Intra-party democracy in Belgium: On paper, in practice and through the eyes of the members

Transcription

1 Intra-party democracy in Belgium: On paper, in practice and through the eyes of the members Bram Wauters 1 University College Ghent and Ghent University bram.wauters@hogent.be Paper prepared for the ECPR Joint Sessions Lisbon April 2009 Workshop 21: Promoting Internal Party Democracy: A Selling Point, A Serious Danger, Or A Redundant Exercise? 1

2 1. Introduction Political parties are thought of as being in crisis. This is not entirely true, though. Mair (1994) proposed to split up the party organisation into three parts in order to facilitate analysis: the party in public office, being the party in parliament and in government, the party on the ground, i.e. the party members and also the party voters, and the party in central office, being the central party organization. Rather than talking about a general decline of parties (Wattenberg, 1984 ; Webb, 1995) it is better to call recent evolutions a change within parties where the balance between the three components has been altered to the detriment of the party on the ground. The two other components have grown in importance due to the generous system of public financing and to the intra-party professionalisation. This implies that when one is speaking of a decline of parties, one refers mainly to the party on the ground, featuring a decline of trust in parties, a reduced identification with parties and a decline in the number of party members. The decrease of the number of party members is a phenomenon that occurred the last decades all over Europe (Mair & Van Biezen, 2001). This decrease is also apparent in Belgium. In 1980, all Belgian parties together counted members, while in 1999 this had dropped to , which is a decrease of 24 % (Biondi et al., 2000). Also in trust rankings, Belgian political parties always lag far behind other actors and institutions such as Parliament, the press and the judicial system (Elchardus & Smits, 2002) Most political parties continue, however, to make every effort to attract as many members as possible. This can be explained by the benefits that members offer to parties, that still outnumber the costs (Scarrow, 1994). One of the means to attract members is giving them a voice in the internal functioning of the party. By giving the rank and file a greater say, parties have tried to increase the involvement of the members with the party and hoped in this way to stick them to the party. Owing to secularization and de-pillarization, the traditional reasons to join a party have lost much of their power. Moreover, the last decades new social movements and new parties such as the green parties, who value involvement and participation highly, have gone into competition with traditional organizations and parties (Deschouwer, 2

3 1994). These new movements succeeded in attracting a lot of people, especially young people, mainly at the expense of more classic organizations (Katz, 1990). These latter movements tried to anticipate this trend by providing more participatory procedures in their internal functioning (Scarrow, 1999). In addition, the rise of postmaterialistic values in most Western societies has had as a consequence that participation and involvement have become important issues. Inglehart (1990) has found a clear shift in the culture and attitudes of Western societies from materialistic to postmaterialistic values, focussing on the quality of life and self-expression. This shift coincided with an improvement in education and the extension of communication means and mass media. Thus, people have not only the desire to participate, but also the knowledge and skills to become involved. Procedures granting party members more involvement include among others party leadership elections (LeDuc, 2001 ; Denham & O Hara, 2007 ; Wauters, 2009), giving members the right to vote on party conferences (Dewachter, 2003) and involvement in the composition of the candidate lists for elections (Pennings & Hazan, 2001 ; Katz, 2001 ; Rahat & Hazan, 2001). The question arises whether this introduction of participatory procedures constitutes a real empowering of the rank and file. Most studies about intraparty participation demonstrate only a limited influence of the rank and file and even see these procedures as instruments to reinforce the power of the party elite (Mair, 1994 ; Scarrow, 1999 ; Wauters, 2003). In this paper, it is our aim to give an overview of intra-party democracy in Belgium by bringing together existing data and presenting new material about this topic. Belgium has witnessed the rapid introduction of intra-party democratic procedures in the course of the 1990s, following the rise of the green parties and the radical transformation of the Flemish liberal party (De Winter, 2000). Four participation procedures will be scrutinized in ths paper: party leadership elections, member participation in party conferences, party member involvement in the composition of candidate lists and intra-party referenda. For each of these procedures, three types of analyses will be run: an analysis of the changes in the party constitution (formal rules) and the motives used to undertake these changes, an analysis of the functioning in 3

4 practice of these intra-party procedures and an analysis of how the members perceive these procedures. Our approach is summarized in this figure: Party leadership elections Right to vote in party conferences Compostition of candidate lists Intra-party referenda Formal party rules Practice View of members 2. Change in formal party rules 2.1 An overview of the changes a) party leadership elections The party leader is a crucial and powerful actor in Belgian politics (Fiers, 1998) who is responsible for both the internal organisation of the party and the external relations and activities, such as acting as spokesperson in negotiations and appointing ministers. Formerly, Belgian political parties used to assign their party leader by a decision of the party executive or the party conference containing only a selected elite of party members. In the 1970s only the French-speaking 2 christian-democratic party PSC elected their party leader in internal elections. The members of the francophone nationalist party FDF elected their leader in that decade on a party conference where all members could participate in. At the end of the 1980s, the French-speaking liberal party PRL followed the example of the PSC by organising leadership elections. In the beginning of the 1990s, the radically transformed liberal party VLD introduced as first party in Flanders internal elections to designate the party leader (De Winter, 2000). Soon, most other parties followed and by now, almost all Belgian parties use internal elections with member suffrage to assign their party leader 4

5 There are three different means to collect members votes in leadership elections (Maes, 1990 ; Fiers, 1998 ; Biondi et al., 2000) : - by asking a member to vote by post or via electronic way (internet, ) - by arranging polling booths in every local party section - by inviting all members to attend a party conference where they can vote We include only elections using the first two means in our overview, since attending a conference clearly involves much more effort for the rank and file (Rahat & Hazan, 2001). In this way, internal elections in a.o. FDF and the green parties are excluded from the analysis. (for an overview of all party presidential elections until 2004: see Appendix). While in the 1970s only 8 % of the party leadership appointments were decided by internal elections, this figure had risen to 38 % in the 1990s and it seems very likely that it will have further increased in the first decade of the new millennium. Table 1: Number of party leader appointments and number of party leadership elections 3 in Belgian political parties ( ) (Maes, 1990 ; Biondi et al., 2000 ; Wauters, 2003) Decades Number of party leader appointments Number of party leadership elections Percentage party leadership elections All parties 4 represented in Parliament give their members nowadays a direct say in the appointment of the party leader, be it on a party conference be it via internal elections. The exception is formed by the Flemish extreme right party Vlaams Belang where the party council holds a vote and the preferred candidate is subsequently acclamated by a party conference. We should note, however, that parties limit the number of competitors by (to a varying degree) imposing conditions on potential candidates. 5

6 Table 2: Means of assigning the party leader in Belgian political parties 5 (current situation) Party Party leader assigned by Conditions for candidates CD&V All members - Nominated by either at least 1 provincial section, 2 regional section or 3 local sections SP.A All members - Nominated by at least 10 local sections from at least 2 provinces, who together count at least 5000 members - Candidate deputy party leader of the other sex OpenVLD All members - Member of the party since at least 2 years - Nominated by at least 2 regional sections from different provinces or by at least 1 local section in every province or by 500 members Groen! Party conference with No conditions participation of all members N-VA All members No conditions S-LP All members or Party No conditions conference with participation of all members 6 LDD All members 7 No conditions Vlaams Party council - Member of the party council Belang CDH All members - Member of the party since at least 3 years PS All members - Member of the party since at least 5 years - Member of a party committee on the arrondissemental level MR All members - Member of the General Council Ecolo Party conference with - Member of the party since at least 6 months participation of all - One member of the Brussels region and one of members the Walloon region 8 The Flemish liberal party VLD introduced as first and only party up to now also internal elections for the national party executive. b) right to vote on party conferences The party conference is, according to the formal party rules, the supreme party institution (Dewachter, 2003). Although their role in intra-party decision making has recently been challenged, they continue to play a crucial role in the life of political parties (Faucher- King & Treille, 2003). 6

7 As for the party conferences in Belgium, there has been in the 1990s an extension of the number and type of members that has the right to intervene and to vote on these conferences. The traditional system of delegation has recently been replaced by a system where every individual member is entitled to attend party conferences and to express a vote. An analysis of the formal party rules 9 in the 1980s shows that all Flemish parties except the green parties Agalev and Ecolo used a system of delegates for their party conferences. It should be noted, however, that there still were large differences between the systems of delegation used at that time. In the social democratic party SP, for instance, there was only one delegate per 300 members and this delegate was assigned on the arrondissemental level, which groups several local party sections. In the christiandemocratic CVP on the contrary, there was already one delegate for 50 members and they were designated at the local level. It is straightforward that the participation threshold for grass-roots members is higher when there are more members needed to have a delegate. Table 3: Formal rules for participation to party conferences (1980s) (Dewachter et al, 1988) Party Who is allowed to participate? (both the right to Assigned by intervene and to vote) CVP 1 delegate per 50 members Local sections SP 1 delegate per 300 members Arrondissemental Federations PVV 1 delegate per 1250 votes at the parliamentary elections and Arrondissemental 1 delegate per 150 members Federations Agalev All members VU 2 delegates per section and 1 additional delegate per 25 Local sections members PSC 1 delegate per 500 members Arrondissemental Federations PS 1 delegate per 250 members Arrondissemental Federations PRL 1 delegate per 500 votes at the parliamentary elections and 1 Arrondissemental delegate per 50 members Federations FDF delegation according to the number of members (not further Local sections specified) Ecolo All members 7

8 The radically reformed liberal party VLD enfranhised as first major traditional party all their members on party conferences in Again, soon most other parties followed their example. Table 4: Formal rules for participation to party conferences (current situation) Party Who is allowed to participate? (both the right to Assigned by intervene and to vote) CD&V All members SP.A 1 delegate for each section with a minimum of 50 Local sections members and one additional delegate per 200 members OpenVLD All members Groen! All members N-VA All members S-LP All members LDD All members CDH All members PS 1 delegate per 250 members Arrondissemental federations MR System of delegation of the components of the party 10 Party components Ecolo All members Nowadays, almost all parties accord their ordinary members the right to intervene and to vote on party conferences. When parties still use a delegation system (SP.A for example), participation hurdles are lowered in comparison with the past. Non-members are in general allowed (and increasingly invited) to attend party conferences, but not to intervene nor to vote. c) intra-party referenda The last decades, referenda are worldwide increasingly used (Butler & Ranney, 1995). A referendum can be defined as a process that allows the electorate to approve or reject a policy proposal (Matsusaka, 2004). An initiative process even allows citizens to propose items to be put forward to the population, by collecting a predetermined number of signatures from their fellow citizens. In a party referendum, the party elite presents one or more propositions to the rank and file who approve or disapprove these propositions. These propositions could be policy 8

9 propositions to be included in the party s electoral manifesto or modifications to (intra-) party procedures or more fundamental changes to the party s core elements. An intra-party referendum should be discerned from internal opinion polls about general (VU in 1983 e.g.) or very specific topics (VLD in 1994 on judicial topics e.g.). In a referendum, approval is asked for clear propositions presented to the members, while in internal opinion polls their opinion is asked without any further obligations. A second condition for an intra-party referendum is that all members, and only they, are entitled to participate. The VLD organised in 1995 for instance a referendum where all Flemish voters could participate in. This is not an intra-party referendum. The Flemish green party Groen! is the only party who has included rules about a referendum in its party constitution. 11 If a two-third majority of the participants of a party conference or one fifth of all members or one third of the local party sections demand for a referendum, the party is obligated to organise one. This referendum have to be about political relevant topics and should contain clear and straightforward propostions. The results are binding when a two third majority agrees with a propostion. In other parties, intra-party referenda can be organised as well, but on an ad hoc basis. Given the high cost of an intra-party referendum, to date only 3 such referendums have been organised in Belgium: one by the Flemish social democrats of SP in 1993 and two by the French-speaking christian-democrats PSC in 1996 and in In each case, it was always the party elite that took the initiative to organise a referendum. In none of these referenda, there was bottom-up involvement of the members from an early stage, as is the case in a popular initiative. Table 5: Intra-party referenda in Belgium Date Party Subject Number of participants Participation rate November 1993 SP 20 topics of the party % manifesto June 1998 PSC renewal of the party % April 2002 PSC / cdh new name and new party concept % 9

10 d) composition of candidate lists Despite the increasing percentage of preferential votes and a recent electoral reform halving the impact of list votes on the allocation of seats, the list order, and consequenly the list composition, remains to a large extent decisive in determining who will be elected in Belgian Parliament (Wauters & Weekers, 2008). The composition of the list and the choice of the list order is important in deciding who will be elected. Contrary to the other procedures granting members involvement in the party, Belgium has a long tradition of member participation in the composition of the candidate lists. Already in the 19 th century, the composition of candidate lists of Belgian parties was discussed and decided by assemblies open to all members (so-called polls ). In the 1960s and 1970s, however, these polls did not attract a large number of members anymore or were even no longer organized. Whereas in 1958 still 90 % of the members of the House of Representatives were placed upon a candidate list by a poll, this has declined in 1978 to only 10 % (De Winter, 1988). The composition of the candidate lists was then mostly decided by assemblies of delegates or by a selected party institution, and no longer by all members. This has, however, changed again and Belgian parties, together with a large number of other European parties (Pennings & Hazan, 2001), have re-introduced member involvement in the candidate selection process at the end of the 1980s and in the course of the 1990s. We should note, however, that formal membership involvement in the composition of candidate lists is less prevalent than in leadership appointements for instance. We restrict our analysis here to the composition of the candidate lists for the elections for the Senate and the European Parliament. These elections are organised in region-wide electoral districts and consequently the candidate selection process also takes place at the level of a region, i.e. Flanders and Wallonia. The limitation of our research object allows us to control for local differences. It appears from the statutory rules that the influence of the rank and file in this candidate selection process is considerable. Table 6: Formal rules concerning the composition of the candidate lists for the European elections and the elections of the Belgian Senate 10

11 Party Who decides? On what? CD&V All members in a secret poll Model list SP.A Party conference (delegates) Model list OpenVLD All members in primaries Individual candidates Groen! Party conference (all members) Individual candidates or model list N-VA Party council Model list SLP Party council Individual candidates or model list LDD Party executive - Vlaams Belang Party council Model list CDH Party conference (all members) Model list PS Party conference (delegates) Model list MR Electoral commission - Ecolo Party conference (all members) Rahat & Hazan (2001) make a distinction between appointment systems and voting systems. In a pure voting system, all candidates are selected by a voting procedure and no other party organ can intervene in the result. In an appointment system, on the contrary, candidates are appointed by a selected party organ without any further need of approval. Most of the Belgian parties can be located in between these two extremes: the party elite designs a model list, which must be approved by the party members (ranging from all members to those at a party conference or a party council). 2.2 Reasons for these changes Now the reasons why parties introduced these changes will be described. We categorize them into four groups. For this analysis, we rely on newspaper articles and party publications about intra-party democracy. a) attractiveness for (new) members Granting members a say in the policy of the party is seen as a strategy to make party membership more attractive (Scarrow, 1999). The underlying hypothesis is that all (future) members are keen on participation - which is by the way not the case (Wauters, 2009). Parties use their democratic functioning as an argument to convince people to join the party. Parties that were amongst the first to introduce internal party democracy 11

12 stressed the unique characteristic of it. In 1993 when the VLD held for the first time leadership elections, Bart Somers (now party leader, but at that time only a junior party official) wrote in the party member magazine: This weekend we organise something completely new for Belgian politics: internal elections in which every member is allowed to participate ( ). In no other party in Western Europe, members have a greater say than in our party! (Somers, 1993:1) In the early years of the new millenium, both the Flemish social-democrats (SP.A) and the Flemish liberals (VLD) set up an advertisement campaign stressing their internally democratic profile in order to attract new members. Now, several parties mention on their website the involvement of the grass roots members in determining the party s policy as one of the motivations to join the party. Some illustrations: Enough of waiting, I want action! I ll join the S-LP and I ll have the right to vote so that I can determine how our party project will further be built up (S-LP, 2009). On the sp.a-website, four reasons to become party member are summed up. The fourth one is: because my opinion counts within the party (SP.A, 2009) On the OpenVLD website, the party gives four motives to join the party. One of them is An unequalled internal democracy: VLD members are to large extent involved in the party s internal functioning. The VLD was the first party that allowed their members to elect directly the party leader and the party executive. The rank and file have the final say (and recently also the first say) on the compostion of the candidate lists. We welcome our members with open arms on party conferences ( ) to decide actively on the general party line and on governmental participation. (...) (OpenVLD, 2009) b) reinforcing the bottom-up communication Parties are increasingly loosing their link with society (Mair, 1994). In order to keep abreast of the problems and needs of their followers and of society as a whole, parties rely on intra-democratic procedures. The opinion of the grass roots is used to test whether the party s policy is in line with the public opinion. The SP party leader at that time Frank Vandenbroucke formulated the motivation of his party to organise an internal referendum in 1993 as follows: it was decided that, in order to get tuned in with the rank and file 12

13 and to become a modern political member movement, we had to adapt our internal procedures drastically. It was our aim to become more involved in politics together with our members. We succeeded. Party s viewpoints that were formerly approved by some hundred party activists at a party conference, are now read, weighed, discussed and deliberated by more than 40,000 members. (SP, 1993: p. 85) Internal democratic procedures are instruments allowing the party elite to keep track of evolutions in society. c) attractiveness for voters Parties like to spotlight their internal democratic character also to the voters. A leadership contest for instance can be a spectacular opportunity to advertise itself to the electorate as an open and democratic party (Punnett, 1992). The internal democratic functioning is clearly used as an element of electoral competition between parties. In 1993 when the VLD introduced party leadership elections, the party s spokesman Guy Vanhengel wrote: Last month, Frank Vandenbroucke was elected party leader of the SP. He obtained 294 of the 394 expressed votes. Only 0,27 per cent of the SP members were entitled to vote. We, as party of the citizen, on the contrary, invite all the 80,000 VLD members to express their vote." (Vanhengel, 1993: 1). The preceived success at that time of the VLD internal elections, and of this reformed Flemish liberal party in general, incited other Flemish parties to introduce also leadership elections. As a consequence of this contagian effect, most parties changed their formal rules in order to give members a greater say in amongst other the choice of the party leader. In the communication of the results of the internal democratic procedures, parties always stress over-enthusiastically how well these procedures functioned. After the SP referendum, the party described it as a dialogue without predecessor and an immense succes. The PSC called their 2002 referendum a brilliant proof of the members interest in their party. This not a particular Belgian phenomenon: because parties are convinced that voters are in favour of these new procedures, the party elite pays lip service to these reforms and make efforts to publicize them for a wide audience (Scarrow, 1994). 13

14 The question arises whether the introduction of internal democratic procedures produces really any effect in electoral terms. A condition for such an electoral effect is the awareness of the existence of these intra-party democratic procedures among voters. As the internal functioning of parties often remains hidden for non-members, this condition seems not to be met. This is confirmed by a survey of the KU Leuven 13 conducted among the entire Flemish population. Respondents were asked to indicate for four organisations whether the members could vote in internal elections for either their leader or their national council. Two political parties were included in the analysis: one that organises such internal elections (VLD) and one that does not organise such elections, but operates in a hierarchical manner (Vlaams Blok) (Deschouwer, 2001). The questionnaire contained also two organisations that are organising internal elections but that are not political parties (the semi-public catholic health insurance company CM and the disciplinary organisation for doctors Order of Medecins).. Table 7: Could you indicate for each of the following organisations whether the rank and file could elect directly via internal elections their leader and/or their council? (Student survey KU Leuven) Organisation Election No election Don t know N % N % N % VLD Vlaams Blok CM Order of Medecines No less than 78 % of the respondents correctly states that VLD members can elect their party leader in internal elections. This is significantly higher than the CM and the Order of Medecins who organise also internal elections. Although Vlaams Blok functions in a hierarchical way without internal elections, about two-third of the respondents are convinced that their members can elect directly either the party leader or the party council. From these results, it appears that citizens do know that political parties organise internal elections (but not that other organisations are organizing them as well). Citizens do, 14

15 however, not make a clear distinction between parties that do organise such elections and parties that do not. This begs the question how parties can benefit in electoral terms from their intra-party democracy if citizens hardly make a distinction between the internally democratic parties and more hierarchical functioning parties. d) legitimisation of the party leaders A party leader devoting himself to a renewal of the party, but fearing resistance from long-established party figures or party groups can search backing from the rank and file in order to realise his ambitions. This support can be made public by using intra-party democratic procedures. The PSC Referendum of 1998 for instance was held in a period when the party was internally divided and was performing badly in the opinion polls. In order to be able to renew the party with a powerful hand, the candidate party leader Philippe Maystadt asked the grass roots-members to express their opinion about him as a party leader in party leadership elections, but also about his ideas of party renovation which were bundled in the Contract de relance (Recovery Contract). He took himself the initiative to organise an intra-party referendum about this Contract in order to settle down possible opposition against it. Another example is provided by Elio Di Rupo, who declared after his election as PS party leader in the first party leadership elections in that party in 1999: I will use the mandate you gave me to further renew the party. I would like to see more women in our party, we will work more open than in the past, and so on. (Di Rupo, 25/10/1999) Opposition from people or intra-party groups who have something to loose with party renovation can be overruled by referring to the mandate one was awarded by the rank and file, either by being elected party president either by the endorsement of a particular program of renewal. Sometimes, the party elite introduces internal democratic procedures especially for these purposes. These procedures accord legitimation to the party elite, which can be used to further their goals more powerfully. 15

16 3. In practice 3.1 Observations Formal rules are not always a good indicator of the functioning in practice of participatory procedures. More than half a century ago, Lipset, Trow & Coleman (1956: 1) already wrote about member participation in trade unions: In few areas of political life is the discrepancy between the formal juridical guarantees of democratic procedure and the actual practice of oligarchic rule so marked as in private and voluntary organisations. This also applies to political parties. Consequently, it is important to consider not only the formal rulesn but also the practice of member involvement in parties. It might be that the party elite is strengthening its influence by limiting in practice the extensive influence on paper of the rank and file. As Mair (1994:17) states: democratization on paper may actually coexist with powerful elite influence in practice. It seems that members are often only giving a formal stamp of approval to decisions made elsewhere. a) party leadership elections The democratic character of internal elections is measured here by three indicators: the number of (serious) candidates, the difference between the contenders and the percentages of votes casted. Firstly, up to 2007 there have been 41 party leadership elections (see also Appendix) in Belgium. In 20 of them (48.8 % of all leadership elections), there was only one candidate. This means that in almost half of all internal elections grass roots members can only approve or disapprove the candidate put forward by the party elite. It is evident that their impact can only be limited then. Secondly, when there are more candidates, there often is no real contest between the candidates. Most of the time there is a candidate preferred by the party establishment who only has to struggle with a minor, often not well-known contender who obtains only a limited share of votes. In elections with more than one candidate, the winning candidate has on average 36.8 % more votes than the runner up. Up to now, only in two internal 16

17 elections the difference was smaller than 10 %: in 1996 when Charles-Ferdinand Nothomb was elected president of the PSC against Joëlle Milquet (a difference of only 0,01 %) and in 2000 when Geert Bourgeois beated Patrik Vankrunkelsven in the VU leadership contest with a difference of 7,8 %. Thirdly, the average participation rate in Belgian party leadership elections is only 40,09 % (expressed on the number of party members). In general, less than half of the party members participates in internal elections. A further statistical analysis of all Belgian party leadership elections reveals some factors that influence this participation behaviour. These results show that there is a (negative) correlation between the percentage of votes for the winner and the participation rate in internal elections. The higher the score of the winner, the lower the participation rate. There is also a significant correlation between the difference between the winner and the runner-up, and the participation rate (when there is more than one candidate). Table 8: Correlation coefficients between participation rate and the score of the winner, and between participation rate and the difference with the runner-up in Belgian intra-party elections Correlation between participation rate and N Correlation coefficient Percentage votes obtained by the winner ** Difference with the runner-up * * p < 0.1 ** p < 0.05 *** p < 0.01 Also the mean participation rates differ considerably between elections with only one candidate (average participation rate of 36.92) and elections with more than one candidate (average participation rate of 43.10). We can conclude by stating that voters will participate more to internal elections when there is more than one candidate, when the score of the winning candidate is not extremely high and when there is a runner-up with a considerable share of support. In most Belgian party leadership elections, however, these elements are absent. b) member participation in party conferences 17

18 Although in most parties all party members are enfranchised on party conferences, their real influence seems not to be increased. There are three indications for. First of all, party conferences tend to be more open and more visible than in the past. Given the extensive media coverage of these events, parties make every effort to show up well. As a result, party conferences are to a large extent directed by the party elite who ensures that the party is depicted in a positive manner. Consequently, profound and heated discussions are avoided. The testimony of former MP Herman De Loor (sp.a) speaks volumes: Formerly, we used to hold in-depth discussions on party conferences. Now, they put on a show, where it is decided beforehand who can intervene and how." (De Loor, 2002) Secondly, an analysis has shown that the party elite has an influence over 13 different stages of the party conference which enables them to take control of the conference (Dewachter, 2003). These stages range from deciding whether to organise a party conference or not, over a.o. choosing the topic to be discussed, designating the conference leadership and establishing the voting procedures, to delivering an extensive speech themselves. Despite the apparent increase in power of the rank and file, the party elite retains much power about party conferences in its own hands, even when members are granted the right to vote on these conferences. A third remark that should be made about party conferences is that they in practice almost never take binding decisions on important issues. A conference sets the party ideology and the general party line, but does not impact on the way this general line is day-to-day put in practice (Faucher-King & Treille, 2003). As long as the party in central office determines the daily functioning of the party, the final effect of members involvement about for instance party ideology remains limited. c) intra-party referenda There are two weaknesses in the functioning in practice of intra-party referenda in Belgian political parties (apart from the sparse use of them): almost all propositions are approved with a large majority and the party elite remains autonomous in interpreting the results. 18

19 First of all, the high approval rates of the propositions strike the eye. In the SP member referendum, all 20 propositions were endorsed, and the lowest score for a proposition was 65 %. In the 1998 PSC referendum, all 10 propositions except one were supported by the members and yielded results ranging from 75 to 97 %. Only one proposition ( the party leader can intervene in the candidate selection process in order to give non-members a place on the list ) obtained only 46,9 % of the votes and was consequently disapproved. These high acceptance rates can be explained by two factors: Firstly, parties select only non-controversial issues to be put to the members. In the review report of the results of the SP referendum, it was said that three types of propostions were included in the referendum: propositions where the position of the party was already longtime known, positions that are already some time defended by the party elite but that were not yet approved by a party conference, and new propositions. Including only non-controversial issues limites the impact of the rank and file, and allows the party elite to prove that they are tuned in with the grass roots: party leader at that time Vandenbroucke concludes: The fact that these 20 propositions are approved by a large majority, proves that these are in line with the common sense of social and democratic people (SP, 1993: 85) A second factor that can explain the high acceptance rates is the inclusion of one-sided arguments in the propositions. In the SP referendum there was an issue on compulsory voting, which was formulated as follows: At elections all people must go to the polls. Compulsory voting guarantees that the voice of as many people as possible is heard. Without such an obligation, particularly people with social problems will abstain. This is proven by the practice in the US. (SP, 1993: 66) This could be a textbook example of how the phrasing of a questionnaire could influence the results. It is difficult to be against compulsory voting with this partial information in favour of it in the phrasing of the proposition. In the 2002 PSC referendum about a change of the party name, the following question was included: Do you think that, in order to have a chance to strengthen the party and in order to propagate our ideas in the population and in particular among the young generation and in order to have the capacity to defend our values in practice, the PSC should find a new acronym that continues nevertheless to identify the party with its 19

20 particular political project? The arguments in favour of a change of name are mentioned, while possible drawbacks are not included in the question. Moreover, the crucial question, the new acronym itself and whether there needs to be a c of christian(- democratic) 15 in it, was not put to the members. Secondly, the party elite keeps the hands free for the interpretation and consequences attached to the results of the referendum. Since intra-party referenda are almost never included in the party s formal rules, the party elite can set autonomously the rules on an ad hoc basis. In the 2002 PSC referendum about a change of name, only 47 % of the participants of the intra-party referendum agreed that the party should adopt a new acronym. The party changed its name despite having only obtained a relative majority in the referendum. The party elite justified this decision by referring to the fact that only 42 % of the members 16 did not agree with a change of name and that 60 % of the elected representatives of the party did agree. From this analysis, it appears that in intra-party referenda the consent with the proposals of the party elite is considerable. Intra-party referenda seem to look for confirmation with the views of the party elite rather than granting the rank and file a real say in the determination of the party program or the positions of the party. d) compostion of the lists Unlike Anglo-Saxon countries, there is no legal obligation to put candidate lists to the members or the voters for approval. Consequently, parties in Western Europe have more leeway in determining the rules of the candidate selection process and as a consequence a greater chance to control this process (Pennings & Hazan, 2001). Although most parties have granted their rank and file according to the statutory rules more power in the composition of the candidate lists, the party elite remains the most powerful actor in this process. The recent comments of incumbent Flemish MP Anissa Temsamani (SP.A), who was granted an inferior position on the model list, confirm this view: [The provincial party leader] ( ) advises me to dispute the model list on a party conference, but I know beforehand that I will bite the dust then. Everyone knows that the 20

21 decision on candidate lists by a party conference is entirely a matter of form. (Temsamani, 24 February 2009). Most of the time, the central party office daws up a proposition for the whole list or for the most crucial positions on that list. The party members can most of the time only approve or disapprove this model list designed by the party leadership. The latitude of members to disapprove is further limited by the press coverage of this model list. The party elite already communicates about the list composition before the model list is endorsed by the rank and file. When party members reject a model list, this constitutes a public disapproval of the party elite and exhibits a detrimental image of a divided party in the build-up to the elections. Parties themselves make every effort to avoid intra-party struggles becoming public. For the upcoming regional elections of June 2009, the liberal party OpenVLD decided in the East-Flanders district to reach an agreement with the three contenders for the top position of the list (Senator Jean-Jacques De Gucht, former Flemish minister Fientje Moerman and former MP Filip Antheunis), rather than allowing the rank and file to express their preference as is required by the formal party rules. A model list is nearly always approved by the members. An exception is formed by the green party Groen! were the composition of the list is publicly announced only after the poll results and where model lists (if used) sometimes are rejected. In 1999 the model list for the European Parliament was disapproved and all members could express their vote for the two contenders for the top position, Isabelle Vertriest and Patsy Sörensen. It also happens in other parties sometimes that model lists are eventually altered, but it is apparent that the rank and file do not play any role in this process. 17 A final remark about the low impact of party members in the candidate selection process refers to the number of participants. Participation rates of internal democratic procedures in this selection process are generally low. For the primaries for the Senate in 2003 for instance only 14 % of the VLD party members casted their vote (Lucardie & Voerman, 2004). These figures are clearly lower than those for party leadership elections (see above) and provide an indication of the limited impact of party members in this process. 21

22 3.2 Explanations Three explanations for the limited involvement in practice can be formulated. a) Individualisation of involvement A remarkable feature of the changes in the statutory party rules is the strong focus on the individual party member, rather than for instance on local sections or on intra-party groups. In their classic study about internal democracy in unions, Lipset, Trow and Coleman (1956) have named the existence of organised sub-groups within an organization the most crucial factor in producing real internal democracy. Such sub-groups of members permit organisations, contrary to what Michels stated in his iron law, to function democratically, Lipset et al claim. This also applies to political parties. Intra-party democracy tend to flourish in parties where members make up organised and structured intra-party groups, while maintaining at the same time loyalty towards the party. Faction has long had a negative connotation because it has been perceived as representing divisiveness (stemming from power battles between personalities rather than between ideas). It has, however, also been recognized that factions have positive functions (Beller & Belloni, 1978). These intra-party groups are playing in this perspective a role comparable with that of parties in society: bundling claims, raising grievances, recruiting candidates, mobilizing members to vote, communicating about the points at stake, and so on (Wauters, 2004). Consequently, parties that are approaching the individual party member and that are neglecting groups of party members (local sections, factions, etc.) tend to be detrimental for the members impact in practice. b) law of curvilinear disparity We will add here to the above reasoning about individualisation that parties do it intentionally. Individual party members are more docile and tend to follow the party leaders more easily than committed party groups or party activists (Mair, 1994). We can refer in this respect to the law of curvilinear disparity (May, 1973; Kitschelt, 1989; Mulé, 2001). According to this law, the views of party activists will always be more radical than those of the party elite and those of the ordinary members. The leading 22

23 politicians of a party are dependent upon the voters for their re-election, therefore they are going to moderate their points of view in order to enlarge the electorate. Meanwhile, party activists have no mandate to lose, so there is no need for them to pursue a more pragmatic policy. Moreover, they are not involved in any daily or regular negotiations with other parties. As a consequence, rank-and-file members are often ideological hardliners who may have problems with the more pragmatic attitude of salaried politicians. Although contested (Weldon, 2007), this law provides us with useful insights in the motivations of the party elite. By granting the ordinary party members a say in the party s policy on paper, the party elite circumvents these party activists (often backed by a local section or a faction) and keeps much of the power in their own hands. This consultation of the (docile) rank and file-members yields the party elite legitimation and the image of an open and democratic party, while avoiding the drawbacks of participation which would impinge negatively on their own power and on the party s efficiency. For instance for the candidate selection process which requires a balancing of a.o. ideological, geographical, age and gender factors, it is much more efficient for the central party office to have full control rather than empowering the different opinions of a broad range of party activists. In sum, parties grant participation opportunities deliberately to individual members instead of to delegates of local sections or factions. This allows them to enjoy the advantages (legitimation, image of open party) while avoiding the drawbacks (inefficient internal party functioning, too much influence for extremist party activists) of intra-party participation. c) pillarized parties in a consociational democracy A final reason for the lack of impact of the rank and file is particular for Belgium. Belgium is generally known as a textbook example of consociational democracy (Lijphart, 1969). This concept refers to the cooperative government of party elites in order to establish and maintain a stable democracy in divided societies. In such a system, the leading elite of pillarized parties (i.e. parties as part of structured network of organisations) is supposed to diminsh the society-threatening tensions by negotiating with party elites from other pillarized parties. Elite control, a great amount of freedom for the 23

24 elites and apathy of the masses are essential characteristics of such a system (Huyse, 1971 ; Deschouwer, 2006). Moreover, state reforms that turned Belgium into a federal state introduced a complex institutional framework with a.o. mutual vetoes, that obligates party elites to continue to cooperate with each other. Political parties, and in particular a limited number of top party figures, play a major role in the negotiations out of political crises. As apathy of the masses and a free hand for party leaders are key features of this kind of political system, it becomes clear why party elites in Belgium are not very keen on giving a great say to the rank and file. This bottom-up participation could undermine the delicate system of elite accomodation and eventually the whole fragile system of Belgian s consociational democracy. 4. The view of the members In the previous section, the involvement in practice of the members has been described as inglorious. This raises the question how members themselves perceive the intra-party procedures granting them greater involvement: are they aware of the limitations of these procedures, or do they continue to believe that by participating they can have a large impact upon the party s policy? The view of the members is of considerable relevance since the support of the rank and file for measures of intra-party democracy is crucial for the success of these measures (Young & Cross, 2002). We indicated already in the introduction of this paper that parties created intra-party procedures in order to attract (new) members. When members are not convinced they can have a real impact, it seems that these intra-party procedures cannot produce any effect. In this section, we will use data from various Belgian surveys among party members and party activists. There is first of all, the 1999 party activists project of the KU Leuven (further referred to as the party activists project ). 18 Non-salaried active members of Flemish political parties were interviewed about their activities in the party and about the perception of their influence within the party. Secondly, there was a party member survey conducted at the 2001 party leadership elections of the Flemish liberal party VLD (referred to as VLD project )

25 We start with a general assessment of the power the rank and file enjoy in practice. There was a list of 10 party actors or institutions put to the party activists and they were asked to indicate which of these actors or institutions possessed the least influence in the party. In order to make the question more concrete, the phrasing of that question referred to the influence in the composition of the party manifesto. Table 9: Who has the least influence within the party 20 (party activists project) (N = 296) % Rank and file Local party section Arrondissemental party section 6.42 Study- and workgroups 5.41 Party conference 4.05 Despite the changes in the statutory rules in favour of the grass roots members, party activists estimate that the rank and file have the least influence of a series of party actors. More than half of the party activists indicate the rank and file as having the least influence in the party. a) party leadership elections For the opinion of party members about party leadership elections, we will rely on an analysis of the VLD leadership elections. The VLD members were asked to assess the influence they enjoy by participating in intra-party elections. Table 10: To what extent do you have via these internal elections an influence on the party policy? (VLD project) (N = 330) % Cumulative % Very limited influence Limited influence Average influence Much influence Very much influence Don t know 7.27 No answer