Abstract
New-Flemish Alliance (N-VA) burst on the scene barely a decade ago and is now Belgium’s largest political party. One explanation for this success is that N-VA is not brand new but rose from the ashes of a dissolved party. How exactly should we differentiate between new and old parties? We use Barnea and Rahat’s (2011) analytical framework to assess dimensions of N-VA’s newness and capture the party at two stages – start-up and more developed. This shows that N-VA is a successor party, building on its predecessor’s ideology and programme, its electorate, activists and organization. However, we also find indicators that the party actively renewed in terms of ideology and party organization. The empirical evidence illustrates that newness of political parties should be conceived of as multi-dimensional, which allows for a more subtle approach to questions about the origin and varying success of new political parties.
Introduction
Since voters began to choose (Rose and McAllister, 1986) and traditional political parties became increasingly unable to capture votes, a considerable number of new parties have forced their way into party systems. Several aspects of these new parties have been researched, including the reasons for their emergence and breakthrough (Harmel and Robertson, 1985; Tavits, 2006), their specific organizational choices (Poguntke, 1987), their ideological profile (Lucardie, 2000), their survival chances (Bolleyer, 2013) and their fate after entering government for the first time (Bolleyer et al., 2012; Deschouwer, 2008).
One of the interesting debates in this literature is the seemingly simple question of how to clearly distinguish a new party from an old one, i.e. how exactly to define a new political party. One possibility is setting a cut-off point, like 1960, and then counting all parties that have appeared after that point (e.g. Mair, 1999). Yet that procedure is not uncontested, as the distinction it creates is too crude. It assumes that newness is one-dimensional and that the only dimension that matters is date of birth. But parties can be new in different ways. And some of the ways in which a party is new can make it less new than other new parties. A new party can be the result of the split of an existing party or of a few MPs breaking away from their party and forming a new one. A new party can also be the result of a merger of existing parties. In all these cases the newly created actor is then not completely new.
Even if one only counts parties without any partisan history at the moment of formation, some can still be less new than others. That is the case when it is new as a political party, i.e. as an actor participating in elections, but not organizationally new (Bolleyer, 2013). For instance, many green parties started off as social movements and decided only later to field candidates in elections. Yet another way in which parties can be new is their ideological profile when they enter the scene. Parties can defend established, ‘old’ ideologies or alternatively compete on new themes and issues (Lucardie, 2000).
Barnea and Rahat (2011) provide a summary of the debates on how to define a new party and present a systematic checklist of dimensions along which parties can be considered as new. In a similar vein, Arter (2012) builds on the concept of ‘successor party’ to make a distinction between new and pseudo-new parties. The concept originated in the literature on East Central Europe and refers mainly to former Communist parties that could rely on old resources and personnel to compete in Europe’s new democracies (Ishiyama, 1998; Waller, 1995). Arter suggests using the term for any party that builds on an older party.
Although this debate about newness and pseudo-newness might sound like a sterile battle of definitions, it is highly relevant. The ways in which or the degrees to which a party is new can have important consequences for its chances of success and survival in the long run (Bolleyer, 2013). Discussing and refining the classification schemes and trying to understand to what extent empirical cases are new and what the consequences of varieties of newness are is thus a relevant exercise. The focus in this paper is on New Flemish Alliance (N-VA / Nieuw-Vlaamse Alliantie). This party defending the independence of Flanders from Belgium is indeed an intriguing example of a ‘new’ political party with a fascinating track record. It was created in 2001, on the remnants of the regionalist party Volksunie (VU). VU dissolved itself after 47 years in parliament, although it had won 5.6 percent of federal and 9.3 percent of regional votes in 1999. There was an internal members’ referendum to decide on its future. None of the factions obtained an absolute majority, but the largest group founded N-VA. The party only managed to elect one MP in the 2003 federal elections. In the 2004 regional and again in the 2007 federal elections it did not run alone, but in an attempt to save itself had formed an alliance with the Christian-democratic CD&V. That alliance broke down in 2008, after which the N-VA ventured out alone again. It polled 13 percent of the votes in Flanders in 2009 (regional elections), 28.5 percent in 2010 (federal) and 32 percent in 2014 (both federal and regional).
The question is then whether this is a spectacular success story of a new party, or to the contrary a success that can be attributed to its pseudo-newness. It is, in other words, an ideal case to investigate, using the tools put forward in the literature to differentiate between forms and degrees of newness of political parties. This paper is further organized as follows. In the next paragraph Barnea and Rahat’s framework (2011) is presented in more detail, which is then used to discuss the case of N-VA. To that end we use a variety of methods and rely on various datasets: party manifestos, national election studies, membership figures and surveys, a local chair survey, biographies and financial reports. The conclusion summarizes the findings and defends the use of a multidimensional approach for the analysis of new political parties.
The multiple dimensions of newness
Political parties can be born in many different ways. And parties can also die in many different ways. The two processes are not unrelated, since the end of one political party can be the beginning of a new one. If both the beginning and the end of a party’s life are not clear-cut events, we need a more subtle way to assess the newness of political parties, for example the contribution of Barnea and Rahat (2011). In their attempt to conceptualize newness as a non-dichotomous variable, they suggest Key’s analytic breakdown of the party in three aspects or faces (Key, 1942; see also Katz and Mair, 1995): the party-in-the-electorate, the party-as-organization and the party-in-government. For each of these they list a couple of criteria, which results in a checklist to assess the newness of a political party (see Table 1).
Analytical framework for measurement of party ‘newness’.
Source: Barnea and Rahat (2011: 306).
While recognizing that parties can be new along several dimensions, Barnea and Rahat (2011) also suggest a dichotomous measurement to separate new from old parties. All parties are to be treated as old, unless their newness can be proven; in order to qualify as new, a party must have adopted a new label and “no more than half of its top candidates (top of candidate list or safe districts) [must] originate from a single former party” (Barnea and Rahat, 2011: 311). This second – dichotomous – way to measure newness seems less convincing. It decides rather arbitrarily that label and candidates are necessary conditions for newness, and it introduces soft and ambiguous criteria for the assessment of the candidates. One can indeed identify ‘safe districts’ for parties that compete in single-member districts, but there is no hard criterion to decide when a seat is really safe. In times of high electoral volatility even seats that have been safe for a very long time might not be safe anymore. For a party that participates in elections for the first time, there are probably no safe seats at all.
In his definition and measurement of a successor party, Arter (2012) puts forward a cut-off point. A successor party is any party that builds on a previously existing one, but it can only be considered as new if no more than two-fifths of all its candidates have previously run for the original party (Arter, 2012: 812). Again the focus on the candidate list as the crucial indicator and the arbitrary threshold of two-fifths of the candidates impose an unnecessary reduction of the empirical variation. If parties can indeed be new along several dimensions, it is useful to keep these dimensions on board. If newness is to be used as an independent variable explaining new parties’ survival chances, their long-term success or their impact on the functioning of party systems, it is important to be able to check which of the dimensions of newness has the expected effect. Some authors have stressed for example that organizational newness – as opposed to ideological newness or self-declared newness (a new party label) – influences new party survival and the way in which they cope with participating in government (Bolleyer et al., 2012; Bolleyer and Bytzek, 2013).
Instead of borrowing these dichotomous measurements, Barnea and Rahat’s (2011) richer checklist of eight criteria is used to describe in a more nuanced way the degree to which and the way in which a political party can be considered as new. In this paper we try to measure the newness of N-VA, created in 2001 in Belgium’s region of Flanders.
The birth of a new party
The first appearance of New-Flemish Alliance (N-VA, Nieuw-Vlaamse Alliantie) on the ballot paper was for the 2003 Belgian federal elections. The party presented candidates in the Flemish region and in Brussels only, but that is normal in Belgium. Belgian parties belong either to the Flemish (Dutch-speaking) or to the francophone language group; so all parties limit their electoral presence to the areas where their potential voters live. For N-VA the choice could also be linked to the fact that the party defends a further devolution of the Belgian state, with increased autonomy and possibly – in the long run – independence for Flanders.
N-VA is not the first party to mobilize on a regionalist agenda in Flanders. There is a long history of parties defending the Dutch language and its speakers in a country that was originally a francophone state. From 1918 onwards several consecutive parties were built on this centre-periphery cleavage (Deschouwer, 2012; Van Haute, 2005). N-VA’s immediate predecessor on this issue was the People’s Union (VU, Volksunie). VU had entered parliament in 1954 and reached its electoral peak and government participation in the 1970s, after which electoral decline began (De Winter, 2006; Noppe and Wauters, 2002; Van Haute and Pilet, 2006: 299; Wauters, 2005). In the 1990s doubts set in about the raison d’être of the party, which exacerbated friction between the party’s left-wing and right-wing factions. Belgium had by then been reformed into a federal state with far-reaching regional autonomy, thereby partially satisfying the party’s core programmatic demand. In response, one wing of the party, guided by the party elite, preferred a move to the left to develop a platform beyond the question of regional autonomy and to enter coalition governments again. Another wing of the party, mainly the middle-level elite and activists, preferred a more right-wing position with a clear demand for expanding Flemish autonomy, if need be to the detriment of executive power. Both positions clashed when in 2000 a new Constitutional reform was agreed and supported by VU in the Flemish regional government but rejected by VU (in opposition) at the federal level.
In 2001 VU held a referendum among its members to decide its fate. There were three options. The one defending the pursuit of more Flemish autonomy, led by Geert Bourgeois, turned out to be the most popular with 47 percent of members supporting it. The others were split between a group defending the more leftist path, led by former party president Bert Anciaux (23 percent) and a group that wanted the party to remain united (30 percent). With no option winning the majority of the votes, VU officially dissolved in June 2001. The left-leaning Anciaux group transformed itself into a short-lived new party – Spirit – that never participated independently in elections and broke up in 2009; members and leaders joined the Social Democrats (SP.a) or the Greens (Groen) (Van Haute, 2011). The Bourgeois group founded N-VA in October 2001. The new party was not allowed to keep the former name (Volksunie) because it had not convinced an overall majority of VU members.
At the 2003 federal elections N-VA did rather poorly: 4.8 percent of Flemish votes and only one seat in the federal House of Representatives (party president Geert Bourgeois). After this disappointing result the party accepted the Flemish Christian-democratic CD&V’s offer for an alliance at the 2004 regional polls. CD&V had been in opposition since 1999, after leading most Belgian governments since 1958, and it now counted on the few percentages N-VA might add to regain its position. The gamble worked. The alliance became the largest group in the regional parliament. CD&V headed the new regional government coalition; N-VA secured six seats (of the 124) in the regional parliament and joined the Flemish government. The 2007 federal elections were another electoral success for the alliance between CD&V and N-VA with 29 percent of Flemish votes. But they also marked the beginning of the end. The alliance had campaigned on a hard regionalist agenda, while the francophone parties refused the profound institutional reform this implied. A CD&V-led federal government could only be formed in March 2008 by deciding to postpone all discussions about state reform. N-VA denounced this as betrayal. It subsequently broke up the alliance – presenting itself as the only true alternative to the traditional parties which preferred power to keeping electoral promises – and went to the voters independently in the 2009 regional elections. The campaign led by the party’s popular new president Bart De Wever resulted in a surprisingly high vote share of 13 percent and government participation. In the early federal elections of 2010 N-VA, again in a standalone capacity, polled an astonishing 28 percent of Flemish votes, also making it the largest party in Belgium (with 17 percent of Belgian votes). The failure to find a compromise between N-VA (largest in Flanders) and the francophone Socialist Party (largest in French-speaking Belgium) on both state reform and socio-economic issues resulted in a coalition formation process that lasted a record-breaking 541 days (Hooghe, 2012). In the end a grand coalition of the traditional parties was formed (Socialists, Liberals and Christian-Democrats), without N-VA. The regional, federal and European elections of May 2014 were again a triumph for N-VA, now winning 32 percent of Flemish votes. N-VA currently leads the Flemish regional government (Prime Minister Geert Bourgeois) and participates in the federal government (for which it left the leadership to Charles Michel of the francophone Liberals MR).
Undoubtedly N-VA is an example of a party that is in some aspects a new party, albeit it one with a partisan history. It has not been built from scratch. The next section will assess in a systematic way how new it is by relying on Barnea and Rahat’s criteria. We will capture N-VA’s newness at roughly two stages: the start-up stage in foundation year 2001 and the more developed stage at the time of its real breakthrough in 2009.
The party in the electorate
The party label
Barnea and Rahat (2011) suggest three elements to measure the degree of newness of the party-in-the-electorate: label, ideology and electorate. For the party label the outcome is obvious. In VU’s members’ referendum to decide its future, the agreement was that the winner could only keep the old party’s name if it polled 50 percent of votes. That did not happen, although the Bourgeois faction came close with 47 percent. Whether keeping the old name would have been a real option for the party is difficult to say. The new name N-VA (chosen in October 2001) has offered the party the opportunity to present itself as different from the old and conflict-ridden VU. On the party’s website – but only on the Dutch pages – N-VA explains that it was born out of the ‘ruins’ of the VU (http://www.n-va.be/over-n-va/geschiedenis). Subsequently the N-VA label has become a familiar and strong name, prompting the party to ensure that local sections always use the national name and not local variations of it (see below).
On this first criterion N-VA is thus clearly new. The name of the party has remained the same throughout its (short) history. Also when allied to the Christian-Democrats, the label remained visible in the way in which the alliance presented itself: CD&V – N-VA.
The party ideology
The obvious continuity between VU and N-VA is the demand for increased Flemish autonomy, which both parties put at the centre of their party programme. Yet one of the reasons why VU collapsed was the discussion about the future of this Flemish nationalism and its combination with other ideological choices. VU’s leadership had very much been looking into a more leftist and green direction, with also some emphasis on institutional changes to improve the quality of democracy (a ‘new political culture’ – Maesschalck et al., 2002), rather than only on increased territorial autonomy. The group that won the internal referendum and that created N-VA positioned itself on a harder line for the nationalist theme and on a more right-wing approach to social and economic policy choices. This should then also be translated into the party programme. To assess the newness of N-VA, we analyse the electoral programme of VU for 1999 (VU, 1999) and manifestos of N-VA for 2003 and for 2010 (N-VA, 2003, 2010).
State reform
N-VA’s mission statement – adopted at formation and still in place – leaves no doubt about the party’s goal: it is a “democratic Flemish-national party which strives for an independent Flanders as a Member State of the European Union” (N-VA, 2001: 1). But how this end goal should be reached remains vague in the 2003 (the party’s first) manifesto. While VU in 1999 already refers to confederalism – i.e. a loose association of substates with few common policies – as its model for the future, N-VA does not even use that term in 2003. In that start-up stage, the party seems to keep its options open. The 2010 manifesto however describes its confederal goals frankly and in detail: N-VA wants to abolish the Senate and turn the Chamber of Representatives into a Bundesparliament consisting of representatives from the regional parliaments (N-VA, 2010: 36). Both VU in 1999 and N-VA in 2010 agree which level should have priority: it is up to the regions to decide which competences are transferred to the confederal level and how they will be funded (N-VA, 2010: 6; VU, 1999: 23, 25). In 2014 the N-VA elaborates on its envisioned confederalism and includes such details as the projected number of civil servants to be retained after reforming government services, for example (N-VA, 2014: 74).
It is quite evident that the N-VA has not introduced a new ideology into the Belgian or Flemish party system. There are nuances and variations in precision on the short-term and mid-term goals, but N-VA is a ‘purifier’ rather than a ‘prophet’ (Lucardie, 2000). That purifying aspect is also important in the relation to the radical right party Vlaams Belang, which aims for Flemish independence sooner rather than later and combines this institutional goal with a classic right-wing and populist approach to immigration and law and order (Billiet and Witte, 1995). When N-VA presents itself explicitly as a democratic Flemish-national party it marks the distance between N-VA and Vlaams Belang that has for a long time – and also against VU – claimed to be the only pure defender of Flemish nationalism.
Democratic renewal
VU had in its final years invested quite a lot in discussions on democratic renewal, a salient issue at the time. This included the possibility to organize referendums, expand citizen’s initiatives, limit the combination of political mandates, reduce clientelism, etc. VU – or at least its left-wing leadership – got along well with the ecologist party and they shared many concerns and ideas on these issues. On this point N-VA does not follow the old VU line, but adopts a more conservative position. In its 1999 program VU defended binding referendums and the direct elections of the local mayors (VU, 1999: 9); in its 2003 manifesto N-VA explicitly states that such proposals devalue representative democracy and should therefore be abandoned (N-VA, 2003: 26). In 2010 the topic of democratic renewal is absent from N-VA’s manifesto, but by then the topic had also disappeared from the political agenda.
Economic policies
On economic and social policies, N-VA also immediately shows that it does not want to keep the centre-left positions adopted earlier by VU. The 1999 VU manifesto and the 2003 N-VA manifesto for example both discuss the so-called unemployment trap – the unemployed have no incentive to accept a low-paying job if the wages are not (much) higher than unemployment benefits. In 1999 VU suggests a solution of lower taxes for the lowest incomes – thus making jobs more attractive (VU, 1999: 46). N-VA in 2003 copies this verbatim, but adds that welfare and benefits should be submitted to the same tax regime as low wages (so raise taxes for the unemployed) (N-VA, 2003: 16). Another telling example is the difference in proposals on fraud. In 1999 VU vows to battle tax fraud and tax evasion. The N-VA’s 2003 manifesto however pleads for a more favourable tax regime for the self-employed so they are no longer tempted to hide their total income from the revenue service (N-VA, 2003: 18). In 2010 N-VA states that fighting fraud by the unemployed should be a priority (N-VA, 2010: 26). The guilty parties thus change from tax evaders (VU) to the government taxing the self-employed too severely (N-VA, 2003) to the moonlighting unemployed (N-VA, 2010). This illustrates a clear shift towards more right-wing positions on economic issues.
Overall the evidence of N-VA’s ideological newness is mixed. On the nationalist issue it follows in the footsteps of its predecessor VU. It wants an independent Flanders and sketches the steps (confederalism) towards that end point. On other issues and especially on economic policies N-VA moves away from VU and positions itself very close to the liberal parties. That is already clear in 2003 and is strengthened towards 2010. The combination of free-market liberalism and Flemish nationalism is new, and with this new offer N-VA has been able to create a few electoral earthquakes already.
The electorate
The third criterion in Barnea and Rahat’s list is the difference between the new and the old party’s electoral base. Assessing this requires survey data. These are available for the 1999 federal elections (ISPO-PIOP, 1999), the 2003 federal elections (ISPO-PIOP, 2003) and the 2009 regional (and European) elections (PartiRep, 2009). The 2009 voter survey was a three-wave panel, with two pre-electoral waves and a post-electoral one. The survey’s panel structure indicates that N-VA picked up steam during the campaign, and we will therefore present the data for each of the waves to show how the characteristics of the voters change as the party expands its electoral reach.
We look at two variables present in the three surveys. The first measures voters’ preference for the distribution of competences in the Belgian federation. This has been asked with an 11-point scale, where 0 represents the opinion ‘all authority to the substates’ and 10 ‘all authority to the Belgian federal state’. Table 2 presents the average scores for the electorates of the most important parties and for the Flemish electorate as a whole. Two results stand out. First, on average, the position of Flemish voters evolved between 2003 and 2009, a period during which tensions between the two language groups in the country increased. Between 1999 and 2009 it dropped one full point on the scale, and crossed the symbolic midpoint with an average favouring the substate (Flemish) level.
Average position of voters on a scale measuring the preference to see all authority concentrated at the Flemish substate level (0) or at the Belgian federal level (10).
Note on VU / N-VA: VU dissolved in 2001, so only the results of 1999 refer to VU-voters. In 2003, Spirit supporters could only vote for the alliance SP.a-Spirit, but respondents were separated in the survey.
Source: ISPO-PIOP for 1999 and 2003; P
The second finding is that VU (in 1999) and then N-VA (in 2003 and 2009) have an electorate that clearly favours a transfer of authority to the substates, more so than the average Flemish voter. VU voters in 1999 stood apart as supporters of a strong regional autonomy (average of 3.38 on the scale). They are positioned even more towards Flanders than Vlaams Belang (VB)’s voters, who favour more Flemish autonomy, but who predominantly vote for the party because of its discourse on immigration and law and order (Swyngedouw et al., 1993). In 2003 N-VA voters were still staunchly defending regional autonomy (average of 3.27 on the scale), completely in line with the party founders, the former Bourgeois group in VU. Conversely, Spirit, established by the more left-leaning wing of its VU dissolution, attracts voters much closer to the mean position of the electorate, which matches the party’s choice to moderate its stance on the issue. This signifies that N-VA retained those VU voters who were most convinced of the Flemish-nationalist agenda. The strong anchorage of N-VA’s electorate on the more regionalist side of the spectrum remained stable in 2009. N-VA remains the party that – like its predecessor – attracts voters who believe that more authority (and possibly all authority) should be given to Flanders. In terms of its electoral base, N-VA is clearly VU’s successor, much more than Spirit ever was. We also find evidence of N-VA attracting a new electoral base, as the (slight) softening of the profile towards Election Day in 2009 indicates.
Another variable that assesses the degree of continuity between VU and N-VA is the voter’s self-placement on a left-right scale. On average, for the entire period, Flemish voters locate themselves slightly right of centre (Table 3). VU voters in 1999 are located almost exactly on the Flemish average. At that time, the party was very divided on the left-right issue. Part of the party elite was trying to move more to the left and signed an electoral alliance with a civil society movement called ID21 (‘Ideas for the 21st century’). In 1999 the party actually ran as VU-ID21, referring to renewal and more focus on New Politics rather than on Flemish nationalism. This push by the party elite created tension and ultimately led to the party’s collapse in 2001 (see above).
Average left (0) – right (10) self-placement of the voters in Flanders (1999, 2003, 2009).
Note on VU / N-VA: VU dissolved in 2001, so only the results of 1999 refer to VU-voters. In 2003, Spirit supporters could only vote for the alliance SP.a-Spirit, but respondents were separated in the survey.
Source: ISPO-PIOP for 1999 and 2003; P
VU’s dissolution led to the formation of two parties, each embodying one side of the left-right spectrum. This is reflected in the composition of N-VA’s electorate in 2003, who position themselves on average slightly more to the right than VU voters in 1999. Spirit, the more left-leaning wing of the former VU, formed an alliance with the Flemish Social Democrats (SP.a). With an average score of 4.18 on the left-right scale, Spirit voters are clearly on the left side of the scale, much closer to social-democrat SP.a and ecologist Groen than to N-VA. The profile of N-VA’s electorate moves further to the right in 2009. While in 2003 they were close to centre-right parties CD&V and Open VLD, at the start of the campaign in 2009 they are on average closer to VB (Vlaams Belang), the most right-wing Flemish party. In the course of the 2009 campaign, this right-wing profile softens with the expansion of its support base. Yet, for those who reported voting for N-VA in 2009, the average placement (5.70) is still clearly more to the right than for 1999 VU or 2003 N-VA.
The comparison of VU’s and N-VA’s electorate confirms the evolution of the party platform. On the regionalist issue the new party is very close to the old one: both defend a further devolution of powers and both attract voters who favour this policy. On the left-right scale voters move to the right, reflecting the party platform’s right-wing evolution on social and economic issues.
The party as organization
The formal and legal status
N-VA’s formal status leaves no doubt: the party is fully new in this respect. N-VA was created after VU’s dissolution, a formal ending which made it necessary for the new party to formally create a new organization. Belgium has no requirements for an organization to be accepted as a political party, but a formal and legally recognized organization is needed to receive financial support from the state and the substates (Weekers and Maddens, 2009). N-VA created such an organization.
The party’s institutions
The next question to be answered is whether N-VA’s institutions are separated and differentiated from those of the old party (see Table 1). To what extent is N-VA built on the former VU party organization and personnel and how has it evolved since? Four indicators are used: continuity of financial resources, local chapters, party membership and representatives.
N-VA in its start-up phase relied heavily on VU resources: it won the internal referendum and as such was allowed to keep most of the party’s assets (including real estate and personnel). VU’s financial resources were split proportionally according to the referendum’s results and to the number of representatives supporting each group (van Haute, 2005). The first criterion was quite advantageous for N-VA, since it had convinced almost half of the party membership. There was also a deal that those elected for VU in 1999 would formally remain members of the party group, which would continue the state subsidies to the party. These would then be distributed proportionally according to the group to which the MPs had committed themselves. Here the outcome was less positive for N-VA, since only seven of the 26 members of the regional and federal parliaments went to N-VA. Seventeen representatives went to Spirit and two of them decided not to remain in either party. However, N-VA’s impressive electoral growth (first in 2009 and then again in 2010 and 2014) meant it could rely on large amounts of state subventions (Weekers and Maddens, 2009). N-VA is now a rich party that does not owe its wealth to VU.
Local chapters are crucial in Belgian politics. Especially for new parties, the local party organization facilitates electoral sustainability and organizational institutionalization. Local members offer a stable vote share and are the ‘workhorses’ of the party during electoral campaigns. The municipal level is also the breeding ground for new political talent; a high number of regional and federal MPs in Belgium are actually also local councillors, aldermen or mayors. Municipal elections are held every six years, which makes them ideal for ‘surviving’ the lean years between elections on higher state levels.
The first local elections in N-VA’s lifetime were held in 2006, which is considered here as the start-up stage because this level was only then activated. A survey among local party presidents has been organized every six years since 1994 (Buelens et al., 2008; Deschouwer et al., 2013). Table 4 presents the number of local sections for both VU and N-VA. That number is derived from the number of questionnaires that was sent out, after having received the list of local sections from the party headquarters. The table shows that N-VA started out with fewer local sections than VU in 2000, but this drop was relatively small compared to the decline of local VU sections between 1994 and 2000. By 2012, well into the party’s developed stage, the number of local chapters doubled to an impressive 306 (on a total of 318 local municipalities in Flanders) and they are shown to be more active, too. One of the questions in the survey is about the frequency of local meetings, and the N-VA local sections indeed show not to exist on paper only. Almost all (85 percent) local sections report in 2012 that they have at least one general assembly per year.
Local party organization of VU (1994, 2000) and N-VA (2006, 2012).
Source: Belgian Local Chair Survey 1994, 2000, 2006, 2012.
Another indicator of the party’s strength is the number of local sections that participate in the local elections using the national name. It is quite common for local sections to opt for a local name, but this is not the case for the N-VA sections in 2012. All of them used the national name and label and campaign material, which is quite an impressive achievement. And it is an achievement of a party that has decided to set up a strong local party organization. The widespread presence of N-VA at the local level in 2012 resulted in large numbers of elected councillors, aldermen and mayors. These could in turn become candidates for the 2014 regional and federal elections. Although N-VA originally capitalized somewhat on VU’s local roots, its level of organizational development in 2012 outshines previous VU accomplishments. This shows that N-VA in its developed stage did more than just continue in its predecessor’s footsteps. This is N-VA functioning as a new party, and not as a party relying on the old VU organization.
The party activists
The third and final indicator under the heading of party organization is the newness of the rank and file. In Figure 1 we present the evolution of the VU and N-VA membership numbers. VU built a relatively large membership base in its first 20 years: from 2500 members in 1960, the party grew to more than 50,000 members in the 1970s, after which it continuously lost members, passing below the 20,000 mark in the 1990s. During its early years N-VA displays membership figures similar to those during VU’s first years. Lean at first, N-VA’s electoral growth in 2009 was accompanied by a fast rise in membership. Today it has a strong membership base, especially for a new political party.
Party membership figures over time (VU, N-VA: 1960–2012).
Whether the N-VA members came straight from VU is a question that can be answered by using membership surveys. There was no such survey during the start-up stage, but a 2013 survey among N-VA members (Table 5) shows that 22.3 percent are former VU members (Wauters, 2013). Knowing that N-VA counted 36,439 members in 2013, this amounts to 8126 people, very close to the number of members that N-VA claimed to have in 2002 (9931) and to the VU membership figure in its last year of existence (11,543). It is even higher than the proportion of VU members (5448) who voted for the ‘Flemish nationalist’ option defended by Bourgeois (founder of the N-VA) at the fatal VU membership referendum. This means that N-VA attracted former VU members beyond the core supporters of the Bourgeois option. Other information in the membership survey corroborates this idea of VU members moving seamlessly to N-VA. Respondents were asked their year of joining, and 22.6 percent of them declared having joined N-VA 12 years ago or more, which is technically impossible and probably corresponds to them joining VU. These elements are a clear indication that N-VA has established its membership base by recruiting heavily among former VU members. However, its membership base has also rapidly expanded. Today, these former VU members represent less than a quarter of the party’s grassroots, which means that for three quarters of the members, N-VA is a new home.
Previous party membership of N-VA members.
Source: Further calculations based on Wauters (2013: 20 (research note)).
The party in government
Barnea and Rahat (2011) put forward two indicators for measuring the newness of the party in government: the origin of the top candidates and the party’s policies. For the second indicator we cannot provide evidence. N-VA joined the Flemish regional government in 2004 as part of the alliance with the Christian Democrats. That government was a coalition of the alliance with the social-democrat party SP.a and the liberal party Open VLD. VU had been part of the Flemish regional government formed in 1999, and that was a coalition of Open VLD, SP.a, the green party (called Agalev then, now Groen) and VU. Policies of both governments are agreed on in a coalition agreement and are implemented collegially. In both cases the weight of VU and N-VA was obviously quite small. It is therefore impossible to produce a meaningful comparison between the policies of the old and the new parties. We will thus leave that indicator open.
The other criterion looks for the origin of the top candidates. When N-VA participated for the first time in the federal elections of 2003 it had to produce six candidates to lead the lists in each of the electoral districts. None of these positions could be considered as safe seats. The party campaigned with the slogan ‘De N-VA erin’ (‘vote in N-VA’), which marked the ambition to cross the 5 percent threshold in at least one of the districts. That is indeed what happened: the only candidate elected was Geert Bourgeois, former VU president and then N-VA president. The five other top-of-list candidates, who were not elected, were all former VU activists; three of them had been members of parliament and two of them had been active in the party organization. In this very first phase, N-VA relied completely on the old VU personnel.
This link between VU and N-VA personnel persists in the later stage of the party’s history. In 2013 N-VA counted 20 representatives at the national, regional and European levels out of a total of 58 who have been formerly engaged in VU (often at the local level). This does not even include sons and daughters of former VU activists, who were too young at the time to hold a representative mandate but are now involved in N-VA. Furthermore, the highest positions are all granted to former VU (and thus more experienced) politicians: Bart De Wever (party leader), Geert Bourgeois (Prime Minister in the Flemish Government), Jan Jambon (federal Minister of the Interior) and Jan Peumans (Chair of the Flemish Parliament). When N-VA was created, it was built more on VU members than on VU’s elected representatives. After initially breaking away from the VU top, it went on to build itself, relying heavily on former VU middle-level elite.
Conclusion
We have argued that the newness of a political party is not a straightforward and dichotomous affair. The analysis of the first years of the N-VA has illustrated that newness is indeed a multidimensional characteristic of a party. The findings are summarized in Table 6. For each of the three faces of the party, we have found mixed evidence. In some respects N-VA is clearly and fully new, while in others its newness is relative, and we see a party that was able to put itself on the map by relying on some voters, activists, party organization, representatives and financial resources inherited from the dissolved VU. If we were to follow the rules put forward by Barnea and Rahat (2011) for deciding whether the N-VA is really new (party label and top candidates), we would have to label it as ‘old’ because the candidates of the 2003 lists all had a VU history. If we were to follow Arter’s definition of a successor party, N-VA would certainly qualify. It is “a party which is nominally and legally a new entity that takes the place of, and fills some of the political space vacated by a single, defunct party of origin” (Arter, 2012: 807). Yet the aim of this contribution was showing that we should not stop there.
Summary of results.
Source: Based on Barnea and Rahat (2011: 306).
The mixed evidence supports the idea that using a clear and unambiguous cut-off point between new and old parties would mean that we fail to see the empirical nuanced picture. Whether N-VA or actually any other new party is truly and fully new cannot be answered. Labelling the party as old because it strongly relied on the dissolved VU’s party personnel for selecting its top candidates, as suggested by Barnea and Rahat (2011) and by Arter (2012), would favour that aspect of newness and see it as more important than, for instance, ideology, party membership or party organization. Especially if one wants to understand why some new parties perform better in winning votes, seats or office, or why some new parties are able to survive while others disappear after one or a few electoral cycles, the differentiation between dimensions of newness might actually offer us much more nuanced insights.
One other finding from our analysis of N-VA – related to these survival chances – is the importance of party agency. We have seen how N-VA could originally rely on some resources inherited from its predecessor VU (Volksunie). Yet if N-VA is towards 2010 much more a new party than at the time of its creation, that is the result of a conscious choice to really build a new party. Both in ideological and in organizational terms N-VA has deliberately followed its own path. By combining Flemish nationalism with a centre-right position on social and economic policies it has put a new programmatic offer onto the electoral market that has so far proven to be extremely successful. By allying itself to the Christian-Democrats for one regional, one federal and one local election it bought time and fully seized that opportunity to build a solid party organization and to attract a fairly broad party membership. It is, in other words, by combining newness and continuity, heritage and conscious strategic choices that N-VA has been able to write its own success story. The case of N-VA therefore illustrates convincingly that we need a multidimensional definition of the newness of political parties to move forward in our understanding of the dynamics of these parties.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Stefanie Beyens was supported by the Research Foundation – Flanders (grant number GO26611 N).