Ideas and the Study of Political Parties: The Added Value of the Discursive Institutionalist Approach^*

Las ideas y el estudio de los partidos políticos: el valor agregado del institucionalismo discursivo

]]>

Johannes A.A.M. van Gorp^**

Abstract

There are four methods often used to study party locations along salient cleavages in party systems: expert surveys, content analysis of party manifestos, media analyses and broad opinion surveys of electorates. Most of the literature that explores the relationship between political parties and these cleavages has two shortcomings: (1) it treats parties as cohesive units, and (2) while the combination of policies adopted by a party is considered important, almost none consider how they are communicated to the voters. Both matter: if politicians within a party remain torn over an issue, parties will have a difficult time agreeing on what position to take and cannot effectively communicate their ideas to the public. That is why one should examine the coordinative and communicative discourses of parties. Coordinative discourse encompasses the process whereby political actors agree on a policy program, while communicative discourse is the process through which this program is framed. To provide evidence for this argument, communicative discourse in the Netherlands is discussed.

]]>

Keywords: Discursive institutionalism, party politics, discourse, Dutch politics, party cohesion

Resumen

Hay cuatro métodos usados a menudo para estudiar los posicionamientos de los partidos políticos sobre los clivajes prominentes en los sistemas de partidos: encuestas de expertos, análisis del contenido de los pronunciamientos partidarios, análisis de los medios masivos de comunicación y encuestas de opinión del electorado. La mayoría de la literatura que explora la relación entre partidos políticos y dichos clivajes tiene dos deficiencias: (1) trata a los partidos como unidades cohesivas, y (2) mientras la combinación de políticas adoptadas por un partido es considerado importante, casi nadie considera cómo han de ser comunicadas a los votantes. Ambos asuntos son importantes: si los políticos de un partido permanecen divididos sobre un tema, los partidos tendrán dificultad para ponerse de acuerdo en la posición que deben tomar y no podrán comunicar efectivamente sus ideas al público. Por eso uno debe examinar los discursos de coordinación y los discursos comunicativos de los partidos. Los discursos de coordinación engloban el proceso por el cual los actores políticos acuerdan un programa de políticas; mientras que el discurso comunicativo es el proceso por el cual este programa es enmarcado. Para proveer evidencia en respaldo de este argumento, el discurso comunicativo en los Países Bajos es discutido.

1. Party Politics as a Field of Study

Within the study of party politics, four methods are most often used to study party locations along salient cleavages in party systems: expert surveys, the content analysis of party manifestos, media analyses and broad opinion surveys of electorates (Gemenis and Dinas 2010: 181; Volkens 2007). The approaches link parties to ideas, placing them on an imagined spatial plane composed of cleavages that best explain existing divides within a state. According to Bartoloni and Mair (1990: 215), a “cleavage” consists of a socio-political fault line between social groups that is strong enough to structure party political conflict between them. It must contain at least three elements:

An empirical element, which identifies the empirical referent of the concept, and which can define in socio-structural terms; a normative element, that is, the set of values and beliefs which provides a sense of identity and role to the empirical element, and which reflects the self-consciousness of the social group(s) involved; and an organizational/behavioral element, that is, the set of individual interactions, institutions, and organizations, such as political parties, which develop as part of the cleavage.

Accordingly, for politically contentious issues to form a cleavage, they have to not just result in conflict. They have to, “generate group consciousness or political organizations that structure debate and behavior in an enduring fashion” around these issues (Webb 2002: 116).

The four approaches most often used all have one major common shortcoming: while they are concerned with the combination of policies a party holds, none really are concerned with how these positions are communicated to the public. If parties cannot communicate these ideas effectively to voters, voters will have a difficult time linking parties to the policies they stand for (Van Gorp 2012: 17). Political parties, after all, are supposed to connect ‘leaders and elected officials around common programs’ (Norris 2005a: 3). If politicians within the same party differ in their views, this will be difficult to achieve. Disagreements, in turn, will make it much more difficult for the party to clearly communicate its position on important issues to voters. This is why much of the literature on political parties argues that party cohesion, over time, is the norm.

While the above methods used to study the positioning of political parties do not allow one to measure such discord, a discursive institutionalist approach does. The virtues of this approach have been extolled by Vivien Schmidt (2006a; 2008; 2010) and Van Gorp (2012). Discursive institutionalism examines how discourse is affected and incentivized by institutional settings. It is especially interested in “coordinative” and “communicative” discourses. While the former is concerned with how epistemic communities and political elites communicate with one another on what policies to adopt, the latter is concerned with how those positions are communicated to the broader polity. This article argues that by examining the coordinative discourse of political parties, one can better understand why cohesion is sometimes lacking on issues that are salient to voters. To examine this coordinative discourse, 51 Dutch politicians were interviewed and asked the same nine questions about socio-cultural issues. Answers were then coded to create a socio-cultural index for each party. By discussing evidence from the Dutch case, it shown that despite the salience of socio-cultural issues to Dutch voters, traditionally dominant parties have remained divided on these issues. Internal discord within these parties helps explain why the communicative discourse of these parties surrounding socio-cultural issues has left much to be desired. Accordingly, the positioning of these parties on these issues remains unclear to voters.

The Netherlands serves as an especially interesting case because most of the newer parties do not conform to the discursive behavior one would expect from political parties in a compound polity (Schmidt 2002: 243-246). In such polities, coordinative discourse is described as being elaborate, while communicative discourse is described as being “thin” (Schmidt 2006: 258-262). For parties such as the PVV, the opposite is true. Having seen how internal discord brought down both the LPF and the short-lived General Elderly Alliance,¹ Wilders opted to create a party with only one member – himself (Vossen 2013: 179-181). This means that the party leadership can focus on establishing a clear communicative discourse on issues it deems important and does not have to worry about convincing potential dissidents within the party to tow the party line. Those that do publicly challenge Wilders, like Hero Brinkman, are simply forced to leave the party. Similarly, parties such as the SP, D66 and the Greenleft have much clearer communicative discourse on socio-cultural issues than the VVD, CDA and PvdA have. Considering how important socio-cultural issues have become to Dutch voters, the decreasing vote share of traditionally dominant parties might be partially explained by the continued inability of these parties to clearly communicate their socio-cultural policies to voters.

2. The Existing Literature on Party Cohesion

]]>

The existing literature on party cohesion assumes that if there is disagreement within a party over an issue, these disputes will be short lived (Hazan 2006; Jensen 2000: 210-236; Lucardie, Marchand and Voerman 2007; Özbudun 1970; Thomassen and Andeweg 2007). There is a strong rational choice logic underlying this argument, with the assumption being made that politicians are ultimately self-interested and desire to be reelected above anything else (Thomassen and Andeweg 2007: 168-173). In proportional representation systems, most owe their job to party leaders. Once elected to parliament, party leaders decide who will fill sit on what committee. Leaders also decide the placement of members on the party list prior to an election. If a Member of Parliament has deviated too far and too often from the party line, he or she runs the risk of being placed lower on that list. One could even be placed on an unelectable position.

As a last resort dissidents can choose to leave their party. However, defectors face low odds of being reelected. In a study of the British, German, Belgian and Dutch party systems from 1945 to 2005, Lucardie, Marchand and Voerman (2007) found that defectors often did not get reelected. Belgian defectors had the most chance of being reelected (40.5 percent). In Germany and Great Britain, only 25.2 and 22.8 percent, respectively, returned to parliament following the next election. Dutch politicians faced even worse odds. Only 8.3 percent –three out of 36 defectors– were reelected. Because of all these tools available to a party through which it can enforce unity, it is assumed that most politicians within a party will ultimately agree on issues that really matter to voters (Thomassen and Andeweg 2007: 173). After all, presenting a unified front to voters come election time is important to parties.

3. Discursive Institutionalism and Party Cohesion

]]>

The turn towards institutionalism as a way of explaining the political world has been well documented (Hall and Taylor 1996; Lichbach 2002; Schmidt 2010; Aspinwall and Schneider 2000). The three “new institutionalisms” that developed independently from one another ─historical, sociological and rational choice institutionalism– sought to move beyond the behavioralism of the 1960s and 70s by refocusing on the role that institutions play in shaping political outcomes. Institutions in their simplest definition are seen as rules. Their study can vary from those on organizations and formal rules, to norms and informal rules (Streeck and Thelen 2005; Steinmo 2008).

However, explaining change in the political world through “static institutions” can be problematic (Schmidt 2006a: 11). That is why one has recently seen a turn to ideas in all three institutionalist approaches (for a discussion see Béland and Cox 2011). Some have stayed within the confines of their approach, while others have moved into what Schmidt has labeled “discursive institutionalism”. She sees this institutionalism as “an umbrella concept for the vast range of works in political science that take account of the substantive content of ideas and the interactive process by which ideas are conveyed and exchanged through discourse” (Schmidt 2010: 3). As an approach, it is closely related to sociological institutionalism. The difference between the two is more one of degree than kind: discursive institutionalism treats ideas as “dynamic constructs”, while sociological institutionalism treats them as “static structures” (Schmidt 2008: 320). Through the tracing of discourse, discursive institutionalism allows one to trace when change happens within a political system. Discourse, as defined by Schmidt, is “whatever policy actors say to one another and to the public in their efforts to generate and legitimize a policy programme” (Schmidt 2002: 210). The institutional context of a country also matters, for it frames the discourse.

The emphasis in this approach is on examining the discursive interactions that take place in any given system. It focuses on agents, or ‘who is speaking to whom’. There are two ideal types of discourse: (1) “Coordinative discourse” and (2) “Communicative discourse”. While the first examines how political elites and epistemic communities decide to adopt certain policy positions, the latter is concerned with how those policies are communicated to the public. Schmidt has used this approach to explain differing outcomes between countries, distinguishing between “simple” and “complex” polities. Also referred to as “single-actor” and “multi-actor” systems, Schmidt argues that due to a stronger executive and the limited need to compromise with others, simple polities are characterized by stronger communicative discourse and weaker coordinative discourse. The opposite holds for compound polities. They will see stronger coordinative discourse because political actors need to coordinate and compromise with one another (Schmidt 2006b). This makes their communicative discourse thinner, especially for those involved with governing. This is because those governing likely do not want to have the compromises reached behind closed doors publicly scrutinized.

While Schmidt has employed the discursive insitutionalist approach to explain outcomes at the national level vis-à-vis the welfare state (2002) and perceptions of the European Union (2006) in Europe, this article argues it is also well suited to study party cohesion and communication. The literature on political parties takes ideas, institutions and interests seriously. It is useful to consider it in terms of discursive institutionalism in order to be able to theorize with greater precision the ideational and discursive processes. Though they do not identify themselves as ideationalist, the four main methods used to study party politics are all concerned with linking parties to ideas. However, while all are concerned with the combination of policies each party holds, almost none consider how they are communicated to the public.²

If parties do not effectively communicate their positions to voters, this will hurt those parties at the polls. This is well established in the political communications literature. Within this literature, two different types of framing are identified: “emphasis framing” and “equivalence framing”. While the latter examines how the way in which the same issue is framed influences outcomes, “emphasis framing” analyzes the, “process in which competing frames emphasize different messages and arguments in a policy debate” (Shaffner and Sellers 2010: 2-3). What is called emphasis framing in the communications literature is referred to as communicative discourse in the discursive institutionalist approach. In essence, it is the study of which frame wins out over another.

Before a political party can even communicate its ideas to the public, it first needs to agree on what ideas to promote. It needs to “connect leaders and elected officials around common programs” (Norris 2005a: 3). Put differently: it has to decide what the content of their frame is going to be regarding an issue. This is easier to do if its politicians hold similar views. If they differ substantially in their opinions regarding an issue, disagreement is likely to ensue. In the literature on party cohesion discussed above, it is assumed that such discord will not last long on issues that matter. After all, parties have the means to remove those politicians that do not adhere to a party’s views. In the discursive institutionalist approach this process of reaching agreement on what ideas to communicate to the public is described as coordinative discourse. As can be seen in Figure 1, both types of discourses do not take place in a vacuum: they are shaped and constrained by the institutions that underpin a political system.

According to Schmidt’s ideal types, the Netherlands would be categorized as a compound polity. One should expect a stronger coordinative discourse to be present between the various political parties and epistemic communities. However, when it comes to communicative discourse, one should expect it to be weaker, since “political leaders are schooled in communicating only in vague terms on agreements reached among the wide range of actors involved in the coordinative discourse of policy construction” (Schmidt, 2006b: 45). Even when in opposition, politicians in a compound system are expected to speak vaguely. While today the governing parties are their foes, tomorrow they may be their coalition partners. As a political actor you do not want to burn any bridges unnecessarily. Reaching a consensus in such a fragmented system takes time and is difficult. This discourse framework goes a long way towards explaining why traditionally dominant political parties were so slow in forming a discourse concerning immigration and Dutch identity once the issue became more salient.

Upon closer examination of political discourse in the Netherlands these last two decades, the above theoretical framework should be amended to distinguish between “newer” and more “traditional” political parties. Schmidt contends that communicative discourse becomes more pronounced in compound polities only during election times and during periods when coordinative discourse breaks down (Schmidt 2006b: 262). However, politicians such as Geert Wilders (Freedom Party - PVV), Pim Fortuyn (LPF), Alexander Pechtold (D66) and Jan Marijnissen (Socialist Party - SP) have continued to employ polarizing discourse even during non-election times (Van Gorp 2012). With his party serving as a minority partner of the governing coalition from 2010 to 2012, Wilders for a while got the best of both worlds. More stringent immigration and integration policies were proposed, while at the same time the PVV had no reason to moderate its discourse. As for the other parties, D66, the SP, and the GreenLeft also have displayed clear communicative discourse vis-à-vis socio-cultural issues during both election and non-election times. Parties such as the CDA, VVD and PvdA have not and continue to conform to Schmidt’s expectations.

Despite the increased importance of socio-cultural issues to voters, parties such as the CDA, VVD and PvdA have remained internally torn over them. Socio-cultural issues became politicized in the 1990s. Following the 1960s, the religious cleavage that had been so important to explaining post-war Dutch politics began to lose much of its salience amongst voters (Lijphart, 1975). In the 1990s a socio-cultural one began to take its place as a significant segment of the electorate not only became anti-immigrant, but also voiced that it found it one of the most important issues facing the nation. For example, close to half the voters identified “immigration” as the most important issue facing the nation in the 1994 elections (Boomgaarden 2007: 104). Parties now had to decide what position to take on these issues. In the Netherlands, this is decided through coordinative discourse that involves politicians and the think-tanks attached to most parties. They decide what the party manifesto will be prior to each election.

Initially only the leader of the VVD, Frits Bolkestein, was willing to employ frames that not only were critical of immigrant integration, but also questioned existing immigration policy and whether Islamic culture was really equal to that of the West. He first voiced such criticism in the early 1990s and repeated these claims in the 1994 and 1998 elections. Even after Bolkenstein broached the subject, most other politicians continued to avoid it.⁴ Though a considerable segment of his own party remained critical of his positioning on socio-cultural issues, voters rewarded the VVD handsomely in 1994 and 1998. This changed when Bolkestein left for a position in Brussels following the 1998 elections. With Bolkestein gone, the VVD under the leadership of Hans Dijkstal refrained from further politicizing these issues and moved further to the left.

With Pim Fortuyn capitalizing on the vacuum left on the right of the political spectrum as a result of the VVD’s moderation on socio-cultural issues, the national elections that followed in 2002 were a watershed moment with regard to the discussion of immigration and integration in politics. Taking an even harsher stance against immigrants and the perceived problems associated with their integration than Bolkestein had done resulted in large electoral gains for the LPF. In an analysis of voter motivation in the 2002 elections, Van Holsteyn and Irwin found that LPF voters cited “sending back asylum-seekers”, and “foreigners should adapt” as the two main reasons for voting for the party (2003: 41-68). Despite Fortuyn’s assassination by a lawyer with close ties to the animal rights movement, the campaign he had run resulted in an unprecedented 26 seats out of 150 in the Second Chamber for the LPF. Ranked on the Pedersen index of electoral volatility, these elections ranked “fourth among all West European general elections in the period 1900-2002.” (Van Holsteyn and Irwin: 2003). Though the LPF as a party collapsed soon after the 2002 elections⁵ –in the 2003 parliamentary elections the party fell to 8 seats and gained 0 during the 2006 elections– the subsequent rise of the PVV has shown that anti-immigrant sentiment can still lead to large electoral gains.

Parties willing to take a critical stance towards immigration and integration policies –and communicate that position effectively– have made significant electoral gains. Neither the VVD nor the CDA were able to step into the void left by the LPF, as both parties remained publicly internally divided over socio-cultural issues. Instead, Geert Wilders left the VVD after disagreements over the party’s socio-cultural positioning and established his own party, the PVV. That party made impressive gains in two successive elections. Mainly because of its critical framing towards socio-cultural issues, the PVV became the third largest party in parliament in just five years. Following the June 2010 elections, the party became the third largest in parliament. Through their support for the VVD/CDA led minority government, the PVV also played an important role in the shaping of policy from 2010 to 2012. As for the parties that used to dominate Dutch politics, the CDA, VVD and PvdA: they received just over 50 percent of the vote in the last election.

The more pro-immigrant, pro-European and anti-nationalist a politician’s answers were, the higher their individual score was. Inversely, if answers were anti-immigrant, anti-EU and pro-nationalistic, the politician’s individual score was lower. Examples of how answers were coded are given below for two of the questions asked. One of the questions asked of the politicians was whether it matters where immigrants come from for their integration, or if the country of origin does not matter at all? This question was coded on a scale from -1 to +1. Politicians who argued that the country of emigration mattered for the integration of immigrants into Dutch society received a score of -1. Examples of such responses were: “Cultural differences matter. If the culture in the country of emigration is closer to that of the Netherlands, then you have a much easier time integrating.”⁶ Another: “We have most problems with non-western immigrants.”⁷ Politicians who responded that the country of emigration matters but that other variables are just as important for the integration of immigrants, received a score of 0. An example: “The country plays a role, in various manners, but it is always in combination with other factors. Educational level, socio-economic position... whether you stay with your own group here in the Netherlands [are all factors that matter].”⁸ Finally, respondents that said the country of emigration did not matter at all for integration into Dutch society received a score of +1. Amongst these responses was the following: “The background of immigrants matter, but the land of origin does not…It matters whether people studied or not, whether they are from rural regions or from cities. It matters how wealthy immigrants are and how much knowledge of the world they have, how cosmopolitan one is… The country itself is not an explanatory variable.”⁹

Some answers were coded on a scale from -2 to +2. For example, one question asked the following about integration policy: “In January 2000, Paul Scheffer published an article entitled “The Multicultural Drama” in De Volkskrant. In it he was very critical of integration policy in the 1980s. Had that policy really failed, or were there also positive aspects to the policy of ‘integration with retention of one’s own identity’?” Respondents who argued integration policy of the 1980s was a total failure and had no redeeming qualities received a score of -2. Such answers included the following: “It was failing policy. The emphasis was too much on the ‘new Dutchmen’, while little thought was given to the ‘original Dutchmen’, the man that lives in an apartment building in Rotterdam and was just fired from his job because of increased competition… and encounters the problems [associated with immigration] daily.”¹⁰ Another: “It was totally failed policy. We wouldn’t have the mess we have now if the policy had been successful.”¹¹ If politicians identified the policy as failed, but then identified just a few positive aspects, their answer was coded as -1. Two examples are: “The policy failed in the sense that it failed to stimulate a large group of people to actively participate in society... But the thought process behind the policy was a positive one,”¹² and, “While the policy was failing, the integration itself has been relatively successful.”¹³ The answer was coded as 0 if the emphasis of the answer was on equally identifying both positive and negative aspects of integration policy in the 1980s. Answers were coded as +1 if respondents claimed there were problems with integration policy in the 1980s, but spent most of their time identifying its good qualities. Such answers were, “The Netherlands is a multicultural society…You do need a common language and values that people follow,”¹⁴ and, “I enjoy living in a country where you see different identities and cultures…At the same time I don’t live in a neighborhood where you encounter 20 different cultures a day.”¹⁵ If respondents refused to identify the policy as failed and claimed identity could still be retained while integrating, their answer was coded as +2. An example: “When integration is successful, you don’t have to relinquish your whole identity. It’s about your behavior, that’s what matters.”¹⁶

Figure 2 shows how individual scores were then compared within parties to see how similar opinions were within a party over socio-cultural issues. This serves a strong indicator as to how much internal discord remains within parties over socio-cultural issues and how fraught their coordinative discourse is likely to be. The placement of political parties on this figure from left to right is based on how Van Kersbergen and Krouwel (2008) positioned parties on the socio-economic cleavage in their piece on the Dutch centre-right. In their study parties rank as follows on this cleavage: GreenLeft, SP, PvdA, CU, SGP, D66, CDA, VVD, and PVV. Despite having been salient to voters for over two decades, scores made clear that the CDA, VVD and PvdA, as of 2011, remained internally torn over socio-cultural issues, while politicians of the SP, D66, GreenLeft and PVV were unified in their views towards these issues. Therefore, it is not surprising that these parties have not had major public disagreements over socio-cultural issues. Accordingly, they have also been very effective in communicating their position on the socio-cultural dimension. In other words: they have had both successful coordinative and communicative discourses. This stands in stark contrast to the CDA, VVD, and PvdA. These three parties have remained internally divided over socio-cultural issues. They continue to be so today, despite the importance of these issues to a significant portion of the electorate.

One can speculate as to why these traditionally dominant parties continue to remain more ]]> The Brawlers Party (De Vechtpartij), Niemantsverdriet (2014) argues that internal discord within the PvdA is to blame on its “ideological insecurity” that has dominated since 1986, the unpleasant internal culture that exists within the party, and the temptation of Dutch labor politicians to publicly self-chastise themselves (256). Alternatively, despite their ]]> I find it important that the ]]> PvdA remain a broad people’s party and that the tensions ]]> that are apparent in society... also remain within the Labour ]]> Party. That is the way it has always been” (Spekman ^]]> 2011). After sliding in the polls throughout 2011, the leadership of the CDA decided it was time to analyse why this was happening. Despite all the changes proposed, the report that this produced concluded that: “The CDA is and remains a ]]> Christian-Democratic people’s party” (Du Pré 2012). Some of those that were interviewed had held leadership positions in the PvdA and VVD. They were asked directly about this discord in their parties. One PvdA politician emphasized that since his party was a “people’s party”, it would naturally have a ]]>