Party Representation and the Organization of Eastern European Parliaments

Transcription

1 Party Representation and the Organization of Eastern European Parliaments Royce Carroll Rice University Monika Nalepa University of Notre Dame A Research Note Prepared for Whither Eastern Europe? Changing Political Science Perspectives on the Region. University of Florida, January 9-11, Introduction Upon winning seats in parliament, political parties face constraints in pursuing the policy agenda they advocate in their electoral campaigns. Parties in government may not be able to pursue legislation preferred by their supporters as part of their coalition compromise. Parties in the opposition are often faced with a legislative agenda they do not control and may find themselves allying with parties on the floor in a manner at odds with their ideological profiles. Mediating how the process of party representation unfolds is legislative organization, which can constrain the voting behavior of members and parties in at least two ways. First, parliaments constrain parties by regulating the opportunities to vote on particular issues i.e. agenda control (Cox and McCubbins, 2011; Döring, 2001); and, second, government membership constrains parties by raising the costs of coalition conflict on the floor (Huber, 1996; Diermeier and Feddersen, 1998). 1 In this paper, we focus on how the legislative arena structures the relationships among parliamentary parties. Specifically we examine how their ideological profiles are reflected in parliamentary voting and the possible constraints on that behavior. To understand this question, we use two sources of data. The first source are roll-call data from four post-communist legislatures: the Czech Republic, Poland, Romania, and Slovakia. We use multidimensional scaling to locate legislators and parties in the policy space based on their observed voting behavior. Second, we use data from the Chapel Hill Expert Surveys (CHES) (Bakker et al., 2012), where respondents were asked to place parties on a series of 10-point scales corresponding to potential cleavages in society. These dimensions are Economics and what we will refer to as Post-Materialism versus Traditionalism. 2 Figure 1 below shows these party positions across the three waves of surveys included included each of these countries for each of these policy areas. 1 Another important mediating factor in legislative voting is how intraparty organization and party recruitment enables party unity, which we examine elsewhere (Carroll and Nalepa, 2012). 2 The authors of the CHES describe this dimensions as GAL-TAN because it captures the green/alternative/libertarian pole versus the traditionalism/authority/nationalism pole (Bakker et al., 2012) 1

2 Figure 1: Party Positions on CHES Economic Left/Right and Post-Material/Traditional Dimensions: Czech Republic, Poland, Slovakia, and Romania Post material/traditional LPR CSSD S PSL UP PRM AWSP PiS SLD PSD KDU CSL ODS US DEU UW PUR CDR 2000 PD PNL UDMR PSNS KDH HZDS HZD SMK Smer SDKU ANO PO LPR PiS S KDU CSL CSSD Nezavisl PRM PSL PSD SZ SDPL SLD PC Smer SNK ED UDMR FDGR SNS ODS PO PD PD PNL KDH LS HZDS SMK SDKU SF CSSD PRM KDU CSL LPR PiS S SZ SDPL SLD PSL PC PSD SNS Smer VV ODS TOP0 PO SD UDMR PD L LS HZDS SMK PNL KDH Most SDKU DS SF SaS cz pol rom slo Left Right Economics 2

3 Data from Eastern Europe is well suited for studying legislative party systems because, compared to parliamentary regimes with longer periods of continuous democratic experience, there are several reasons for discrepancies between partisan electoral competition and conflict in legislative policy. In part, this is due to the fact that Eastern European party systems are less institutionalized than their western counterparts (Grzymala-Busse, 2002a; Moser, 1999). Furthermore, these systems have inherited post-authoritarian legacies that shape elite conflict. For instance, even where successor-communist parties share policy preferences with successor dissident parties on a salient issue, such as redistribution, to take one example, they have avoided forming governing coalitions across the regime divide (Grzymala-Busse, 2001a). This is typical for countries like Poland or the Czech Republic, where this regime divide was particularly stark, but even in cases such as in Romania, where the emergence of such a coalition in 2008 continued to be costly (Stan and Zaharia, 2010). Characteristics of young party systems, such as these authoritarian legacies, may lead to parties that are in fact ideologically distant to cooperate in contradiction to some of their programmatic appeals. Potential problems between coalition partners in established parliamentary regimes are known to be mitigated by restricting the agenda for instance, using chamber leadership to shape an agenda that avoids issues counter to coalition interests which may reinforce or contradict the programmatic identities of partners. Legislatures in Eastern Europe present an opportunity to trace the institutionalization of these institutions using relatively recent records of behavior, particularly in the form of recorded legislative voting, an indicator of the joint products of party policy preferences, coalition discipline, and legislative organization. The purpose of this paper is to initiate an empirical investigation into how legislative institutions and parliamentary coalitions influence party behavior. Our topic of interest is the relationship between parliamentary floor voting and the reputations of parties. We take this relationship to be a reflection of the effects of legislative organization especially agenda setting in structuring the choices parties make and the way in which they represent their constituents. For current purposes, we sketch several ideas from the literature on parliamentary legislative organization on which we can gain leverage by studying Eastern European parliaments. We then describe the empirical patterns in cases of Poland, Czech Republic, Slovakia and Romania and provide some preliminary impressions of these parliamentary party systems and their relevance for legislative studies. 2 Party Reputations and Agenda Setting in Parliaments Despite becoming active for the first time during founding elections in the late eighties and early nineties, parties in Eastern Europe did not develop and organize simultaneously. In many cases the successor-communist parties emerged with organizational advantages over parties based on dissident movements. With better managerial resources, more technocratic skills and experience with intraparty elections, they were better prepared for the challenges associated with elections and democratic procedures. A vast literature has documented the causes and consequences of this variation in the organizational features of these parties, including their members experience as legislators, cabinet members, and party bureaucrats 3

4 (Grzyma la-busse, 2001b; Grzymala-Busse, 2002a; Ishiyama, 1999, 2001). In contrast, the dissident groups from which other parties emerged were formed in pre-transition conditions where weak organization (informal networks, and lack of transparency) was a necessity for evading infiltration (Stan and Zaharia, 2010). As a result, some communist successor parties did well in post-transition elections though often facing resistance from other parties in forming governments (Grzymala-Busse, 1998, 2002b, 2001a). Meanwhile, poorly organized dissident groups were also able to capitalize on presenting an alternative to communistsanctioned candidates, often at the cost of developing programmatic identities. Consider as an example the Polish ex-dissident party Election Action Solidarity (AWS). AWS was the Polish right s response to the Democratic Left Alliance (SLD), the coalition of successor communist parties. Prior to the 1997 election, the post-solidarity right had failed to unite in time for elections and ended up splitting the post-solidarity vote. In 1997, close to forty ex-dissident groups created the AWS, an electoral coalition that secured enough seats in the legislature to lead the first post-solidarity cabinet since 1993, when a government of four post-dissident parties lost a non-confidence vote. To unify, AWS had to temporarily suppress any programmatic divisions that could create conflict in the postdissident camp and unite around the main thing dissidents had in common: preventing the SLD from returning to office. While the AWS was successful at winning enough seats to create a cabinet, its organization perpetuated legislative divisions among groups with diverse policy interests on the basis of dissident roots. By the end of the electoral term, separate legislative caucuses proliferated among AWS own membership and it became impossible to suppress the programmatic differences among its members. The new caucuses became the prototypes of parties that started forming in anticipation of the upcoming elections among them Law and Justice (PiS) and Civic Platform (PO), the parties that later came to dominate the party system as with starkly opposing platforms. Issues such as the regime divide and other factors leading to weakly institutionalized party systems make the region a unique place to examine the simultaneous emergence of legislatures and patterns of party competition. The typical Western European left-right cleavage separating libertarian and traditionalist parties, on the one hand, from parties favoring social liberalism in conjunction with tight state regulation of the market and redistribution, on the other, is rare in post-communist Europe. Specifically the extent to which these tendencies play out depends on the behavior of the communist successor party in question and the specific type of the preceding communist regime (Kitschelt, 1995, 1999). Meanwhile Eastern Europe presents just as much variation in terms of legislative institutions as Western Europe (Zubek, 2011). Together, the region s parliaments allow us the opportunity to understand where legislative organization fits into the representation functions of the party system and how this changes. Much work has examined the relationship between parties and legislative organization (Cox and McCubbins, 2011; Cox, 2006), including in parliamentary regimes (Döring, 1995; Akirav, Cox and McCubbins, 2010; Chandler, Cox and McCubbins, 2006; Cox, Heller and McCubbins, 2008; Cox, Masuyama and McCubbins, 2000) in general and Eastern Europe in particular (Zubek, 2008, 2011). Yet we know considerably less about the general relationship between legislative organization and the dimensions of party competition in these countries. Existing literature argues that parties can take positions and generate a record of behavior that develops their policy reputations (Snyder and Ting, 2003; Snyder Jr and Ting, 2002; 4

5 Woon and Pope, 2008). At the same time, the visible positions of parties in parliament (based on their voting) are a function of both the political cleavages dividing parties and the organization of the legislature. 3 That is, when parties organize their chambers they constrain how party competition in elections translates into legislative activity. This in turn feeds back into party reputations. There is considerable literature on the US Congress linking together liberal and conservative ideology and the organization of the House and Senate. Most prominently, Poole and Rosenthal (1997) trace the phases of party system change through the roll call record. They argue that the main policy conflicts among parties (e.g. economic left and right, civil rights) can be represented as behavioral relationships among legislative parties. In this note, we take a similar approach to interpreting legislative voting behavior as representing spatial policy positions, comparing the revealed preferences of parliamentary parties with their policy reputations and examine the factors that mediate how party system cleavages manifest themselves in parliaments. Unlike in the U.S., where the institutional independence of Congress from the president allows it to function as distinct arena, parties in parliamentary regimes are inextricably linked to executive politics. Thus, governing status (that is, cabinet membership) has a direct impact on parliamentary voting (Laver, 2006). When parties form majority governing coalitions, they are especially likely to constrain behavior by avoiding the issues on which partners have systematic differences. As Chandler et al. note, in parliamentary regimes, notwithstanding any direct effects of cabinet membership, high levels of coalition discipline observed in roll call votes are as much a function of governmental agenda control specifically, the ability of the coalition to prevent bills that would split its members apart from being voted on the floor (Chandler, Cox and McCubbins, 2006). Compared to opposition parties, which are influenced by the agenda but usually do not directly control it, we would expect governing parties to vote together even on important dimensions of difference that is dimensions on which experts consider them in conflict. In such cases, roll call voting will deviate from parties policy positions. Naturally, parties choose their governing partners based on policy concerns. But even if a coalition pays off on one policy dimension for office-seeking parties, it nevertheless may result in them getting closter to their coalition partners on other dimensions in a manner that is at odds with the reputation they seek to reinforce. Still, in other instances, being in a government with strong control over the legislature may mean reinforcing their parties positions by forcing most conflict in the chamber to revolve around the matters already salient to the governmentopposition divide. 4 Two factors should be especially important here. First, majority coalitions can more easily control the agenda compared to minority governments. That is, if a cabinet is in place 3 The source of these party positions on political cleavages may vary but can be taken to reflect their electoral campaigns, which in turn are designed to represent divisions in society. For present purposes, we largely avoid the question of how party reputations are cultivated 4 Note that, because this visible behavior is a function of the agenda, voting records indicating that ideologically distant parties are spatially close (or vice versa) are not necessarily a result of parties actively contradicting their reputation by voting on legislation. The legislative agenda setting process can structure the legislature to distort the ability of parties to reinforce their reputations in parliament. As a result, some parties end up voting closer to parties we would expect them to diverge from and voting distant from apparent allies on salient dimensions of the political system. 5

6 that has rewarded parties comprising a majority of seats with cabinet portfolios, they should be able to organize the chamber for their mutual benefit and they should have the votes to execute that agreement on the floor. By contrast, minority governments tend to involve more ad hoc voting coalitions (Strom, 1990; Lijphart, 2012). Thus, while majority parliamentary coalitions can easily induce legislative behavior that either reinforces or contradicts party reputations, parties operating under minority governments should be better positioned to generate a record of parliamentary behavior that reflects their policy positions. This is because a broader range of items will be considered on the agenda, even those that the government would prefer to avoid, making the roll call record more likely to reflect party ideological conflict. Second, the ability of a governing majority to control the agenda should also be connected to the formal rules of the chamber. Generally, centralized legislative rules may have the effect of enhancing the power of the government to avoid conflict among governing parties even for minority governments. Decentralized legislative rules, meanwhile, may allow opposition parties to greater influence over the agenda even when facing majority governments. In sum, while party policy reputations exist exogenously to parliamentary behavior, elite conflict in parliament will often diverge from these expectations when coalitions organize the legislature. To begin an examination of where this occurs and why, the sections below present data visualizations constructed using legislative roll call votes in four Eastern European countries in conjunction with information on party reputations. 3 Poland For the Polish case we present the parliamentary voting patterns in most governments from As we will in each of our explorations, we make use of roll call data from a governing period and transform them into ideal point estimates for legislators estimated relational locations in space via the Optimal Classification algorithm (Poole, 2000). This method unfolds binary choice data to determine the underlying dimensionality of the behavior and locate legislators within that policy space. Here, the locations reflect the tendencies of parties to join in similar or different voting coalitions on the first (and dominant) dimension and a second (orthogonal) dimension of variance. When parties are closer together, this means that their behavior was more similar on that dimension. In order to compare the parliamentary behavior to party reputations, we label each party with a color indicating their location on the CHES policy dimensions of economic left-right (left panel) and postmaterialism-versus-traditionalism (right panel), linking each of the surveys (2002, 2006 and 2010) to the most proximate government period. 5 These data for Poland are displayed in Figure 2. Darker shading on the economic scale corresponds to the right and wanting the government to play a reduced role and support lowering taxes, privatization, and limiting welfare and redistribution. Lighter shading corresponds to parties who prefer the government to play and active role in the economy. On the second panel, darker shading corresponds to 5 Note that only parties with such survey data are depicted. This excludes small parties in most cases and all independents. In several instances, parliamentary parties are linked to parties in the survey data from after name changes mergers or splits. With the exception of Slovakia s 1998 government below, The CHES name is shown in all figures. 6

7 the most traditionalist, nationalist and authoritarian parties (i.e., valuing law and order and believing that government should take a firm stance on questions of morality), while lighter shading represents to the most post-materialist or socially liberal parties (e.g. favoring expanded access to abortion, same sex marriage, or euthanasia). The recent CHES data suggests that these two dimensions capture most substantial policy debates. In Poland, as the most salient cleavages reported in these surveys are civil liberties, cosmopolitanism/religion, and redistribution. By 1997, the Speaker of the Polish Sejm had discretion over the agenda, limiting the Sejm s collegial steering committee to an advisory role. Following this reform, any governing majority with control over the Speakership could largely restrict the legislative agenda to those items it preferred to see passed. These changes in legislative procedure were actually put in motion by the AWS itself at the beginning of the term and had the potential to strengthen the government-opposition divide. That AWS did not fully exploit this centralized legislative arrangement is apparent when comparing the Buzek I (majority) with the Buzek II (minority) cabinet above. A majority cabinet provided little advantage to unifying AWS and UW by avoiding matters of difference relative to the minority cabinet that came when UW left the cabinet entirely. Even as a majority, the AWS-led government did allow a strong second dimension influenced to some extent by both economics and traditionalism to emerge, each contributing to the fragmentation of the government. Under Buzek I, the Law and Justice (PiS) faction (the part of AWS that was to go and form PiS) had already started voting with PSL on a second dimension of legislative conflict one that tracks their agreement on the postmaterialist-versus-traditionalism cleavage. The government-opposition axis of conflict during the Buzek governments is best understood as a reflecting the crosscutting regime divide, which was likely reinforced independently of legislative institutions. The Buzek I cabinet, although enjoying a nominal majority, was likely too internally divided within its leading party to operate similarly to typical majorities in terms of constraining parliamentary behavior. 6 The Civic Platform (PO) and Law and Justice (PiS) remained apparent allies during their time as component circles within the AWS, as well as shortly after. The movement of PO and PiS into defining opposite sides of the first dimension of legislative conflict was a gradual process. In 2001 (under Buzek II and Miller I), just after breaking away from AWS, the PiS and PO members remained at least adjacent on both dimensions of voting patterns. According to Jacek Janiszewski, a former AWS MP, members of the two parties tried to form another overarching coalition of the post-solidarity right, based on the Conservative Peoples Party (SKL), but went their separate ways when this attempt failed. The coalition would have had to accommodate within one organization leadership styles as distinct at that of Maciej Plazynski, House Speaker, who was being groomed to serve as SKL s Chairman of the Political board and that of the Kaczynski brothers. Ideological dissonance was also present. The new libertarian PO platform influenced by Donald Tusk and Andrzej Olechowski was difficult to reconcile with the Kaczynski brothers social conservatism. 7 Nevertheless, in 6 Elsewhere we argue that the inability to efficiently control the agenda stemmed from the appointment of a Speaker that, while powerful, did not share the interests of the coalition as a whole. In subsequent terms, particularly the last government discussed here, the recruitment of the Speaker became more institutionalized as a means to extend government control over the parliament (Carroll and Nalepa, 2013). 7 In the past, they had frequently conflicted with the goals of the Gdansk-based libertarians, who were 7

8 Figure 2: Legislative Roll Call Positions and Expert Surveys on Economic Left-Right and Post- Material/Traditional Dimensions: Poland (Buzek I AWS,UW) SLD PSL AWS UW (Buzek II AWS) (Buzek I AWS,UW) SLD PSL AWS UW (Buzek II AWS) PSL SLD PiS AWS UW (Miller I SLD,PSL,UP) PSL SLD PiS AWS UW (Miller I SLD,PSL,UP) Dimension 2 Samoobrona LPR PSL SLD UP PiS PO (Miller II SLD,UP) PSL Samoobrona LPR SLD UP PiS PO (Marcinkiewicz I PiS) SLD PO PSL Samoobrona LPR PiS L R Economic Dimension 2 Samoobrona LPR PSL SLD UP PiS PO (Miller II SLD,UP) PSL Samoobrona LPR SLD UP PiS PO (Marcinkiewicz I PiS) SLD PO PSL Samoobrona LPR PiS Post mat v. Traditional (Kaczynski PiS,SRP,LPR) (Kaczynski PiS,SRP,LPR) SLD PSL PO LPR Samoobrona PiS PR SLD PSL PO LPR Samoobrona PiS PR (Tusk I PO,PSL) SLD PSL PiS PO (Tusk I PO,PSL) SLD PSL PiS PO Dimension 1 8 Dimension 1

9 the beginning of the fourth term, PO and PiS began talks to form an electoral coalition and even considered the possibility of a future cabinet coalition. When an attempt at an electoral coalition failed in the municipal elections of 2002, the two parties went their separate ways and the process of developing independent programs for running against the SLD began in earnest. After PiS won a plurality of votes in the 2005 elections, the economic dimension becomes the main opposition divide with PiS on the social left end of the scale and PO on the libertarian right, a trend that is maintained through Kaczynski s majority cabinet and becomes even starker under Donald Tusk s PO cabinet. Thus the 2005 election firmly marks the turning point at which the regime divide expires and programmatic divisions especially on the economy become prominent. The transformation of PO and PiS was accompanied by the shift in ideological positions of SLD and the Polish Peasant Party (PSL). PSL was a successor communist satellite party known earlier as the United Peasant Alliance (ZSL). The successor-satellite status led to its cooperation on one side of the regime divide with SLD. While in coalition with SLD and voting with that party they anchored together the economic left (see Miller I). While the CHES experts perceived it as quite leftist on economics, it nevertheless dramatically shifted voting patterns following the coalition break-up with SLD. In that period (Miller II), PSL began to vote more similarly to Samoobrona on both dimensions. This pattern is accentuated under PiS s first minority cabinet when PSL actually overlaps with the populist Samoobrona, League of Polish Families (LPR) and PiS, providing support for the latter party s minority cabinet. Once PiS invited Samoobrona and LPR to join a coalition, and not PSL, the latter immediately began to gravitate towards PO and the new round of expert surveys in 2006 also started identifying it as moving rightward on the economic position scale. Recall that we expect parties in parliaments without majority cabinets to be better able to represent their ideological positions in parliament. Because of their minority status, these cabinets are less able to control the legislative agenda to restrict issues considered on the floor. While Buzek II did not represent a stark contrast with the Buzek I majority, subsequent governments exhibit the effect of majority status in shifting the salient dimension of behavior. Miller II, for instance, allows a much clearer divide to emerge between the traditionalist ex-dissident parties (PiS and LPR), along with PSL and Samoobrona, on the one hand, and liberal ex-dissident party (PO), on the other. At the same time, economic attitudes shape the second dimension. Marcinkiewicz I suggests the basis for the PiS-led coalition was going to be the parties attitudes to traditional values, as PSL, Samoobrona PiS and LPR were all voting together and known to be similar in terms of those attitudes. Indeed, during the series of minority cabinets, the post-materialist/traditionalist cleavage is the best way to understand the first dimension which is only reinforced by Kaczynski s majority coalition. By the time Samoobrona and LPR left the coalition and Kaczynski was left to lead a minority coalition, PSL is no longer strongly in the traditionalist camp and this cleavage itself begins to give way as the main dimension organizing voting in the legislature. When Tusk s cabinet comes into office, the first dimension aligns better with the economic cleavage with the coalition pulling PSL out of line with its reputation on that dimension. In summary, legislative behavior in Poland s first years following the transition continued at the Platform s core. 9

10 to reflect the successor-communist versus anti-communist regime divide. But particularly during the first (minority) Marcinkiewicz cabinet, the post-materialist versus traditionalist cleavage emerges as the main basis for parliamentary voting differences and is reinforced by the subsequent PiS-led majority coalitions. However, following PiS loss of power, the economy has become a more important cleavage defining the first dimension of voting in parliament. 4 Slovakia Scholars argue that Slovakia s main societal cleavage in the early to mid-nineties was defined by attitudes to maintaining the Czechoslovak federation (Deegan-Krause, 2013). This cleavage elevated to power Vladimir Meciar and his Movement for a Democratic Slovakia (HZDS), an upshot of the anti-communist Public Against Violence (VPN). HZDS together with Vaclav Klaus s Civic Democratic Forum (ODS) finalized the velvet divorce between the two component parts of Czechoslovakia in December In the post-divorce years, the main political divide was between supporters and opponents of Meciar s authoritative and charismatic ruling style. When HZDS former allies failed to exceed the 5% threshold and no party wanted to enter into a coalition with them, the cleavage defined by nation-building and attitudes to authority finally gave way to matters of the economy (Deegan-Krause, 2004). This is largely corroborated by the 2006 CHES data, which shows that the most salient cleavages along which parties presented themselves to voters were spending, deregulation and redistribution. Figure 3 shows results from ideal point estimation of roll call voting in four legislative cycles, compared with party scores on the same CHES economic and postmaterial-traditional surveys considered for the Polish scale above. We first contrast the last majority Dzurinda cabinet (1998) with his minority cabinet (2002), made up of the four party coalition of two Christian Democratic parties (KDH and SDK-u), the Hungarian Coalition (SMK), and a the New Alliance Party (ANO). The 2002 cabinet started off as a majority cabinet, but lost its majority status when three MPs from the government coalition parties left their caucuses to become independent. However, despite lacking a coalition majority for much of the period, Dzurinda manages to retain and even increase the degree of coalition unity and the clarity of the government opposition divide on economic issues. This may have been facilitated by the fact that the cabinet composition remained intact from the initial majority and that the Speaker Pavol Hrusovsky of the governing KDH remained in his post. 8 With the loss of the parties associated with what became Smer (SDL and SOP), the remaining parties (KDH, SMK and ANO) had a stronger basis for unity on the main economic dimension, despite including the slightly more centrist SDK-u and the more traditionalist KDH. 9 As a result, a traditionalist-versus-postmaterialist emerged as a second dimension of legislative voting. The constraints placed on parties by 8 In addition, the coalition was dealing with especially aggressive opposition from Smer that emphasized the government-opposition divide. According to Peter Ucen Smer tried to snowball as many antigovernmental votes as possible, while putting aside coalition potential and prospects for its recently adopted social democratic programme to be implemented; the party apparently bet on making itself indispensable to any coalition to come after elections (Učeň, 2006). 9 Note, however, that the survey data for Smer and its predecessors is based on the same 2002 survey. 10

11 Figure 3: Legislative Roll Call Positions and Expert Surveys on Economic Left-Right and Post- Material/Traditional Dimensions: Slovakia (Dzurinda SDK,SDL,SMK,SOP) PSNS SDL HZDS SOP SDKU SMK SNS KDH (Dzurinda SDK,SDL,SMK,SOP) PSNS SDL HZDS SOP SDKU SMK SNS KDH (Dzurinda SDKU,SMK,KDH,ANO) (Dzurinda SDKU,SMK,KDH,ANO) Smer KSS HZDS ANO SMK SDKU KDH Smer KSS HZDS ANO SMK SDKU KDH Dimension (Fico SMER,SD,SNS,LS HZDS) LS HZDS Smer SNS SDKU DS KDH SMK L R Economic Dimension (Fico SMER,SD,SNS,LS HZDS) LS HZDS Smer SNS SDKU DS KDH SMK Post mat v. Traditional (Radicova SDKU DS,SaS,KDH,MOST) (Radicova SDKU DS,SaS,KDH,MOST) Most SDKU DS Smer KDH SaS SNS Most SDKU DS Smer KDH SaS SNS Dimension 1 Dimension 1 11

12 their government status are well illustrated by taking into account KDH apparent similarity to HZDS in 1998 and 2002 on matters of traditionalism versus post-materialism. The difference between the two cabinets suggests that legislative behavior in 2002 started responding to a cleavage change, moving away from the pro-independence versus pro-federation cleavage associated with Meciar s rule. The next elections in 2006 were called early because KDH left Dzurinda s already fragile coalition, degrading its minority status even further. The party Smer ( Direction ) formed during the previous term won a plurality of seats, with Dzurinda s Christian Democrats coming in second. Smer s leader Robert Fico became the Prime Minister and created a cabinet with two junior partners the Slovak National Party (SNS) and Meciar s new People s Party-Movement for a Democratic Slovakia. Although, these three parties voted with one another more often than with other parties prior to entering together into a cabinet coalition, Smer and Meciar s LS-HZDS began to diverge on matters of spending and government involvement according to the CHES 2006 survey. Yet, despite these differences, the three parties appeared together as a unified bloc once in government. In addition, postmaterialism-versus-traditionalism continued as a second dimension in this term, almost orthogonal to the first dimension government-opposition divide. The 2010 elections led to a much more one-dimensional and polarized party system with clearly unified parties organized along the economic left-right dimension. Only six parties won legislative seats. Robert Fico s Smer New Direction secured 12 additional seats, but despite winning a plurality could not form a cabinet, as its former coalition partner LS-HZDS failed to win any seats and SNS barely cleared the 5% threshold. As a result, the runner-up Slovak Christian and Democratic Union (SDKU-DS) formed a cabinet coalition with the Christian Democratic Party (KDH) and two new parties: Freedom and Solidarity (SaS) and the interethnic party Most-HiD. Everything about this coalition spoke to the transformation of the Slovak Party system from one characterized by a gap between elite conflict in the legislature (that is, pro-meciar and anti-meciar) and substantial cleavages in society, to a system where parties began to reflect programmatic party positions more clearly. The transformation of the Hungarian coalition into an inter-ethnic party with the name most ( bridge ) highlights the decline of the ethnic/nationalistic cleavage that had elevated HZDS and SNS. It is also apparent that governing parties who are leftist in their economic attitudes are voting together against parties in the opposition who represent economic views of the right. 5 Czech Republic In contrast to Slovakia or Romania, the Czech communist party failed to reform into a viable major party and remained marginalized while the economic left has been successfully represented by the CSSD. Partly as a result, the regime divide cleavage dissipated shortly following the first democratic elections, which were characterized by a complete defeat for the communists (Grzyma la-busse, 2001b). Meanwhile, the Czech political and post-transition fiscal context enabled a very fast-paced path toward liberal economic reforms led by ODS, making the nation a leading economy in the region. The first two cabinets were led by coalitions created on the basis of the dissident Civic 12

13 Forum (OF). After Havel became president, OF s two main successors ODS and the Civic Forum Alliance (ODA) formed a cabinet and were joined by another, pivotal coalition partner, the Christian Democratic Union-Czechoslovak People s Party (KDU-CSL). The KDU- CSL was an alliance between Czechoslovakia s People s Party (CSL), a party with pre-wwii roots that was allowed a puppet presence in the Communist legislature and government, and the new Christian Democratic Union, intended to emulate the German CDU. Subsequently, the Czech Republic has been one of the more stable party systems in the post-communist region (Mansfeldova, 2013). This stability is largely corroborated by the CHES survey and the roll-call voting patterns we present here. 10 According to CHES, between 2002 and 2009, the three most salient cleavages were spending, redistribution, and deregulation. Legislative behavior data meanwhile, shown in Figure 4, suggest that both the constellation of the parties in the system and the basic dimensionality has remained consistent: the Civic Democratic Forum (ODS) has anchored the economic dimension on the right and the Czech Social Democratic Party (CSSD) on the left. Recall that minority status is reflected in the roll-call patterns in general because they are likely to lack the numbers to enable control over the agenda, allowing a larger variety of floor coalitions to emerge and potentially a clearer reflection of party reputations in roll call voting. This means they are less likely than majorities to systematically avoid votes on which they expect to fail and certainly that they have less reliable floor majorities for passage. Thus, party behavior in such contexts is at least never more constrained than under a majority cabinet. 11 Our data for Czech Republic provides several minority cabinets (one a caretaker) to examine for this purpose: Tosovsky s (in 1998), Milos Zeman s ( ) and Mirek Topolanek s (in 2006). Generally, minority situations appear to have accentuated the second dimension of voting on issues that crosscut economics. In 1997, Klaus cabinet ostensibly fell over financial corruption scandals, though it is noteworthy that the partners were in agreement only on economic matters and constrained significant disagreement on issues of traditionalism. To prevent any economic uncertainty after both ODS and KDU withdrew from the cabinet, Vaclav Havel asked Central Bank leader Josef Tosovsky (1998) to lead a caretaker cabinet that included representatives from parties in the previous coalition. Differences between the parties on a second dimension became clearly apparent, although these differences were also not clearly aligned with CHES traditionalism at the time. The early elections of 1998 were won by the Social Democrats (CSSD), but they failed to create a coalition cabinet with the communist KSCM. But ODS agreed to a confidence and supply agreement to enable a minority cabinet led by Milos Zeman and the CSSD. As we expect from a minority cabinet, Zeman s government allowed a second dimension to emerge fully, dividing the parties more clearly than ever on traditionalism, while retaining a first dimension based on economic differences. The 2006 election produced an evenly split house between ODS, KDU-CSL and the Independent Green Party (SZ) on the one hand, and the CSSD with the Communists, on the other. ODS Mirek Topolanek was appointed to lead the first cabinet, but failed to gain 10 We currently we lack data between 2002 and We say never more, because in some instances, as in the case of AWS described above, even a ruling majority may not be able to capitalize on its numbers and engage in strong agenda control in a way that will produce clear constraints on legislative behavior 13

14 Figure 4: Legislative Roll Call Positions and Expert Surveys on Economic Left-Right and Post- Material/Traditional Dimensions: Czech Republic (Klaus II ODS,KDU CSL,ODA) (Klaus II ODS,KDU CSL,ODA) CSSD KDU CSL ODS KSCM CSSD KDU CSL ODS KSCM 1998 (Tosovsky (Caretaker; KDU CSL,ODA) 1998 (Tosovsky (Caretaker; KDU CSL,ODA) CSSD KDU CSL KSCM ODS CSSD KDU CSL KSCM ODS (Zeman CSSD) (Zeman CSSD) Dimension 2 CSSD ODS KDU CSL KSCM (Topolanek I ODS) SZ CSSD ODS KDU CSL KSCM L R Economic Dimension 2 CSSD ODS KDU CSL KSCM (Topolanek I ODS) SZ CSSD ODS KDU CSL KSCM Post mat v. Traditional (Topolanek II ODS,KDU CSL,SZ) 2007 (Topolanek II ODS,KDU CSL,SZ) CSSD ODS KDU CSL SZ KSCM Dimension 1 14 CSSD ODS KDU CSL SZ KSCM Dimension 1

15 confidence of the legislature and led a care-taker cabinet until 2007 when he finally assembled a right wing-center coalition with KDU-CSL and SZ. When we compare the transitions to minority cabinets under ODS, we clearly see the effect of losing majority agenda control. Between Topolanek I and II, the formation of a majority government induces a much stronger economic pattern on the first dimension of voting, diminishing the cross-cutting issues on which the economic left and right had cooperated. 6 Romania According to the expert surveys, cosmopolitanism (the importance of cosmopolitanism versus nationalism) and positions towards ethnic minorities were the two cleavages dominating politics in Romania in 2004 and 2008, followed closely by matters of spending and redistribution. However, Romania stands out in expert evaluation surveys, because in contrast to the remaining countries considered here, many more potential cleavages were considered salient by experts. In 2004, spending, deregulation, redistribution, civil liberties, urban, cosmopolitanism, and ethnicity all had an average salience score greater than 5 for every single party winning more than 10% seatshare in the legislature. The lack of clarity in what defines salient policy competition could originate in the phenomenon of catch-all (Kirchheimer, 1966) or clientelistic (Kitschelt, 1999) parties in Romania, focused more on short-term vote-getting tactics at the expense of developing a programmatic reputation. Thus, the assumption that parties might face costs for voting behavior inconsistent with their public policy reputations may not be well supported. Relative to other cases considered here, Romania was also an outlier in terms of the communists exit from office. While dictator Nicolae Ceausescu was swiftly executed, the leadership of the National Salvation Front (FSN), Ion Illiescu, was a former Communist Politburo member himself. As the National Salvation Front inherited much of the infrastructure of the Romanian Communist Party, the FSN enjoyed a considerable advantage over the other contenders of the 1990 elections the National Liberal Party (PNL), the National Peasant-Christian Democratic Party (PNT-CD), the Social Democratic Party (PSDR), and the Democratic Magyar Union of Romania (UDMR). Unsurprisingly, FSN dominated in both the parliamentary (68%) and presidential (85%) elections in May 1990, but not without resorting to intimidation of the historical anticommunist opposition parties and even violence. Following the elections, a conflict broke out in the FSN between the Prime Minister Roman, leading a group closer to the historic parties and favoring more rapid reforms, and the President Iliescu, leading a group closer to some newly emerging parties on the nationalistic left, such as the Socialist Labor Party (PSM), the Party of Romanian National Unity (PUNR), and the Greater Romania Party (PRM). Eventually Iliescu started his own party the Party of Democratic Socialism of Romania (PDSR). In the meantime, the historic opposition parties joined forces with other anti-fsn parties to create the Democratic Conventions (CDR). This prolonged the existence of regime divide in Romania. The parties defining the regime divide or at least their leaders remained the main players in power until 2004, which is the first government for which data exist. The main contenders in an election that attracted 64 parties to the competition, were the Iliescu and Ilie Nastase-led 15

16 Social Democratic Party PSD and the Democratic Alliance (DA) and only six parties (PSD, PNL, PD, UDMR, PC, and PRM) won seats in the assembly. The PSD formally emerged as a winner of the election, after it joined forces in a coalition with the Humanist Party of Romania (PUR) obtaining 31% of the votes (the coalition was called the Justice and Truth Alliance), but because the PSD presidential candidate failed to win the Presidential election, Calin Popescu Tariceanu of PNL (the party that recorded the biggest gain in electoral seats) ended up being called upon to lead the cabinet. To be confirmed, Taricaneu had to secure the support of 265 MPs from his own party as well as PD, UDMR, PUR/PC. Figure 5 again combines expert survey placements on the economy (left panel) and on post-materialist versus authoritarian values (right panel) with roll call data. It suggests that the first coalition was most clearly aligned on its pro-market views as all the parties in the grand coalition appear to contrast the parties in the opposition. However the economy is not entirely a precise reflection of the first dimension of legislative voting because of the location of PRM. The scale on the right reflecting involvement of government on issues of morality does little better, also because of the location of PRM. This suggests there is a third dimension in legislative voting in which the PRM votes with PSD, arguably reflecting a persistent crosscutting ethnicity cleavage. The dimensional patterns of legislative behavior are far from stable, however. Shortly following the formation of the cabinet, a number of deputies switched caucuses or formed new ones, slimming down the majority supporting the cabinet even more. Yet even though the alliance between PD and PNL was strained, the parties appear to have substantially controlled the legislative agenda. Their efforts were apparent when the cabinet became a minority of only PNL and UDMR. Its minority status is reflected in the roll call pattern. As the government loses control over the floor agenda, the first dimension separating the government from the opposition (which was also mostly economic) dissipates and allows parties previously in government to separate (see PD and PUR/PC on the one hand and PSD on the other), shifting the first dimension of party behavior away from economic differences. The 2008 elections were held under the new electoral system under which PRM failed to win enough votes to exceed the 5% threshold. This itself changed the roll call patterns in that questions of ethnicity tend not to interfere with other issues. Following the elections President Basescu asked PDL leader Emil Boc to create the cabinet and the latter invited PSD, who had decisively won the elections, to form a grand coalition. This was an unusual coalition, particularly on the Liberals side because, up to 2007, the PD-L statute banned collaboration with the Social Democrats. The first Boc cabinet continues the pattern of inconsistency between parliamentary voting and party positions, especially for PD-L. Following PSD s departure, PD-L entered into a coalition with UDMR, causing PSD to start voting with PNL with UDMR aligning with its new coalition partner. In terms of economics and traditionalism, the government-opposition divide changed substantially yet neither can fully explain legislative behavior at this juncture. Indeed, over the eight years of data on four Romanian cabinets, neither the economic cleavage nor the postmaterialist-traditional cleavage provide a good fit with legislative behavior without substantial deviations. We take this as some evidence in the direction that Romania s parties may have faced lower costs to shifting parliamentary positions and/or that their behavior is particularly sensitive to the legislative constraints. 16

17 Figure 5: Legislative Roll Call Positions and Expert Surveys on Economic Left-Right and Post- Material/Traditional Dimensions: Romania (Tariceanu I PNL,PD,UDMR,PUR/PC) (Tariceanu I PNL,PD,UDMR,PUR/PC) PRM PSD UDMR PNL PC PDL PRM PSD UDMR PNL PC PDL Dimension (Tariceanu II PNL,UDMR) PSD PRM PDL PC UDMR PNL (Boc II PD L,PSD) PSD PNL PDL UDMR L R Economic Dimension (Tariceanu II PNL,UDMR) PSD PRM PDL PC UDMR PNL (Boc II PD L,PSD) PSD PNL PDL UDMR Post mat v. Traditional (Boc II PD L,UDMR) (Boc II PD L,UDMR) PNL PDL PSD UDMR PNL PDL PSD UDMR Dimension 1 Dimension 1 17

18 7 Conclusion The roll call record reflects an important aspect of party behavior at the elite level, one tied directly to their efforts to represent the salient issue cleavages in society. Obviously, roll call voting in parliamentary regimes will be shaped directly by government membership as voting together on certain issues is a necessary part of coalition agreements. Yet roll call records are a function not only of preferences and discipline, but also of the legislative agenda manipulating the choices presented to the entire chamber. However, we expect that governments vary in the degree to which they control the agenda. Here, we focus mainly on one such variation: that minority governments may find it especially difficult to shape the organization of the chamber. This in turn has important implications for the dimensions of politics at the elite level. As we show throughout our four cases from Eastern Europe, depending on who is in government and how they shape the voting arena, the apparent ideal points (i.e., parliamentary voting relationships among parties) of parties can shift dramatically. Our exploration of Eastern European parliaments has also raised questions about the interaction between ideology, party organizations and legislative procedures. We point out two interesting phenomena that stand out as requiring more nuanced explanation in future work. In Poland under Buzek s AWS-UW coalition, we observed behavior that reflected relatively poor agenda control, despite majority status. Consequently, the government opposition divide is so obscured that other axes of political conflict become visible, specifically, the post-communist regime divide. Conversely, in Slovakia under Dzurinda we saw a minority government produce a relatively strong government-opposition first dimension. In the first case, one possibility is that very weak parties (in this case, on the brink of collapse) are not positioned to delegate their power to use agenda control even when they negotiate majorities because they lack the organization needed to appoint loyal agents on agenda setting bodies. We note that most of the region by the time of these data does not have parties as uninstitutionalized as AWS. With regard to Slovakia s persistent government-opposition economic divide, it may be the case that legislative rules may make it easier for minority governments to control the agenda. This is especially likely to be the case where the chamber organization was put in place when the government possesses a majority. In general, this case emphasizes the need to consider in detail the rules of each chamber. In countries where the chamber rules are more majoritarian, party positions under majority governments may be less sensitive to the loss of majority status. A limitation of the data we present owes to our simplified focus on just two major issue areas, though this allowed us to compare across time using the CHES surveys. As remarked in the introduction, these are important areas. Social and fiscal conservatism do not go hand-in-hand in the post-communist region, hence it is useful to see where the traditionalist/post-materialist dimension emerges as one that shapes legislative behavior. Still, a major omission in this note is the lack of comparison with attitudes regarding the European Union, which is often salient in the period under study (e.g., Grzymala-Busse and Innes 2003). Although in many cases EU attitudes overlap with the policy areas we focused on here, these positions undoubtedly provide additional value in understanding some of the behavior we observe. Furthermore, the programmatic orientation of parties must be more fully accounted for 18