1 represents the maximum degree of cartelization. We also weighted the most relevant variables by adding an additional point (therefore they range from 0 to 1 to 2). The weighted variables are presented with an asterisk in table 2. The cartelization index is constructed by adding the scores achieved by each party according to the indicators constituting the selected three variables (openness of membership, rights over obligations, and stratarchy). We have thus calculated a ‘membership openness index’ concerning the recruitment procedure, a ‘rights over obligation’ index, and a ‘stratarchy’ index. We have then added the values of the three indexes in order to obtain a ‘cartelization’ index. However, as the openness of membership index is measured through more items than, for instance, the stratarchy index, we weighted the different indexes by dividing the scores by the maximum score that a party could possibly attain on the corresponding index. The average of these weighted indexes (which can be presented in percentages) constitutes a weighted cartelization index. This weighted cartelization index takes each indicator as equally important and hence is more balanced than the cartelization index expressed in absolute numbers.

We have also classified the selected parties through the belonging to specific party families. This type of categorization gives an exhaustive account of the party location on the left-right spectrum. We regrouped the selected parties using six categories: 1) New left and Greens (Groen!, Ecolo); 2) Social-democrats3 (SP.A, PS, PD, IdV); 3) Christian-democrats (CDH, CD&V, UDC); 4) Ethnoregionalist parties4 (N-

2 We analyzed the party statutes and constitutions (20) and other party documents, i.e. the electoral regulations, the

party congress regulations, and the ‘chartes du militant’, etc. (10). The small survey of secretary generals and other central bodies’ officers was conducted between July and October 2008 (at least one respondent per each party). 3 The case of ‘Italy of Values’ is really peculiar, because the party can be defined as a one-issue party, having based its

electoral programmes on justice and anti-corruption issues. It is important to note that the selected cases have witnessed a complete re-foundation of their respective party systems during the last twenty years. The Belgian political system has been regionalized and the Italian one imploded in 1994. 4 The six categories are ranked on the basis of their ideological location of the political spectrum, from the more leftist position (1) to the more right positions (6). The ethno-regionalist parties, which are positioned on the centre-periphery cleavage and thus are located on very different positions on the left-right continuum according to the specific case, are positioned somewhere in the centre of the political spectrum (position 4). This is considered as an ‘average’ position and not an empirically centrist position on the left-right scale.

8

VA, UV, SVP); 5) Liberals (VLD); Conservatives and populist right (VB, LDD, FI, LN, AN). We have not retained for the evaluation of the balance of rights over obligations the specific indicator measuring the obligation to follow the party program, because all parties seem to have integrated this rule in their statutes. Moreover, we have also excluded from the calculation of the final index of recruitment complexity the indicator assessing whether the rejection of a membership candidature must be justified, because the indicator is not correlated at all with the final index. Also the indicators assessing the presence of a formal procedure for expelling members and the obligation to motivate any expulsion have not be retained in the final index of complexity because the explorative factor analysis underlined the distance between these two indicators and the factor representing the final ‘recruitment procedure’ index.

4. The intra-organizational cartelization of Belgian and Italian parties

4.1. The role of party membership in the cartelization perspective Table 3 summarizes our main findings5. The characteristics of party membership that surface from

our explorative analysis are rather varied, but it is possible, however, to draw some general behavioural patterns. Looking at the degree of openness of the recruitment procedure (table 3), we can see that the party endowed with the largest organizational reach is surprisingly enough the Flemish extreme right party Vlaams Belang, even though the other parties belonging to this family seem in general to be less open in terms of membership than other party families. Moreover, there are wide differences between Belgian populist parties (VB, FN, and LDD) and Italian ones (LN, FI and the conservative AN) in terms of openness of the recruitment procedure. It is far more easy to become a member of the FN than of the LN, of the AN or of FI. In the latter cases, the conditions for entry are more extensive: incoming members may not be simultaneously a member in another party or in an organization not approved by the central organs (except in the LN), they must undergo a probation period and are compelled to follow the party program. Generally, they cannot become members through an online application.

The differences in the degree of openness of party boundaries of populist parties reflect the general differences in their organizational structures which range from very weakly to tightly organized structures (Bolleyer, 2007). Even though these parties do not refer to a specific ‘classe gardée’ for recruiting members and voters and are generally new parties seeking to enlarge their membership base for legitimization reasons, the former show very limited, hierarchical organizational structures constructed around the charismatic leader. For example, the VB is one of the less stratarchical parties of our database (1.3 points, table 3). On the other hand, FI has been founded on the rejection of the formal characteristics of political parties such as active membership and has always counted on limited membership figures (5100 in 1994, 190000 in 2006). Obligations are generally more extensive than members’ privileges.

On the other hand, in all the selected right parties it is an intermediate level (regional or provincial) that decides upon applications and expulsions and the rejections or expulsions must be justified. Actually, the only party statute where this rule is not prescribed is the one of the Christian-democrat party CDH. Moreover, the populist and conservative parties’ statutes devote detailed rules for identifying the grounds upon which members can be expelled. In these parties, it is quite easy to become a member, but then it is more difficult to remain so. Internal loyalty is considered generally as fundamental because members’ rights are significantly less extensive than obligations and the internal decision power is tightly held in the hands of party elites. The LN has achieved the lowest score possible in the stratarchy index and the VB and FI are among the parties scoring lower, while the LDD and the FN attained higher levels of organizational stratarchy. Undoubtedly the party age plays here an important role, being AN, LN and FI more institutionalized parties than the newly created LDD.

Except for the two extremes of maximum and minimum openness of party boundaries, witch both belong to extreme right or conservative parties (VB and AN respectively), the other responses to societal heterogeneity are confounded among the other party families. There is no clear ideological pattern in the degree of openness of membership. However, even having generally abandoned the traditional

5 For the cases of LDD and CDH, the data reported on our dataset still have to be cross-checked because the answers

given by respondent do not fully correspond to the provision detailed in the party statutes.

9

organizational models of the mass party, current social-democratic and Christian-democratic parties seem to tend to restrain the boundaries of their membership. Except the SP.A, the social-democrats have scored very low in the membership openness index. Nevertheless, they are all positioned around the median value (5) of the rights over obligation index. In fact, PD, SP.A and PS hold quite a significant membership ratio (6.8, 8.1 and 9.6% respectively) and seem to have maintained hierarchical organizational structures, with the remarkable exception of SP.A which is the most stratarchical party in our database. Social-democratic party members can seldom vote in the party congress nor elect the party executive, even though they often enjoy the right to choose directly the party president. The distinction between members and supporter is rather blurred in comparison to the other party families, considering that both PS and PD reserve specific rules for simple supporters and voters for participating to party’s activities and decisions. Supporters are often invited to participate to the party congress, even though without voting rights.

Moreover, the candidate selection procedure is generally controlled by central organs (except for the PD, where party list are elaborated by an organ elected through party primaries) and the local branches are not endowed with veto powers over the final shape of candidate lists, but for the SP.A. The IdV is a rather peculiar case, with ‘personal party’ features (narrow membership ratio and hierarchical structures organized around the charismatic leader) and very few privileges attached to membership. On the other hand, the Belgian Christian-democrats (CDH, CD&V) seem to be organized more stratarchically (3.3 points on the stratarchy index, table 3) and to grant extensive privileges to their members. Nevertheless, the UDC shows a rather narrow organizational reach (with a very complex recruitment procedure) and a more hierarchical organization. These membership features are undoubtedly linked to the limited electoral size of the party and to its surprisingly high membership ratio (22%), even though the official data on UDC’s membership have often been controversial (Pizzimenti, 2008). In fact, contrary to our expectations, social-democrats seem to represent a weakly cartelized party family with regard to the membership role (with the exception of the SP.A), while Christian-democratic parties scored higher in the cartelization index.

The most cartelized party family is represented by green parties, which show in fact a rather blurred distinction between members and supporters, even though these parties traditionally insist on the important contribution of political participation and allocate to the party on the grounds an important position within the party’s organizational structures. Nevertheless, the party organizational reach is rather wide and members are endowed with extensive rights, while the internal power distribution is rather wide (local branches detain a veto power over candidate lists, but central organs do build the list that members after can endorse). On the contrary, ethnoregionalist parties seem less cartelized: while the NV-A shows more stratarchical structures, the internal organization of the two Italian regionalist parties is rather hierarchical. In all the cases, members’ obligations are more extensive than privileges and party boundaries are narrow. This is clearly linked to the fact that these parties address to specific social segments and are characterized with the highest membership ratio of the database: 10% and 11% for the UV and the NV-A and 37% for the SVP, which is widely known in the literature as the regionalist party with the strongest militant base (Tarchi, 1998 and Pallaver, 2006). The liberal party integrated in our database meets perfectly the defining features of the cartel party model: highly stratarchical in organizational terms (candidate selection powers are shared between local branches and central organs), the party shows rather open membership boundaries (the recruitment procedure is way less complex that in other cases).

10

Table3. Membership role’s indicators and party cartelization integrated index.

Party

Membership Openness

index

Rights over obligations

index

Stratarchy index

Cartelization index

Weighted Cartelization

index (%)

SP.A 11 5 5,7 21,7 74,6

Groen! 7 7 4,3 18,3 65,4

Open VLD 6 9 3,3 18,3 65,0

CDH 6,7 7 3,3 17 59,2

CD&V 8 6 3,3 17,3 58,3

Ecolo 7 8 2,3 17,3 58,0

FN 8 5 3,3 16,3 54,6

LDD 10,7 3 3,3 17 53,2

N-VA 6,7 4 3,3 14 48,0

PD 5,5 5 3 13,5 47,4

PS 6,7 5 2,3 14 46,2

VB 12 3 1,3 16,3 45,0

UV 4,8 5 2 11,8 40,3

IdV 6,5 4 1 11,5 34,8

SVP 5,8 3 1,3 10,1 31,2

LN 6,5 3 1 10,5 31,1

UDC 3,2 3 2 8,2 29,3

FI 4,5 2 2 8,5 28,5

AN 0,8 2 3,3 6,1 27,5 4.2 The different degrees of party cartelization and the effects of the independent variables

The data in table 3 reveal a clear distinction between the two countries according to the scores of the weighted index of cartelization. Belgian parties, and particularly the Flemish ones, appear to be the most cartelized. The main exception seems to be the VB, a member of the populist radical right party family, which in general seems to reject the cartel party model (Bolleyer, 2007). The French-speaking parties are mostly situated in the middle of the table and hence can be classified as mildly cartelized parties. Finally, at the bottom of the table, we find the great bulk of Italian parties. An intuitive analysis of table 3, moreover, suggests the presence of a clear ideological pattern in the distribution of the cartelization index’s scores. Considering each country case separately, Social-democrats and leftist parties seem to be more cartelized than Christian-democrats and liberals. Populist parties are less cartelized within the three selected party systems and Italian conservative and populist parties (FI and AN) have achieved the absolute lowest scores. Only in the case of French-speaking parties, the social-democrats show a lower score than that of the populist radical right. The regionalist parties are positioned in general in the middle of the table and show rather a high variation in their scores. The country and the party family variables seem to play an important role in the definition of the degree of cartelization of the selected parties.

With regard to the complexity of the recruitment procedure index, measuring the organizational reach of the party on the ground and thus the degree of openness of party membership, we see a large difference between Italian and Belgian parties. While, for instance, it is a common rule for Belgian parties to provide online applications for membership, this is not always the case in Italy. The populist radical right party Vlaams Belang is the party most open to new members in our analysis. The statutes of this party mention that the only condition to become a member is to pay the annual fee and to endorse the statues of the party. It is not detailed, for example, whether the potential VB member may also be a member of

11

another party. Individual members can even be recruited anonymously through a postal recruitment procedure, mainly in order to avoid the social stigmatization that is generally reserved to extreme right parties’ members in contemporary societies. The VB’s statutes show significant differences with the rules for joining the Italian right parties, which require very complex conditions to be satisfied in order to become a member, at least among those registered in our survey. The central organs of Alleanza Nazionale, for instance, investigate attentively each admission application through three different levels of evaluation and require not only that potential members do not belong to other parties or to social movements and associations not approved by the party, but also that they prove to behave following the principles of ‘honour, dignity and personal decorum’ and do not show any ‘anti-national behaviour’.

More or less the same country differences can be seen with regard to the balance between rights and obligations of party members. Overall, the members of Italian parties enjoy fewer rights than the party members in Belgium and in the latter case the differences between members and non-members are generally more unclear. In particular, the members of the VLD and the green parties in Belgium (both Flemish and French-speaking ones) are endowed with wider channels of representation and greater possibilities to make themselves heard. The VLD congresses not only integrates all the party members but also the observers and supporters of the party; the party president is elected by all the members, and the latter have also the possibility to elect the executive organ of the party (Jagers, 2002). The opposite features in terms of balance between rights and obligations can be seen in the case of populist right parties. Given the fact that Forza Italia is generally considered as a ‘personal party’ according to the literature (Hopkin and Paolucci, 1999), it is not surprising that the party president cannot be elected by its members. Moreover, only delegates can participate to party congress and be endowed with voting rights. The recruitment procedure is quite simple, also because the consequences of joining the party remain minors, because members’ rights are quite limited and the differences between members and supporters remain to a certain extent blurred.

With regard to the autonomy of local party branches and of lower territorial strata, we see a clear difference between Belgium and Italy as well. In Belgium, both in the case of Flemish and of French-speaking parties, the great majority of parties allow the provincial branches to propose candidate lists for the chamber of representatives. The central organs can then approve the final lists or intervene in case of disputes. In Italy on the contrary, it is generally the Party in Central Office that build the lists that afterwards can be approved by party members within the local branches.6 Moreover, the socialist Flemish party SP.A shows the most stratarchically organizational structure. The party central organs are only endowed with advisory powers with regard to candidate selection procedures. The candidate lists for the Chamber are built by the provincial (i.e. intermediate level) organs and they are finally approved by the party members in the SP.A congress. Italia dei Valori and Lega Nord, on the other hand, show a strongly hierarchical organization, in which the local branches can only propose a limited portion of the candidate places available in each constituency list. On the contrary, Forza Italia shows a hierarchical internal organization, with the decision-making power being held tightly in the hands of the central organs. The degree of vertical stratarchy of different party levels is also exceptionally low in the case of Alleanza Nazionale. The candidate selection procedure is dominated by the central organs and the lower strata do enjoy a very limited autonomy in this field.

There is no clear cut explanation for the difference in the degree of intra-organizational cartelization between Italy and Belgium. Tentatively, we might interpret this difference as a consequence of the differences in the political cultures of the two countries. However, the distance in terms of political culture is not overwhelming enough to explain the major difference in the degree of intra-organization cartelization. Size might be a stronger explanatory variable, as the much larger membership of the Italian parties might restrain them to adopt more cartelized organizational structures. However, the party

6 The main explanation for this wide country difference can be found in the electoral system variations between the two countries. In Belgium, the party system is completely regionalized, as we have seen, and the electoral system for the federal lower chamber is based on separated provincial constituencies. Thus, in Flanders can be presented only lists of Flemish parties and, on the other hand, only French-speaking parties can compete in the provincial constituencies of Wallonia (a mixed system is present in the bilingual Brussels region), while in Italy there are 26 regional constituencies with a completely nationalized party system.

12

cartelization literature is rather controversial on this point (Detterbeck, 2005) and the size variable should be researched more in-depth on the base of more country cases.

It must be noted however that Italy is a quite unusual case because the party system has undergone some serious changes since the collapse of the First Republic in 1993-4. The only parties in our sample that keep strong organizational and ideological linkages with the First republic political system are the PD, heir of the communist party and then of the PDS-DS (even tough constituting a ‘new party’ in organizational terms); UDC, one of the heirs of the Democrazia Cristiana, and the regionalist parties UV and SVP. As Italian parties were, in most cases, adhering intensively to the mass party organizational model, a significant path dependency effect seems to intervene here and to lead these parties (PD, UDC, UV and SVP) towards less cartelized organizational features in recent years. The UDC for instance, shows complex procedures for recruiting membership, limited rights allocated to incoming members and a hierarchically structured organization. In many cases, these parties are composed by highly polarized factions and this means that the threat of internal conflicts or even splits might lead party elites to be more reluctant to enlarge membership boundaries or give members more extensive rights. The new parties that emerged after the collapse of the First Republic are, except for IdV, all populist or conservative right parties which generally reject the cartel party model. Parties like LN, FI and AN are typically endowed with a charismatic leader and personalized internal organization. As populist parties see themselves as the counter-reaction against the formation of an elite cartel on the party system level, it is not surprising that these parties tend to reject the cartel party organizational model.

In Belgium, most parties evolved quite fast from a mass party organizational model towards more cartelized structures. The more recently created parties are Groen!, Ecolo, FN, LDD and VB. Green parties seem to widely adopt the cartel party model, whilst populist right parties appear more hostile, but not completely opposed, to intra-organizational cartelization. Particularly with regard to the membership recruitment procedures, Belgian populist right parties seem more open to new members, quite unlike their Italian counterparts. This fact might be due to the tendency of the formers to search for new members in order to gain social and political legitimization. This tendency seems to be less relevant in the Italian case.

Table4. Intra-organizational cartelization and party families.

New Left

and Greens

Social

democrats

Christian

democrats

Ethno-

regionalist

Liberals

Conservative-

Populist right

Total

Very low cartelization 0 1 (25%) 1 (33%) 1 (33%) 0 3 (50%) 6 (33%)

Low to intermediate cartelization 0 2 (50%) 0 2 (67%) 0 1 (17%) 5 (28%)

Intermediate to high cartelization 1 (50%) 0 2 (66%) 0 0 2 (33%) 4 (22%) Very high cartelization 1 (50%) 1 (25%) 0 0 1 (100%) 0 3 (17%)

Total 2 4 3 3 1 6 18

100% 100% 100% 100% 100% 100% 100% Table 4 explores the relationship between party family and the (weighted) degree of membership

cartelization. The parties in our analysis were divided into six party families. As explained throughout our paper, the number of cases in our analysis is very limited and this makes it difficult to draw valid conclusions on the relationships between the cartelization degree and the ideological variable. Ecolo and Groen! are the only two green parties in our sample. With regard to membership features, both of these parties can be labelled as highly cartelized. This result is consistent with the findings of Bolleyer (2007, 22) who claims that the new left parties “seem to reflect cartel party features even clearer than old parties moving from the mass party over the catch all to the cartel model”.

13

Four parties that belong to the social democratic party family were analyzed: SP.A, PS, PD and IdV, although the latter has a quite diffuse ideological profile and is difficult to classify within the traditional party families. Only the SP.A is highly cartelized in intra-organizational terms, but this is not the case for the other Italian and French-speaking social democratic parties. The PS, for instance, is quite hierarchically structured and is quite balanced with regard to rights and obligations of members, even though the difference between members and supporters is slightly blurred. It shows a rather complex recruitment procedure, at least according to Belgian standards. Moreover, the PS not entirely relies on state resources, and this fact contrasts with the situation of many other Belgian parties. When we analyze the Christian-democratic parties (CD&V, CdH and UDC) with regard to intra-organizational cartelization, the picture is rather unclear. While the Italian UDC is among the least cartelized parties in our sample, the Belgian Christian democrats are much more cartelized. With regard to liberal parties it is impossible to draw any conclusions as we have only one case in our database. Although the liberal VLD is extremely cartelized in organizational terms, we should explore more cases before concluding that this is a dominant pattern within the liberal party family.

An even more interesting case is the ethno-regionalist family, which seems to generally reject the cartel party model. The N-VA, UV and SVP are all positioned from very low to quite low positions on the intra-organizational cartelization index. Age might be an interesting intervening variable to consider here. While ethno-regionalist parties are generally considered as new parties by the literature (Deschouwer, 2007), in Belgium and Italy the parties belonging to this party family can be found among the oldest parties in their respective party systems, both in ideological and organizational terms (although the N-VA went through some sound ideological and organizational transformations since the implosion of the Volksunie in 2001). This might explain why they are still a bit more reluctant towards the cartelization process than other parties. Obviously a rather evident exception is represented by the Lega Nord, which is characterized by a very low degree of cartelization, but which also is generally considered as a new party. We have however chosen to include the latter party among the populist right parties instead of the ethno-regionalist ones as this classification seemed more appropriate with regard to the purpose of our research, even though the classification is rather controversial in the literature (Tarchi, 1998; Gomez-Reino 2002).

Finally, we have analyzed the case of the populist right parties (VB, LDD, FN, FI, LN, AN). The great majority of these parties have achieved very low scores according to the intra-organizational cartelization index. Charismatic and powerful leaders hierarchically guide a collective of followers in most of these parties. Nonetheless two populist right parties in Belgium (LDD and FN) also feature some characteristics of intra-organizational cartelization: the LDD shows a rather stratarchical organizational structure, while the FN seems to have a broad reach of the party on the ground due to a low membership boundaries control and a balance between rights and duties biased towards membership privileges.

Table 5. Bivariate correlations.

Intra-

organizational

Cartelization

Public funding

over total receipts

Membership

ratio

Intra-organizational Cartelization Correlation 0,050 -0,367

Sig. (2-tailed) 0,853 0,135

N 19 16 18

Public funding over total receipts Correlation 0,050 -0,057

Sig. (2-tailed) 0,853 0,839

N 16 17 15

Membership ratio Correlation -0,367 -0,057

Sig. (2-tailed) 0,135 0,839

N 18 15 18

14

In table 5 we explore the relationship between intra-organizational cartelization and two other indicators that are frequently used to assess individual party cartelization. The first is the percentage of public funding over total receipts, which might indicate to what extent parties are ‘migrated’ to the state and are no more dependent upon other resources (e.g. membership contributions ) for their functioning7. The second new indicator introduced here is the membership ratio or the number of party members divided by the voters of the party. What appears evident from the analysis of this table is that there are no strong and no significant correlations between the selected variables. In part this might be due to the limited number of cases selected, which makes it more difficult to find significant relations in statistical terms. This said, it is striking that the intra-organizational cartelization has no correlation whatsoever with the percentage of public funding. It seems that these two aspects of cartelization are less related than often assumed. The intra-organizational cartelization is however negatively correlated with the membership ratio, which supports our initial hypothesis. Parties with relatively many members seem less cartelized. Party size in terms of membership seems then to play some effects on the degree on intra-organizational cartelization, even though (as mentioned above) the cartel party literature is quite controversial on this point. Although this relationship is not significant, this might be due to the fact that the analysis is based on only 19 cases.

There are different aspects and dimensions of cartelization, as showed in table 1. At the individual level, the analysis can be focused on the ideological and party goals dimension, or on the inter-party competition dimension. The latter is applicable mainly at systemic level, but can be translated also at the meso-level. The analysis of individual party cartelization can be focused also on the political role dimension or on the intra-organisational dimension. All these dimensions correspond to different types of cartelization. We have focused mainly on the intra-organizational dimension because it appears to be the most innovative and powerful in explanatory terms in comparison to previous models. Nevertheless, the nature of competition constitutes the most direct measure for the collusion features of inter-party cooperation postulated by the cartel party model. The analysis of this dimension at the individual level may provide very different results in comparison to those obtained by analyzing the intra-organisational features. This is the reason why cartel party literature normally focuses on the systemic level, in order to assess the integrated weight of each of the mentioned dimensions on the total degree of cartelization of a given party system. This level of analysis allows us to balance the different results obtained by analyzing each of the four cartelization dimensions.

On the other hand, if we want to focus on the individual party level, it is more difficult to weight and integrate all the four dimensions of party cartelization. We can obtain very different results for the same party according to the dimension chosen for the analysis. We can therefore conclude that there are different types of cartelization and that the choice of the dimension of the analysis must be carried out under detailed theoretical justifications in order to attain valid causation relations. For example, Conti, Cotta and Tronconi (2006 and 2008) show that Italian parties, taken together at the level of party system analysis, can be considered as highly cartelized with regards to public funding rules, ideological distinctiveness and party competition, especially with regard to media access, while they are far less cartelized in organizational terms.

Given the results of our analysis and the wide differences in the degree of cartelization measured through the analysis of party finances and in the one measured by analyzing the role of membership, we may conclude that, at least on the base of our narrow case selection, there are different types of cartelization. Moreover, we may also conclude that the party finance dimension do not directly concern intra-organizational aspects, as part of the cartel party literature suggests, but rather the political role of parties and the interpenetration of parties and the state (Detterbeck, 2005). The resource cartelization index and the membership cartelization index concerns thus two different dimensions of analysis and can therefore lead to very different conclusions. It would be interesting, then, to assess the influence that one

7 The financial data for Italian parties are collected on the basis of the annual party budgets published on the ‘Gazzetta

ufficiale’ and of Bardi, Ignazi and Massari, 2008. The data concerning Belgian Parties are collected on the basis of Delwit, 2008; Jo Noppe “Morphologie des partis francophones en 2002 et 2003”, Res Publica, 46 (2-3), 2004, pp. 413-452 and Jo Noppe “Morfologie van de Vlaamse politieke partijen in 2003 en 2004”, Res Publica, 47 (2-3), 2005 pp. 349-426. The data concerning membership figures are found through the same sources.

15

dimension can exert upon the other, rather than assessing whether they concur in measuring the same phenomena. Given the limited number of cases of our research sample, it would be difficult at the moment to test the relationship between the resource cartelization index and the membership cartelization index. Nevertheless, we may explore the nature of the relationship between the two main independent variables of our study, the country factor and the party family, and the main dependent variable, the degree of intra-organizational cartelization8.

Table 6. An exploratory linear regression model.

Unstandardized

Coefficients

Colonne1

Standardized

Coefficients

T

Sig.

B Std. Error Beta

(Constant) 0,865 0,056 15,376 0,000

Country -0,214 0,034 -0,762 -6,238 0,000

Party family -0,023 0,009 -0,299 -2,449 0,026

Dependent Variable: cartelization (R²=0,773)

In this regression, both the country effect and party family effect are estimated upon the intra-organizational cartelization. The regression performed seem to confirm the conclusions we have drawn on the base of the cross-tabulations reported in table 4 and on the rank-ordered data reported in table 3. Given the low number of cases integrated in the analysis, we cannot push the explanatory power of this regression model further as supplying a clear assessment of the conclusions intuitively built upon descriptive statistics. The country effect impacts inversely on the degree of intra-organizational cartelization: when the country scores increase (passing from 1, Flanders, to 2, Wallonia, to 3, Italy)9, the degree of cartelization decreases. Given the high R-square (.773), this confirms the intuitive interpretation of table 3, where Italian parties seemed to score less on the integrated cartelization index than Belgian ones, especially Flemish parties. Moreover, also the inverse relation between party family identification and intra-organizational cartelization appears to be confirmed here. Moving from the left (1) to right (6) of the political spectrum the degree of intra-organizational cartelization of parties decreases. As we have seen in table 4, new left and greens parties are highly cartelized, while conservative and populist right parties are generally very low cartelized.

5. Conclusions

Since the seminal article of Katz and Mair published thirteen years ago, in which the authors describe the emergence of the cartel party model, many empirical researches have been devoted to the assessment of the degree of cartelization of European party systems. In this study, however, we focused on the intra-organizational cartelization at individual party level. We explored whether a blurred character of membership and a stratarchical relationship between members and elites can be found in current parties in the selected party systems. We based our analysis on these two dimensions because they constitute the only variables that really transcend the intensification of the previous catch-all model. The first aim of this paper was to develop an analytical instrument which enabled us to assess the degree of intra-organizational cartelization in a quantitative way by means of an examination of party statutes and primary documents as well as of a small survey among the parliamentary parties in Italy and Belgium.

This quantitative analysis provided interesting information and conformed that the role of membership represents a relevant indicator for measuring the degree of intra-organisational cartelization

8The N of the sample is obviously too small to draw significant statistical conclusions and in this case it would be more pertinent to apply the qualitative comparative analysis techniques (QCA, see Rihoux and Ragin, 2007 and De Meur and Rihoux, 2002). However, this is just an exploratory attempt to assess our funding hypotheses and these analyses would be further developed methodologically in the next stages of our research. 9 The two countries have been actually coded as three different regions, given the fact that the Belgian party system is completely regionalized: Flanders, including all the Flemish parties (1), Wallonia, including all the French-speaking parties (2, these parties are active also in the Brussels region) and Italy (3).

16

of individual parties. In fact, a large variation has been found in the degrees of party cartelization between the two countries as well as between different parties within the same country. In general, we saw that Italian parties were far less cartelized than Belgian ones and Francophone parties even less than the Flemish ones. These findings confirm our second hypothesis on the relationship between the country factor and the degree of intra-organizational cartelization. Although we provided some tentative explanations for this difference, ranging from path dependence, to party internal factionalism and to the striking predominance of new populist right parties in Italy, these represent only provisional and ad hoc explanations. In order to investigate which variables really exert a significant effect on the degree of intra-party cartelization, more data should be gathered and different cases must be integrated in the analysis.

The third hypothesis that was assessed in this paper tried to identify which party family has adopted the cartel party model more extensively. In this respect, we found that the green parties tend to adhere most to this organizational form, in line with previous research findings (Bolleyer, 2007). The retrenchment of members’ obligations and the blurring of the supporter-member divide in new parties of the left closely correspond to the expectations formulated by the cartel party hypothesis. Populist right parties generally rejected the model, although the relationship is not so clear cut. The Belgian populist right parties for instance seem to be more open to new members than could be expected. This could be due to the fact that these parties are desperate to legitimize themselves through assembling a considerable number of members, an issue that populist right parties in Italy have less to deal with. The ethno-regionalist party family also seemed to reject the cartel party model in most of the cases.

For the Christian-democrats and the social-democrats, no clear cutting conclusions can be drawn, even though the latter seem at a first glance more cartelized than the former: the average score of intra-organizational cartelization achieved by Belgian and Italian social-democratic parties is slightly higher (16.4) than the one realized by Christian-democrats (14.2). Nevertheless, the country factor and the ideological location of parties seem to exert a significant effect on the degree of intra-organizational cartelization and these hypotheses have been confirmed also by performing a regression analysis. These findings are particularly relevant as they contradict the current belief that describes the organizational cartelization process as irrespective of ideological positions.

Finally, we explored whether the intra-organizational cartelization is correlated with other dimension of cartelization such as dependency of state resources and the relative number of party-members. Interestingly enough, we found no significant correlations whatsoever between these variables. In part, this may be due to the fact that we only have analyzed a limited number of cases, but it cannot be ignored that the parties in our analysis sometimes clearly rejected the intra-organizational cartelization, while at the same time they highly depend upon state resources. For example, Alleanza Nazionale and Forza Italia achieved the lowest scores according to the intra-organizational cartelization index, but are intensively funded through state resources. On the other hand, the Partito Democratico and the Parti Socialiste scored highly on the cartelization index, but the percentage of public funding over their total receipt is among the lowest in the database. Therefore, we might conclude that the role of membership constitutes a separate dimension of intra-organizational party cartelization and must be examined with due attention.

In conclusion, it is possible to assert that, in general, party members are endowed with more extensive roles at the level of local branches and with features that raise only a blurred distinction with party supporters and voters in Belgian parties than in Italian ones. The former thus appear to be rather cartelized. Italian parties’ organizational structures tend to be more hierarchical, and thus less cartelized, although with some remarkable exceptions. The fact that our sample is partially biased because of an overrepresentation of populist and far right parties might concur to the explanation of these findings, as populist parties tend to assume more hierarchical organizational features and clear distinction between members and non-members. Nonetheless, stratarchical organizational features are present in most cases of green and new left parties, of Christian-democratic parties and also of social-democratic ones.

It is thus possible to assert that the model of intra-organizational cartelization is generally verified with regard to the openness of membership boundaries and to the stratarchical features of party organization, but less with regard to the blurred distinction between members and non-members. Even though this characteristic is present to a certain extent in some of the selected parties (PS, PD and partially SP.A), the great majority of our cases presented a rather clear separation in the role of members from the role of simple supporters and voters, either with an emphasis on members’ rights over obligations (in the

17

case of Christian-democratic and green parties) or with the charge of extensive duties over incoming members (in the case of ethnoregionalist and populist parties). Even though the overall balance of rights over obligations can change according to the case, it cannot be generally defined as non important as has been implied by the cartel party model. With regard to this variable, the selected cases seem to correspond more precisely to the catch-all party model, which predicted an emphasis of rights over obligations even in the presence of a rather heterogeneous membership. Party members remain an important composing element of the intra-organizational features of current parties, although membership boundaries tend to enlarge and membership figures tend, on the contrary, increasingly to retrench.

6. Bibliographical references

AUCANTE, Y. and DEZE, A. (2008) (eds.). Les systèmes de partis dans les démocraties occidentales. Le modèle

du parti-cartel en question. Paris: presses de Sciences Po. BARDI, L. (2006) (ed). Partiti e sistemi di partito, Bologna: Il Mulino. BARDI, L.; IGNAZI, P.; MASSARI, O. (2008). I Partiti Italiani. Iscritti, dirigenti, eletti. Milano, Università Bocconi

Editore. BEYME, K. von (1996). Party leadership and change in party system: Towards a post-modern party state?,

“Government and Opposition”, vol.31, no.2, pp.135-159. van BIEZEN, I. (2003a) Financing Political Parties and Election Campaigns. Strasbourg: Council of Europe. van BIEZEN, I. (2003b) Political Parties in New Democracies: Party Organization in Southern and East-Central

Europe. Basingstoke: Palgrave Macmillan. BOLLEYER, N. (2007). New parties: Reflexion or Rejection of the Cartel party Model?, paper presented at the

ECPR General Conference in Pisa, 6-8 September 2007. BOLLEYER, N. (2008). Inside the Cartel Party: Party Organisation in Government and Opposition, “Political

Studies”, vol.56, no.3, pp. BLYTH, M.; KATZ, R. (2005) ‘From Catch-all Politics to Cartelisation: The Political Economy of the Cartel

Party’, West European Politics, n° 28, vol. 1, p. 33-60. CALISE M. (2000). Il partito personale. Roma, Laterza. CARTY, R. K. (2004). Parties as Franchise Systems. The Stratarchical Organizational Imperative', “Party

Politics”, vol.10, no.1, pp. 5–24. CONTI, N; COTTA, M. and TRONCONI, F. (2006). The cartelization of the Italian Party System: One Step

Forward and One Step Backwards, CIRCAP Occasional Papers, no. 17/2006. CONTI, N; COTTA, M. and TRONCONI, F. (2008). Le parti-cartel en Italie. In AUCANTE, Y. and DEZE, A. (2008)

(eds.). Les systèmes de partis dans les démocraties occidentales. Le modèle du parti-cartel en question. Paris: presses de Sciences Po, pp. 195-219.

CROSS W., YOUNG L. (2004). The Contours of Political Party Membership in Canada, “Party Politics”, vol.10, no. 4, pp.427-444.

DALTON, R.J.; WATTENBERG, M.P. (2000). Parties without partisans: political change in advanced industrial democracies. Oxford: Oxford University Press.

DELWIT, P. (2003). Composition, décomposition et recomposition du paysage politique en Belgique, Bruxelles: Labor.

DELWIT, P. (2008). Partis et Système de Partis En Belgique. Une Double Cartellisation À L’œuvre ? In AUCANTE, Y. and DEZE, A. (2008). Les systèmes de partis dans les démocraties occidentales. Le modèle du parti-cartel en question. Paris: presses de Sciences Po, pp. 219-244. DESCHOUWER, K. (2007). New Parties in Government. London: Routledge DESCHOUWER K., DE WINTER L., DELLA PORTA D. (1996), ‘Particracies between crises and reforms: the cases

of Italy and Belgium’, Res Publica, vol. 48, no. 2. DETTERBECK, k. (2005). Cartel parties in Western Europe?, «Party Politics», vol.11, no.2, pp. 173-191. DETTERBECK, K. (2008). Le cartel des partis et les partis cartellisés en Allemagne. In AUCANTE, Y. and DEZE, A. (2008). Les systèmes de partis dans les démocraties occidentales. Le modèle du parti-cartel en question. Paris: presses de Sciences Po, pp. 129-152. DUVERGER, M. (1954). Political parties. Their organization and activity in the modern state, London: Methuen.

18

FIERS, S; VANLANGENAKKER, I.; INGLESE, C. (2007). Who wants to/the party? The meaning of party membership in an eroding particracy. Paper presented at the ECPR Joint Sessions of Workshops, Helsinki 2007, 7-12 May.

GOMEZ-REINO, M. (2002). Ethnicity and Nationalism in Italian Politics. Inventing the Padania: Lega Nord and the Northern Question, London: Ashgate.

HAEGEL, F. (2008). Le « Parti-Cartel ». De la Logique Interne à la Validité Empirique. In AUCANTE, Y. and DEZE, A. (2008). Les systèmes de partis dans les démocraties occidentales. Le modèle du parti-cartel en question. Paris: presses de Sciences Po, pp. 113-127. HEIDAR, K. (2007). What would be nice to know about party members in European Democracies? Paper

presented at the ECPR Joint Sessions of Workshops, Helsinki 2007, 7-12 May. HEIDAR, K. (2006). Party membership and participation, in Richard S. Katz and William Crotty (eds.),

Handbook of Party Politics. London, Sage. HOPKIN, J., PAOLUCCi, C. (1999). "The Business Firm Model of Party Organisation: Cases from Spain and

Italy", European Journal of Political Research, 35(3), 307-339. HOPKIN, J. (2004) 'The Problem with Party Finance, Theoretical Perspectives on the Funding of Political

Parties', Party Politics, vol. 10 , no. 6, pp. 627–51. JAGERS, J. (2002). "Eigen Democratie Eerst! Een Comparatief Onderzoek naar het Intern Democratische

Gehalte van de Vlaamse Politieke Partijen", Res Publica, 44(1), 73-96. IGNAZI, P. (2004). Il puzzle dei partiti: più forti e più aperti ma meno attraenti e meno legittimi, Rivista

Italiana di Scienza Politica, no.3, pp.325-346. IGNAZI, P. (2006). Gli iscritti ad Alleanza Nazionale: attivi ma frustrati, Polis, vol.20, no.1, pp. 31-58. IGNAZI, P., FARRELL, D. M. and ROMMELE, A. (2005) 'The Prevalence of Linkage by Reward in Contemporary

Parties', in A. Römmele, D. M. Farrell and P. Ignazi (eds), Political Parties and Political Systems, the Concept of Linkage Revisited. Westport CT: Praeger, pp. 17–36.

KATZ, R. (2001). The Problem of Candidate Selection and Models of Party Democracy, “Party Politics”, Vol. 7, no. 3, pp. 277-296.

KATZ, R.; MAIR, P. (1992). Introduction. The Cross-National Study of Party Organizations. In R. Katz and P. Mair (eds.), Party Organizations, A Data Handbook. London: Sage, pp. 1-21.

KATZ, R.; MAIR, P. (1994). How parties Organize. London: Sage. KATZ, R.; MAIR, P. (1995). Changing models of party organization and party democracy. The Emergence of

the cartel party. Party Politics, vol. 1, no. 1, 17-21. KATZ, R.; MAIR, P. (1996). Cadre, Catch-all or cartel? A rejoinder, “Party Politics”, Vol. 2, No. 4, pp. 525-534. KATZ, R.; MAIR, P. (2002). The Ascendancy of the Party in Public Office. The Party Organizational Change in

Twentieth-Century Democracies. In R. Gunther et al. (eds.), Political Parties. Old Concepts and New Challenges. Oxford: Oxford University Press, pp. 113-135.

KIRCHHEIMER, O. (1966). The transformation of the western European party system, in LaPalombara, J., Weiner, M. (eds), Political parties and political development, Princeton University Press, Princeton NJ, pp.177-200.

KITSCHELT, H. (2000). Citizens, Politicians, and party cartellization: Political representation and the State failure in Post Industrial Democracies, “European Journal of Political research”, vol.37, no.2, pp. 149-179.

KOOLE, R. (1996). Cadre, catch-all or cartel? A comment on the notion of cartel party, Party Politics, no.4, vol.2, pp.507-524.

KROUWEL, A. (2006). Party Models. In R. S. Katz & W. J. Crotty (eds.), Handbook of Party Politics, London: Sage Publications, pp. 249-269.

Lawson, K. and MERKL, P. (1988). When Parties Fail: Emerging Alternative Organizations. Princeton: Princeton University Press.

LINZ, J.; MONTERO, J.R. and GUNTHER, R. (2002) (eds.). The Future of Political Parties. Oxford: Oxford University Press.

LUTHER K.R. (1999), ‘A framework for the comparative analysis of political parties and party systems in consociational democracy, in Luther K.R. and Deschouwer K. (eds) Party Elites in Divided Societies. Political Parties in Consociational Democracy, London, Routledge, pp.3-19.

19

LUTHER, K., MULLER-ROMMEL, F. (2002), Political parties in the new Europe: Political and analytical challenges, Oxford University Press, Oxford.

MAIR, P. (1997). Party System Change. Oxford University Press, Oxford. MAIR, P. (2005). Democracy beyond parties, Center for the study of democracy, Paper 05-06.

http://repositories.cdlib.org/csd/05-06. MAIR, P., VAN BIEZEN, I. (2001). Party memberships in twenty European democracies, 1980-2000, “Party

Politics”, no.1, pp.5-22. MICHELS, R. (1966).La sociologia del partito politico. Bologna, Il Mulino. MORLINO, L.; TARCHI, M. (2006). Partiti e caso Italiano. Bologna, Il Mulino. NORRIS, P. (2002), Democratic Phoenix: Political Activism Worldwide. Cambridge: Cambridge University

Press). NEUMANN, S. (1956). Modern Political Parties. Approaches to Comparative Politics. Chicago: University of Chicago Press. OSTROGORSKI, M. (1903). La democratie et l'organisation des partis politiques. Paris: Calmann-Levy. PANEBIANCO, A. (1988). Political Parties. Organization and Power. Cambridge: Cambridge University Press. PAOLUCCI, C. (2006). The Nature of Forza Italia and the Italian Transition. Journal of Southern Europe and

the Balkans. Vol. 8, no. 2, pp. 163-178. PELIZZO, R. (2005). The Changing Political Economy of party membership, Quaderni di Scienza Politica, vol.5,

no.2, pp. 211-238. PELIZZO, R. (2007). A subjective Approach to the Study of Oligopolistic Party Systems, Quaderni di Scienza

Politica, vol. 1, no. 3, pp. 393-419. PELIZZO, R. (2008). The Cartel party and the Italian Case, Politics and Policy, vol.36, no. 3, pp. 474-498. RANIOLO, F. (2000). Miti e realtà del cartel party. Le trasformazione dei partiti alla fine del ventesimo secolo,

“Rivista Italiana di Scienza Politica”, vol. 20, no.3, pp. 553-581. PILET, J-B. ; VAN HAUTE, E. (2006). Morphologie des partis politiques francophones en 2004-2005. Res

Publica, vol. 48, n°2-3, 2006, pp. 297-335. SARTORI, G. (2005). Party types, organisation and functions, “West European Politics”, no.1, pp. 5-32. SCHATTSCHNEIDER, E. E. (1942). Party Government. New York: Farrar and Rinehart. SCARROW, S.E. (2000). Parties without members? Party organizations in a changing electoral environment,

in Dalton, R.J., Wattenberg, M.P. (eds), Parties without partisans: Political change in advanced industrial democracies, Oxford University Press, Oxford, pp. 79-101.

SEILER, D-L. (1993). Les Partis Politiques, Paris: Armand Colin. SEILER, D_L. (2008). Théorie des partis et systèmes partisans. In AUCANTE, Y. and DEZE, A. (2008). Les systèmes de partis dans les démocraties occidentales. Le modèle du parti-cartel en question. Paris: presses de Sciences Po, pp. 89-112. SEYD, P.; WHITELEY, P. (eds.) (2004). Special Issue: Party Members and Activists, Party Politics, vol.10, no. 4. SEYD, P.; WHITELEY, P. (2002). New Labour’s Grassroots. Basingstoke: Palgrave. SEYD, P.; WHITELEY, P. (2004). ‘British Party Members: An Overview’, Party Politics, vol. 10, no 4, pp. 355-

366. TARCHI, M. (1998). Italy: the Northern League, in L. de Winter e H. Türsan (eds.), Regional Parties in Western

Europe. London: Routledge, pp. 143-157. VAN HAUTE, E. (2007), Does ideology matter? Evidence from the Belgian case, paper presented at the ECPR

Joint Sessions of Workshops, Helsinki 2007, 7-12 May. VAN HAUTE, E. (forthcoming, 2009). Pourquoi s’engager? , Bruxelles: Editions de l’Université de Bruxelles. WHITELEY, P.; SEYD, P.; (1998). The Dynamics of Party Activism in Britain – A Spiral of Demobilisation?,

“British Journal of Political Science”, no. 8, pp. 113-138. WOLINETZ, S. (2002). Beyond the Catch-all Party: Approaches to the Study of Parties and Party Organization

in Contemporary Democracies. In Juan Linz, Jose Ramon Montero, and Richard Gunther (eds.), The Future of Political Parties. Oxford: Oxford University Press, pp. 136-65