Maintaining Partisan ties

Abstract

Parties coordinate on a range of activities. They invite leaders from other parties to their national meetings, run joint electoral platforms and even form parliamentary factions and coalition governments. The implications of regular cooperation such as the case of pre-electoral coalitions (PECs) for party positioning are unexplored. Parties form PECs to reduce competition for voters with ideologically close competitors and to signal their ability to cohesively govern. Building on this logic, we argue that parties’ preferences converge in PECs to demonstrate their ability to govern together and diverge when parties observe that this tactic has failed to attract voter support in past elections. We demonstrate support for our approach using data on electoral coalition participation, party positions and parties’ internal speeches. Additional evidence from an extreme case of an enduring electoral coalition in Germany shows that PECs have dramatic effects on parties’ positions.

Keywords

coalitions party coordination pre-electoral coalitions preferences sister parties

Party competition creates strange bedfellows because governance incentivizes cooperation between diverse groups. Parties coordinate their electoral and government formation strategies with their erstwhile competitors with tools such as pre-electoral agreements to maximize their potential for attracting votes and controlling government. Election rules often structure the shape and content of this coordination. Small and large parties in France, for example, often reach agreements to avoid direct competition for the same parliamentary seats and support one another’s candidates on the second round of ballots (e.g. Golder, 2006; Spoon, 2011). Despite knowledge of short term or irregular agreements between parties, less is known about the influence of regular participation in pre-electoral coalitions (PECs) on larger parties’ internal politics and policy positions. Why would ideologically distinct electoral competitors tie their future government success? Is it reasonable to treat parties as a single party-in-the-electorate and party-in-government, when each party acts organizationally separate?

We propose an answer to these questions by considering PECs and parties’ consequent electoral success. Our perspective builds on research focused on coalition and electoral coordination (Golder, 2006; Ibenskas, 2015; Ibenskas and Sikk, 2017) to account for varying levels of organizational, electoral and governmental cooperation. From this perspective, the effect of partisan cooperation on intra-party politics depends on the mode of coordination. Given their electoral motivations, parties in PECs that increase their parliamentary support respond by shifting their preferences closer together. When these coalitions fail to increase their electoral support, the parties involved seek to clarify their positions by distancing themselves.

We empirically explore this perspective with evidence on party positions from the Comparative Manifestos Project (CMP) and PECs from Golder (2006). Specifically, we test if parties’ relative positions respond to experiences in PECs and if parties’ preferences move in tandem with those of their coalition partner. We show broad, comparative evidence that parties’ preferences shift closer to their PEC partners in response to electoral gains but deviate when the coalitions no longer serve them in a sample of 20 countries from 1950 to 1998. The results are clearest for governing coalitions. Based on these results, we delve deeper into the process by examining an unlikely case for preference divergence: the long-standing union between the German Christian Democratic Union (CDU) and Christian Social Union (CSU). Despite distinct organizations, leaders, and even regional bases of support, the parties cooperate so closely that they are often treated as a single party by academics or referred to as ‘sister parties’ in the popular press. These parties offer a challenging situation for party researchers as the parties’ distinct goals are usually combined in single, joint manifestos. These extreme examples of party collaboration provide unique evidence on the way that coalitions reconcile their diverse preferences. If these parties exhibit distinct goals, then this evidence would undermine claims for treating the party in electorate and party in government as unified.

The results from our analysis hold important implications for theories of party competition in advanced democracies. Representation requires choice between ideological competitors. Collusion between parties limits citizen choice and confuses the clarity of policy responsibility (e.g. Powell and Whitten, 1993; Whitten and Palmer, 1999), but PECs might provide voters with choices that are more likely to control government. Evidence of a dynamic representation process would demonstrate the means by which parties maintain their ties, yet actively promote the goals of their distinct constituencies.

Partisan cooperation and competition

Although parties distinguish themselves in elections through their statements of preferences (e.g. Adams, 1999; Adams and Somer-Topcu, 2009; Downs, 1957; Ezrow et al., 2011; Spoon, 2011; Tavits, 2007), they also engage in a range of collaborative activities. While most research on election campaigns conceptualize parties as competitors, electoral rules, party system fractionalization and parliamentary governance incentivize cross-party collaboration. The existence of regular electoral alliances and the so-called sister parties mark the most extreme form of these collaborations. Despite the regularity of pre-electoral and governing coalitions as well as other forms of coordination, the effect of cooperation on parties’ policy strategies and intra-party politics is understudied. Party definitions do not preclude cross-party coordination, yet the existence of cooperation suggests that parties see their electoral fortunes as tied to organizations, which in other times might be their ideological competitors. Perspectives on party system change and election strategy offer some tools for understanding party coordination.

More concrete forms of cooperation reflect ideologically proximate parties’ efforts to maximize their joint likelihood of entering government in response to electoral and government formation rules. In mixed electoral systems, such as Germany (prior to the 2013 federal election) and New Zealand (after 1996), for example, smaller parties encourage their supporters to vote for ideologically close larger parties on the majoritarian list and for themselves on the proportional list (Bawn, 1999; Gschwend, 2007). This split ticket voting increases the likelihood of an ideologically close post-election parliament by supporting parties that are more likely to gain a plurality of votes on the single-member district while still contributing to the smaller party’s seat share through the proportional allocation of seats.

Even in single-member district elections, parties coordinate to increase the likelihood of winning an ideologically close majority in parliament. Parties on the ideological left in France agree to not directly compete against each other in select districts to avoid splitting their vote. In the second round ballot, parties then lend their support to the most ideologically close party (Blais and Indridason, 2007; Golder, 2006; Spoon, 2011).

From an electoral standpoint, ideologically close parties competing for voters with similar preferences likely benefit from coordination. Direct competition among ideologically close competitors likely splits both parties’ support (e.g. Downs, 1957). Furthermore, voters perceive coalition parties’ positions as less distinct than their platforms would indicate (Adams et al., 2015; Fortunato and Stevenson, 2013). Even ideologically distant coalition partners’ statements can legitimate challenger parties’ positions (Meguid, 2005, 2008). Candidates regularly pursue issues important to their electoral competitors upon entering office to avoid appearing weak on those policies (Sulkin, 2005). Broadly, perceptions of opposition parties’ competencies depend on evaluations of the government (Green and Jennings, 2012).

Upon entering government, the incentives for parties to coalesce increase dramatically. Multi-party parliamentary systems motivate parties to negotiate lasting bargains on a range of issues to create stable governments when there is no party with a parliamentary majority (Laver and Shepsle, 1996; Müller and Strøm, 1999). Minority governments necessitate coordination and cooperation with parties in parliament (e.g. Huber, 1996). Coalitions incapable of supporting policy compromises lead governments to end prematurely (Martin and Vanberg, 2011; Warwick, 1994).

While research has shown clear evidence of party coordination at the electoral and government levels, few studies theorize on how parties’ preferences relate. In contexts where parties regularly work together, it is unclear how PECs effect each component party’s policy positions. Parties motivated to control government join PECs to reduce conflict for the same voters, reducing electoral inefficiencies, and thus increase the likelihood that the PEC will be able to form a government following the election. Further, parties might join PECs to demonstrate to voters their preferred governing coalition partners (Golder, 2006; Ibenskas, 2015). This research is unclear, however, if parties’ electoral coordination through PECs leads parties to coordinate their policy positions.

Divided preferences and electoral coalitions

Parties’ revealed preferences through electoral campaigns represent their best effort to maximize votes while also representing internal groups (e.g. Harmel and Janda 1994). Past research on PECs shows that parties join them when they are electorally beneficial and increase the likelihood of forming a government post-election (Golder, 2006).¹ Parties are likely to join PECs when the benefits are relatively high and costs low; for example, when ideological compromise is relatively easy (Ibenskas, 2015). Presumably, in complex electoral environments, PECs reduce uncertainty about the post-election likelihood of governing coalitions (Tillman, 2015). Given PEC’s electoral foundations and motivations, parties’ likely evaluate success based on their electoral consequences. Successful PECs might even result in party mergers (e.g. Ibenskas, 2015) such as the eventual merger of the two largest Conservative parties in France as the Union pour un Mouvement Populaire from the Rassemblement Pour le République and the Union pour la Démocratie Française. In other extreme cases, sister parties, such as the German Christian Democratic Party and the Christian Social Union, manage real policy disagreements based on differences in their geographical constituencies, yet form PECs as a rule.

Not all PECs result in such success or lasting relationships. Following electoral agreements for the 1997 elections in France, for example, Les Verts and the Parti Socialiste decided to go separate ways in the proceeding election. For these parties, elections with PECs likely mark a high tide in their political relations followed by the parties distancing themselves.

Research on PECs contrasts studies on parties’ positioning where parties that appear too ideologically close likely suffer electorally. Voters often perceive coalition parties’ preferences as closer to each other than their platforms would predict (Adams et al., 2015; Fortunato and Stevenson, 2013). Indeed, government parties must emphasize issues more strongly for voters to perceive them (Bawn and Somer-Topcu, 2012; Greene, 2015).

In turn, party leaders seek to win elections by attracting distinct constituencies. These may be geographically separate as is the case for the CDU and the CSU in Germany but also likely reflect slight differences in preferences or issue priorities. Appearing too close would limit the effectiveness of parties’ election campaigns. The logic follows that party leaders normally seek to draw clear lines between their preferences. We argue, however, that the conditions leading to PECs also introduce incentives to blur these lines.

Hypotheses

We argue that parties change their policy positions in response to coalition experiences. While governing coalitions are punished electorally for not sufficiently distinguishing their preferences (Adams et al., 2015; Bawn and Somer-Topcu, 2012; Fortunato and Stevenson, 2013; Greene, 2015), PECs likely benefit from reducing ideological differences as these parties compete for ideologically similar voters. PECs reduce the number of ideologically similar competitors to increase the likelihood that those competitors can get into office and form a government. This logic implies that the very success of the PEC depends on an image of ideological cohesion. Voters ideologically closer to one of the parties (such as Les Verts in France) might take issue with a PEC including a party (such as the Parti Socialiste) that is farther from their preferred policies, unless the competitors reduce their differences. PECs also create the opportunity for the composite parties to foreshadow their combined governing aspirations, reducing the uncertainty of the compromises a governing coalition would make. Both of these logics would indicate that parties would shift their preferences closer together as they enter a PEC. Therefore, we predict that parties entering a PEC likely move their manifesto positions towards each other in our first hypothesis.
H1: Parties engaged in a PEC will decrease the distance between their manifesto positions.

We propose that parties in PECs are also responsive to the coalition’s success. Parties that lost votes will see little reason to maintain their support for continued compromise. Distancing their position from that of their past partners likely offers the parties multiple advantages. For parties remaining in the coalition, greater distance likely expands the coalition’s ideological breadth. Instead, parties leaving a PEC can use the opportunity to distinguish their independent identities. The failure of the past election likely demonstrates that the coalition’s position was not viewed as a credible government.

This logic leads us to expect that parties shift positions in response to the electoral success of the PEC. PECs that increased their seat shares demonstrate that the coalition served its purpose. An electorally viable PEC reduces wasted votes for the two parties through coordination, essentially, mobilizing the inherent electoral system bias to the two parties’ mutual benefit (e.g. Golder, 2006; Ibenskas, 2015; Tillman, 2015). Perceiving their future electoral success as connected, these parties will shift their positions even closer, to signal continued cooperation.

Less successful PECs face opposing incentives. PECs losing electoral support have failed to coordinate and mobilize the electoral system’s large party bias in their favour. Parties in these PECs view future costs of collaboration as greater than the electoral benefits. This logic leads us to expect that parties in PECs that lose support in the last election will create distance between their positions to demonstrate distinct, credible identities. Altogether, we predict an inverse relationship between the past electoral success of a PEC and the parties’ positions. Increased support for the PEC leads the parties to decrease their ideological differences, whereas decreased support leads parties to increase the distance between their stated preferences.
H2: Parties engaged in an electorally (un)successful PEC will decrease (increase) the distance between their manifesto positions

In summary, we explain the relationship between electoral coordination and party preferences by considering the motives for entering a PEC. Parties indicate increased coordination by decreasing the distance between their preferences when they join a PEC. Parties’ positions in consequent elections, however, depend on the electoral fortune of the PEC. Does the PEC mobilize the electoral system bias in the coalition’s favour? PECs benefiting from coordination decrease differences, whereas those coalitions failing to decrease electoral inefficiencies seek to distinguish their preferences. We next detail our approach to testing the hypotheses using a cross-national test and a focused analysis of a difficult case.

Data and methods

We pursue two analyses to empirically examine our hypotheses. We first perform a large-N cross-national test of the hypotheses using data from the CMP and Golder’s (2006) PEC data set. We then examine the internal validity of our analysis by examining a difficult or extreme case (e.g. George and Bennett, 2005): the German Christian Democratic Union and the Christian Social Union. This research design allows us to be confident that our hypotheses are correct in that they not only hold for PECs in a broad range of settings, but also that the inferences explain the relationship between parties engaged in long-term electoral and governing commitments.²

Our hypotheses focus on the relationship between parties. Therefore, we construct a dyadic data set including every observation of a party in an election directed towards every other party in that election. This structure allows us to directly test our hypotheses using characteristics of the parties to predict the relationship between them.³ Following Lowe et al. (2011), we then measure ideological distance using the absolute difference between each party combination’s logged left–right score based on the CMP’s RILE scale (Budge, 2001; Klingemann et al., 2006; Volkens et al., 2011). The dependent variable therefore ranges from 0 to just over 28 in our final sample.

We operationalize our primary independent variable, participation in a PEC, using Golder’s (2006) data. The measure is equal to one if the parties are jointly engaged in a PEC. Our analysis includes 139 dyads that form PECs.⁴

We examine our second hypothesis, using parties’ change in the percentage seat share in the last election.⁵ We create a measure of the percentage seats controlled by the dyad and find the difference from the election at time, t−2, and time, t−1, to predict the manifesto distance at time, t. We include an interaction of change in seat share with whether the party was in a PEC in the last election to analyse our second hypothesis.

We include control variables that likely moderate this relationship. In particular, we measure whether the parties were members of a governing coalition prior to the election using Golder’s (2006) coalition data set. Following Fortunato and Stevenson’s (2013) findings, we expect that parties in a governing coalition avoid appearing too ideological similar by increasing ideological distance. Furthermore, following the logic of Van der Velden and Schumacher (2015), we interact the main variables with a dummy variable indicating if the parties were in a governing coalition together. Presumably, the effect of PEC participation will be stronger in a governing context. We also account for the relative difference in seats controlled by each party; smaller parties likely distinguish their positions more clearly to avoid being subsumed by their larger partners. Fixed effects for the party and election year in each country account for additional unobserved heterogeneity, but limit our ability to include additional variables at these levels.

Following the primary analysis, we extend the logic of the nested analysis to study an extreme case. Previous work on PECs (e.g. Golder, 2006; Tillman, 2015) removed cases of extreme, perpetual cooperation from analyses arguing that these parties should be seen as a single entity rather than as separate organizations. Existing measures of party positions using manifestos (e.g. CMP), experts (e.g. Chapel Hill Expert Survey) or public perceptions (e.g. The Comparative Study of Electoral Systems [CSES]) also do not easily distinguish between parties with long-standing electoral agreements. If these past approaches are correct, then we should expect to see very little consistent variation in the differences between the parties’ positions. The regular treatment of parties such as the CDU and CSU in popular datasets and analyses demonstrates the general acceptance by researchers that there should be little difference between their positions or that these differences do not hold implications for behaviour in government. Analysis of these parties’ positions would then offer a ‘most difficult’ or ‘most unlikely’ case for the theory and allow us to further examine the logic of our hypotheses using more detailed data than available at the cross-national level (e.g. Gerring, 2004; George and Bennett, 2005). We therefore evaluate if the results from the primary analysis hold for parties with more extensive cooperation through a detailed case analysis of the German CDU and CSU using unique data on party positions derived from party leader speeches at parties’ national meetings. We first begin with a cross-national analysis of party preferences and PECs.

Party positions, PECs and election results

To test our primary hypotheses in the cross-national sample, we use ordinary least squared regression with fixed effects and a lagged dependent variable to account for the dyadic data structure and the recurrence of dyads in multiple elections. We present the results of this analysis in Table A2 with multiple model specifications.

Our first hypothesis predicts that parties shift their preferences closer together when they join a PEC. As parties increase their coordination to reduce electoral inefficiencies and to put forward a common platform, they likely perceive benefits from decreasing their ideological distance. The results from Table A2 contradict this perspective. In particular, the coefficients for being a PEC prior to an election are positive in each of the models; however, the coefficients are far from statistically significant. The results suggest that electoral coordination does not concurrently lead to decreases in ideological differences between parties. The negative coefficient for the lagged dependent variable suggests that historically close parties might instead be more likely to join PECs.

In contrast to the results for ongoing PECs, Table A2 presents greater support for our second hypothesis. This hypothesis predicts that parties engaged in a PEC in the last election decrease their ideological distance when the coalition performed well but increase their distance when the PEC performed poorly. Consistent with our second hypothesis, the coefficient for the interaction of being in a PEC in the last election and the change in seat shares in the last election are negative and statistically significant at the 90% level in the main PEC model and model including an interaction of cabinet participation and change in seat share. The full results offer further evidence of the theory, although suggesting the relationship is somewhat more complex. The three-way interaction of PEC participation, change in seat share and cabinet membership is negative and statistically significant at the 95% confidence level. This result indicates that PECs in a government coalition that gained seats decrease their distance whereas those that lost seats become more distant.

We present the effect of change in seat share for PECs based in Figure 1. In particular, we plot the predicted change in ideological distance (holding the control variables at their means and dichotomous variables at zero) over the range of change in parliamentary seat share. The first graph in Figure 1 demonstrates the negative effect of change in seat share conditional on being in a PEC based on the main PEC model whereas the second graph shows the conditional effect of change in seat share for PECs in government from the results in the full model. As hypothesized, a negative change in the percentage seats leads parties to increase their distance for both sets of models, whereas a positive change leads to reduced ideological conflict.⁶ Indeed, the effect is substantively meaningful, as a change from one standard deviation below the mean level of seat share change to one standard deviation above the mean leads incumbent parties to increase their ideological distance by 1.47 (0.48, 2.48).⁷ This is a change of approximately 4% of the dependent variable’s range. Altogether, the results suggest a clear relationship between electoral performance and ideological distance.

Figure 1.
Predicted effect of change in seat share. The figure presents the median predicted effect of increasing the change in percentage seat shares for parties engaged in a PEC in the last election over the observed range of seat share change for parties in PECs based on the main and full models in Table A2. We present 90% confidence intervals for the main model and 95% confidence intervals for the full model. The confidence intervals are from 1000 draws of the variance-covariance matrix. Dashes at the bottom present the observed seat share change for opposition (smaller marks) and incumbent (larger, darker marks) PECs. PEC: pre-electoral coalition.

The results for the control variables in Table A2 also largely fit with our expectations. In particular, parties in governing coalitions slightly decrease their distance, although the effect is not statistically different from zero in the simple cabinet model. In coalitions with greater variance in votes between parties in the last election, parties increase their distance, presumably to stake out more distinct positions.

The results from our primary analysis indicate clear support for the contention that PECs respond to their past electoral success, particularly in government. These results suggest that the logic holds across a range of political contexts. But do these results hold for parties, which are so closely linked that they often form joint electoral manifestos? In the next section, we turn to the case of the CDU-CSU. We follow recent studies (Aylott and Bolin, 2017; Ennser-Jedenastik and Muller, 2015; Somer-Topcu, 2017) and focus particularly on the positions of the parties’ leadership in relation to the general party membership.

A brief history of the relationship between the CDU and CSU

The story of the CDU and CSU coalition began at the conclusion of the Second World War. Following the collapse of the Nazi Regime in May 1945, multiple inter-confessional Christian conservative parties independently formed across the occupied zones. The parties largely consisted of former members of the Centrist Party and conservative parties active during the Weimar Republic. By 1950, regional associations in Bavaria coalesced to form the CSU, while Christian parties in the rest of Germany united to form the CDU. Since then, the CSU has operated only within Bavaria, and the CDU has operated in all other states.

The collaboration between both unions dates back to the first national election in the Federal Republic of Germany in 1949. Following the election, members of parliament of the CDU and CSU joined forces to create a permanent parliamentary party. The CDU/CSU union was the largest faction of the first German Bundestag and formed a coalition with the liberal Free Democratic Party (FDP) and the national conservative German Party (DP). Since then, no Chancellor has ever come from the CSU. The two CSU candidates who ran for Chancellor, Franz Josef Strauß and Edmund Stoiber, were both defeated by the SPD in 1980 and 2002.

Underlying the joint governing experience, the CDU and CSU have a long history of collaboration. They share a common youth organization and run a joint manifesto during federal elections. Yet, their relationship has not always been amicable. In 1976, for example, the parties nearly ended their parliamentary union, which is renewed after every federal election. Although the CDU-CSU won the election, they failed to replace the Social-Liberal coalition between the SPD and the FDP. Following the defeat, the CSU Bundestag faction decided to discontinue the agreement. The party eventually recalled the decision after the CDU made some concessions to the CSU and threatened to campaign in Bavaria.

At several occasions, the CSU has also tried to increase its influence outside Bavaria. After the fall of the Berlin Wall in 1989, the CSU supported the newly formed German Social Union(DSU) with financial resources and political know-how to establish a CSU representation in the states of the former GDR. The CSU’s national congress in Leipzig in January 1990 was also the first time that a Western German party has held a big party meeting in Eastern Germany. In June 1990, 200 supporters of the CSU founded a regional association of the CSU in Saxony.

Despite repeated tensions, observers describe the CDU and CSU as sister parties. As organizations, they function as entirely independent parties below the federal level with separate leaderships, decision-making bodies and distinct policy goals. As such, they frequently disagree on policy in their public statements and during electoral campaigns.

Notwithstanding their assumed coherence, the parliamentary factions also act with disunity. Figure 2 shows the level of discipline within the CDU-CSU faction when rolls were called during the 16th and 17th legislature of the German Bundestag. The black line shows the change of factional discipline over time, while the dashed and solid lines show the mean level of agreement and the break between the two legislative periods. The graph illustrates that the level of unity within the CDU-CSU varies over time and was lower, and generally much more volatile, during the grand coalition with the SPD from 2005 to 2009 than during the liberal-conservative government with the FDP from 2009 onwards. This disagreement likely follows from the SPD’s focus on policies highlighting distinctions between the parties.

Figure 2.
Factional Discipline within the CDU/CSU faction. The level of cohesion of the CDU-CSU faction is based on an analysis of all rolls called in the 16th and 17th German Bundestag. The values on the y-axis show the percentage of Members of Parliament (MPs) that voted with the faction majority.

Differences evidence themselves not only at the abstract governmental level. The public also observes disunity. Figure 3 shows public perceptions of divisions between the two parties over time using annual measures from the Politbarometer survey. Since Angela Merkel took over the leadership of the CDU in 2000, voter perceptions of disagreement between the CDU and the CSU increased by nearly 30%. By 2005, two-third of the respondents thought that the Union was internally divided. This number decreased when the two parties entered government in October 2005 only to increase again in the run up to the 2009 election.

Figure 3.
Perceived level of internal divisions within the CDU/CSU. The data presented in figure are based on a survey question that asks respondents to indicate if they think that the CDU and CSU are rather divided or rather united on key political issues.

The history of collaboration and confrontation between the CDU and CSU suggests that their shared government performance occurs despite organizational differences. As these examples further illustrate, divisions between the CDU and CSU are frequent, although scholars treat them as a single organizations in a range of analyses. In the next section, we introduce a new data set that we use to examine the parties’ previously unmeasured preferences.

Using party congress speeches to extract party positions

To further explore the relationship between the CDU and CSU, we collect a new data set of party leader and member speeches over a 20-year period that allows us to estimate differences in the parties’ positions in the absence of distinct party manifestos. In particular, we estimate the parties’ revealed preferences at these meetings by analysing speeches given at the parties’ national conferences between 1990 and 2011. Party congresses in these parties are usually held once a year and offer delegates a relatively unconstrained platform to voice their opinions.⁸

Following recent advancements in automated text analysis, we use WORDFISH (Slapin and Proksch, 2008) to retrieve the relative position of speeches on the primary dimension of conflict. The WORDFISH algorithm uses the frequency of word usage to estimate word and document coefficients according to a Poisson distribution. The model then uses these estimates to group documents according to their most common word usage.

We estimate one model for all 1268 speeches to calculate the actors’ preferences at each meeting on a principle left–right dimension. Based on this model, Figure 4 plots the yearly median position of the CDU (black square) and the CSU (grey square), the point estimates of the respective leaders from both parties, and a mean position across all congresses (dashed line).⁹ At this point, CSU speeches following 1999 are publically unavailable. As an alternate source of evidence, however, the CDU frequently invites CSU leaders to speak at their congresses. We use the CSU leaders’ position at these meetings as a proxy for the party median following 1999.

Figure 4.
Relative party Congress and party leader positions, 1990–2011. The figure shows the change in positions of the CDU and CSU party congresses (black and grey squares) and their respective leaders (black and grey stars with names) over time. The position estimates are generated from party member speeches given at the parties’ national congresses using the scaling technique Wordfish. We pool all available speeches into a single text matrix and thereby assume that we estimate the actors’ positions on a general left–right scale. The CSU party congresses from 2000 onwards are generated from speeches given by the CSU leader at the CDU conventions.

Public opinion and the distance between party leaders

Our discussion implies differences between the CDU and the CSU, despite continued electoral collaboration. Our second hypothesis predicts that parties will move apart following electoral loss and together in response to electoral gains. This prediction contrasts one in which each party reacts in tandem to some external factor such as public opinion in the ‘riding the wave’ hypothesis (e.g. Spoon and Klüver, 2014).

We present evidence to support our approach in Figure 4. Figure 4 illustrates that both parties and their leaders have changed their positions based on their speeches at parties’ national meetings over time. Most noticeably, the plot exhibits a clear leftward trend in the position estimates, especially for the median position of the CSU congress and the party leaders’ position and following a change in leadership (Somer-Topcu, 2017). The graph also shows that the party leaders’ positions frequently diverge from the party median (consistent with Ceron, 2013; Greene and Haber, 2014; Kölln and Polk, 2017).

These results suggest that the parties and their leaders have distinct positions which vary over time. An intriguing development is that shifts in the CSU party leader position at time t+1 often shift in opposition to shifts from the CDU at time t. The CSU party leader’s position in 1995, 1997 and 1999, for example, shifts more to the right from their past position following shifts towards the left from the CDU party leader in previous years. More broadly, the CDU party leadership positions stake a relatively consistent leftwards trend, whereas the CSU party leaders demonstrate a slightly more varied pattern, perhaps reflecting the relative size disparity.

To determine whether the positions of both parties change in tandem, we estimated the change in position for party leaders over time. Figure 5 shows the change in positions of the CDU (solid line) and CSU (dashed line) leaders from 1990 to 2011. The shaded areas indicate the government status, while the dashed lines represent general elections.

Figure 5.
Mean positions of party leaders, 1990–2011. The positions of the party leaders are the same Wordfish estimates introduced in Figure 4.

The two time series show a number of interesting patterns. The difference between party leader positions of both parties decreased over time. This might indicate that the CSU/CSU have become more ideologically similar in government. Second, as our hypothesis might predict, the time series suggests that both parties changed positions in opposing directions prior to the general election in 1998. From their time in opposition, however, until the end of their coalition with the FDP in 2009, the CDU and CSU changed positions more in tandem.

Intriguingly, the leaders’ positions have more recently diverged. These differences are driven largely by Merkel’s move towards the left. This might imply that the parties could be in for a leadership dispute (unsupported by the membership’s positions).

Without extensive time series analysis, it is difficult to draw conclusions about the differences in terms of the size and the direction of the two parties’ changes in positions. Nevertheless, we can test what effect changes in public opinion have on the distance between the party leaders with the data available. We measure public opinion using the monthly voter polls from the Politbarometer data set. The data are based on a public opinion survey that asks respondents which party they would vote for if there were an election on Sunday. Figure 6 shows the relationship between polls, aggregated to yearly averages, and the distance between the party leaders between 1991 and 2011. We predict the effect of public opinion on distance using a simple, bivariate linear regression model with party leader distance as the dependent and the poll from the year prior to a party congress as the independent variable. Consistent with our second hypothesis and our cross-national analyses, the distance between the party leaders decreases when their polls in the previous year are higher. The result is significant at the 95% level but should be interpreted with caution due to our small sample.¹⁰

Figure 6.
Expected vote share and distance between party leaders.

Conclusions

We have argued that the relationship between party cooperation and party positions is more complex than traditional descriptions take into account. Using evidence on PECs and party positions as well as a case study of two historically close parties, the German CDU and CSU, we find that the electoral success of a PEC predicts the parties’ future difference in preferences. The logic of electoral competition suggests ideologically close parties distinguish their preferences in the face of electoral losses, but they shift closer when the PEC has increased its support. These results imply that parties engaged in PECs act more consistent with traditional theories of party positioning when they face electoral losses (e.g. Adams, 1999; Adams and Somer-Topcu, 2009; Spoon, 2011). Like van der Velden and Schumacher (2015), our results indicate that coalitions use their electoral success to determine their future positions. Unlike other forms of governing coalitions though (e.g. Fortunato and Stevenson, 2013), PECs do not further distinguish their preferences when they increase their votes in the last election. Analysis of party manifestos from a broad cross-sectional time series and speeches from party national congresses in Germany support our hypotheses.

This evidence invokes a number of directions for future research on PECs. Indeed, it is possible that the conditions leading to the past electoral success of PECs also relate to PECs’ ideological proximity. We account for parties’ past positions and success, but the relationship is likely more nuanced. Further, competing logics suggest that the relative ideological closeness of PECs might influence their future electoral success as more diverse PECs attract a broader range of support (e.g. Somer-Topcu, 2015) or unified PECs indicate a more responsible and negotiated governing coalition.¹¹

The results from this analysis hold clear implications for the study of party politics. Researchers often treat sister parties as if they are a single party. Common datasets, such as the CMP only include a single manifesto for the German CDU and CSU in most elections. Our analysis suggests that studies not fully taking account of the complex relationship between these two organizationally distinct, but regular collaborators might arrive at unexpected results. Furthermore, PECs likely influence how party leaders manage intra-party factions and likely introduce the incentives for greater preference incongruence between party leaders and members (see also Ceron, 2017; Greene and Haber, 2014; Kölln and Polk, 2017; Somer-Topcu, 2017).

The relationships between parties likely hold important consequences for political representation. Classic formulations of the linkages between citizens and party government require that citizens have clear parties to choose between. PECs and sister parties confuse that linkage by letting parties appear ideologically closer to each other but also reduce the choices available to increase the chance that either they will control government. Evidence of parties’ efforts with slightly varying preferences to collectively manage their separate constituencies indicates that the representation process is not only more complex than often envisioned, but also still responsive to various constituency demands.

Supplemental material

Supplemental Material, pp-2016-0100-file002 - Maintaining Partisan ties

Supplemental Material, pp-2016-0100-file002.pdf for Maintaining Partisan ties by Jonathan Polk, Ann-Kristin Kölln, Zoe Greene and Matthias Haber in Party Politics

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: This work was supported by the Collaborative Research Center 884 Political Economy of Reforms (Project C2) at the University of Mannheim, funded by the German Research Foundation.

Supplemental Material

The online appendices are available at .

Notes

References

Aylott

Bolin

(2017) Managed intra-party democracy: precursory delegation and party leader selection. Party Politics 23(1): 55–65.

Adams

Somer-Topcu

(2009) Moderate now, win votes later: the electoral consequences of parties’ policy shifts in 25 postwar democracies. The Journal of Politics 71(2): 678–692.

Adams

Ezrow

Wlezien

(2015) The company you keep: how voters infer party positions on european integration from governing coalition arrangements. American Journal of Political Science. Epub ahead of print 25 November 2015. DOI: 10.1111/ajps.12231

Adams

(1999) Policy divergence in multicandidate probabilistic spatial voting. Public Choice 100(1-2): 103–122.

Bawn

(1999) Voter responses to electoral complexity: ticket splitting, rational voters and representation in the Federal Republic of Germany. British Journal of Political Science 29(3): 487–505.

Bawn

Somer-Topcu

(2012) Government versus opposition at the polls: how governing status affects the impact of policy positions. American Journal of Political Science 56 (2): 433–446.

Blais

Indridason

(2007) Making candidates count: the logic of electoral alliances in two-round legislative elections. Journal of Politics 69(1): 193–205.

Budge

(2001) Mapping Policy Preferences: Estimates for Parties, Electors, and Governments, 1945-1998. Oxford: Oxford University Press.

Ceron

(2013) Brave rebels stay home: assessing the effect of intra-party ideological heterogeneity and party whip on roll-call votes. Party Politics 24: 1–13.

10.

Ceron

(2017) Maintaining Partisan ties. Party Politics 23(1): 7–17.

11.

Downs

(1957) An economic theory of democracy. New York: Harper.

12.

Ennser-Jedenastik

Müller

(2015) Intra-party democracy, political performance and the survival of party leaders Austria, 1945–2011. Party Politics 21(6): 930–943.

13.

Ezrow

De Vries

Steenbergen

. (2011) Mean voter representation and partisan constituency representation: do parties respond to the mean voter position or to their supporters? Party Politics 17(3): 275–301.

14.

Fortunato

Stevenson

(2013) Perceptions of partisan ideologies: the effect of coalition participation. American Journal of Political Science 57(2): 459–477.

15.

George

Bennett

(2005) Case Studies and Theory Development in the Social Sciences. Cambridge: MIT Press.

16.

Gerring

(2004) What is a case study and what is it good for? American Political Science Review 98(2): 341–354.

17.

Golder

(2006) Pre-electoral coalition formation in parliamentary democracies. British Journal of Political Science 36(2): 193–212.

18.

Green

Jennings

(2012) The dynamics of issue competence and vote for parties in and out of power: an analysis of valence in Britain, 1979-1997. European Journal of Political Research 51(4): 469–503.

19.

Greene

(2015) Competing on the issues: how experience in government and economic conditions influence the scope of parties’ policy messages. Party Politics 1–14. DOI: 10.1177/1354068814567026

20.

Greene

Haber

(2014) Leadership competition and disagreement at party national congresses. British Journal of Political Science FirstView 1–22. Available at: https://doi.org/10.1017/S0007123414000283.

21.

Greene

Jensen

(2016) Manifestos, salience and junior ministerial appointments. Party Politics 22(3): 382–392.

22.

Gschwend

(2007) Ticket-splitting and strategic voting under mixed electoral rules: evidence from Germany. European Journal of Political Research 46(1): 1–23.

23.

Harmel

Kenneth

(1994) An integrated theory of party goals and party change. Journal of theoretical politics 6(3): 259–287.

24.

Huber

(1996) Rationalizing Parliament: Legislative Institutions and Party Politics in France. Cambridge: Cambridge University Press.

25.

Ibenskas

(2015) Understanding pre-electoral coalitions in Central and Eastern Europe. British Journal of Political Science FirstView: 1–19 Available at: https://doi.org/10.1017/S0007123414000544.

26.

Ibenskas

Sikk

(2017) Patterns of party change in Central and Eastern Europe, 1990-2015. Party Politics 23(1): 43–54.

27.

Kölln

A-K

Polk

(2017) Emancipated party members: Examining ideological incongruence within political parties. Party Politics 23(1): 18–29.

28.

Klingemann

H-D

Volkens

Bara

. (2006) 2 Mapping Policy Preferences II: Estimates for Parties, Electors, and Governments in Eastern Europe, European Union, and OECD 1990-2003. Oxford: Oxford University Press.

29.

Laver

Shepsle

(1996) Making and Breaking Governments: Cabinets and Legislatures in Parliamentary Democracies. New York: Cambridge University Press.

30.

Lieberman

(2005) Nested analysis as a mixed-method strategy for comparative research. American Political Science Review 99(03): 435–452.

31.

Lowe

Benoit

Mikhaylov

. (2011) Scaling policy preferences from coded political texts. Legislative Studies Quarterly 36(1): 123–155.

32.

Martin

Vanberg

(2011) Parliaments and Coalitions: The Role of Legislative Institutions in Multiparty Governance. Oxford: Oxford University Press.

33.

Meguid

(2005) Competition between unequals: the role of mainstream party strategy in niche party success. American Political Science Review 99(3): 347–359.

34.

Meguid

(2008) Party Competition Between Unequals: Strategies and Electoral Fortunes in Western Europe. Cambridge: Cambridge University Press.

35.

Müller

Strøm

(1999) Policy, Office, or Votes?: How Political Parties in Western Europe Make Hard Decisions. Cambridge: Cambridge University Press.

36.

Powell

Jr Whitten

(1993) A cross-national analysis of economic voting: taking account of the political context. American Journal of Political Science 37: 391–414.

37.

Slapin

Proksch

S-O

(2008) A scaling model for estimating time-series party positions from texts. American Journal of Political Science 52(3): 705–722.

38.

Somer-Topcu

. (2015). Everything to Everyone: The Electoral Consequences of the Broad-Appeal Strategy in Europe. American Journal of Political Science 59(4): 841–854.

39.

Somer-Topcu

(2017) Agree or disagree: How do party leader changes affect the distribution of voters’ perceptions. Party Politics 23(1): 66–75.

40.

Spoon

J-J

Klüver

(2014) Do parties respond? How electoral context influences party responsiveness. Electoral Studies 35: 48–60.

41.

Spoon

J-J

(2011) Political Survival of Small Parties in Europe. Ann Arbor: University of Michigan Press.

42.

Sulkin

(2005) Issue Politics in Congress. New York: Cambridge University Press.

43.

Tavits

(2007) Principle vs. pragmatism: policy shifts and political competition. American Journal of Political Science 51(1): 151–165.

44.

Tillman

Erik R

. (2015). Pre-electoral coalitions and voter turnout. Party Politics 21(5): 726–737.

45.

Van der Velden

Schumacher

(2015) Do coalition partners drift apart or stick together? An analysis of party platform changes in 11 Western European Countries. Presented at the Annual Conference of the Midwest Political Science Association (April 2015): Chicago, IL.

46.

Volkens

Lacewell

Regel

. (2011) The manifesto data collection. Manifesto Project (MRG/CMP/MARPOR.). Berlin: Wissenschaftszentrum Berlin für Sozialforschung (WZB).

47.

Warwick

(1994) Government Survival in Parliamentary Democracies. Cambridge: Cambridge University Press.

48.

Whitten

Palmer

(1999) Cross-national analyses of economic voting. Electoral Studies 18(1): 49–67.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB