Party Organization and Electoral Volatility in Central and Eastern Europe

Enhancing voter loyalty

Sergiu Gherghina

• Publisher: Routledge
• Year: 2014
• Print ISBN: 9780415737166
• eBook ISBN: 9781317618973
• Edition: 1
• Topic: Politics & International Relations
• Subtopic: Politics
• Index: Politics & International Relations

Political parties in post-communist countries have very high levels of electoral volatility. In these environments, political factions fail to establish long-term connections with the electorate and thus regularly rise and fall from the political arena.

This book provides an organizational explanation for the variations in party-level electoral volatility. It looks comparatively at 29 political parties in six Central and Eastern European democracies between 1990 and 2008 to examine how political parties can influence their electoral environment. Using empirical evidence, Gherghina tests the effect of candidate selection procedures, membership organizations, and re-nomination of incumbent MPs on voters' loyalty, and in doing so, demonstrates how party organization greatly affects electoral stability.

Including case studies from Bulgaria, Czech Republic, Hungary, Poland, Romania, and Slovakia this book will be of interest to students and scholars of comparative politics, party politics, democratization, elections, and Central and Eastern European politics.

1 Layers and sources of electoral volatility

Introduction

Many factors shape a voter’s decision to attach loyalty to a particular political party. These factors range from the socio-psychological characteristics of the voter to institutions and the outcomes they generate. In the attempt to identify the determinants, previous research on advanced democracies has tended to explain electoral volatility according to three major categories of factors: cleavage structures; electoral institutions; and economic performance. The resulting picture is complex and blurred, characterized more by mixed evidence than by clearly identifiable explanatory patterns. In general, these explanations appear to be context-dependent, as they are often case (e.g., country or region) or time specific.

The inception of this type of explanation was prompted by Lipset and Rokkan (1967), who argued that parties mobilize voters on the basis of cleavage structures that have existed for decades in West European countries. Two studies that followed sought to add nuance to Lipset and Rokkan’s theory. In their attempt to explain the stability of post-1945 West European countries, Rose and Urwin (1970) presented refined evidence of the strong social connections between voters and parties. Pedersen (1979) argued that electoral volatility after 1960 was rooted in the dealignment of class and party. All these studies reveal one basic conclusion: longitudinal change in electoral preferences is a result of the transformations that take place in terms of values and social structures (Dalton et al. 1984, p. 451).

Building on these studies, the research that followed put forth yet more explanations of electoral volatility. In his analysis of several West European countries, Crewe (1985a, p. 129) argues that ideology plays a role in voting behavior. The way in which people vote and change their vote can be explained in terms of the relationship between their own positions and the perceived party positions on major divisive issues. This emphasis on the ideological sources of vote commitment is complemented by Bartolini and Mair’s focus on the party system format (i.e., the number of competing parties), the electoral system, and the political system (i.e., voter turnout). The complexity of providing an exhaustive explanation of volatility prompted them (1990, p. 282) to categorize the causes of volatility into five main groups: changes in the electoral system; the format of the party system; cultural segmentation; the density of organizations connected to political parties; and short-term factors.

Lane and Ersson (1999, p. 198) conceptualize and operationalize this last “short-term factors” category by breaking it down into party membership, votes for the left, convergence on the economic right–left scale, and economic vulnerability due to openness. Using such detailed accounts, their results explained almost 70 percent of the aggregate variation in electoral volatility compared to the less than 50 percent explained by Bartolini and Mair. To sum up: over the course of three decades, the number of explanations applying to a limited number of countries significantly increased. Similarly, the type and nature of explanations became diversified. At the same time, the quest for determinants has shown that electoral volatility is a multi-layered concept with complex analytical and empirical dimensions.

This chapter argues that variation in vote shifts can be best understood at party level and reviews the determinants of this variation at both the macro (party system) and the meso (party) levels. The first section provides an analysis of the layers of volatility and explains the reasons why volatility in CEE should be regarded as a function of individual parties and not of the party system. In spite of this analytical differentiation, the party system and party levels are interconnected. This is why the sources of electoral volatility can have an influence on both layers. Hence, the next three sections present explanations of electoral volatility in a sequential manner, approaching general (party system), mixed, and individual party levels in that order.

Party system and party volatility

Often calculated and analyzed at the aggregate party system level, electoral volatility emphasizes changes that occur at different levels of representation (Crewe 1985b, p. 8). Bartolini and Mair (1990, p. 25) distinguish between the party system, party block, and individual parties as three distinct levels at which volatility can be calculated. At the party system level, volatility encompasses electoral variation as a whole and includes the sum of parties’ net electoral gains or losses in consecutive elections. It is a reliable indicator of relative strength and social rooting of political actors in consecutive elections (Krupavicius 1999, p. 8).

With respect to the changes identified at the party block level, these can refer either to a party family or to parties grouped along several dimensions: left– right, old vs. new, or opposition vs. government (Pennings and Lane 1998). We can distinguish between intra- and inter-block volatility. The first refers to electoral changes between parties from the same block, whereas the second captures vote swings between blocks. At the party level, electoral volatility is a reliable indicator of the acceptance of a party in a society (Lane and Ersson 1999, p. 127) and is often included in measurements of party institutionalization. Electoral volatility at the party level provides an indication of the number of voters gained or lost by individual parties between elections. Calculating volatility at the party level diminishes the major shortcoming of volatility calculated at the party system level: its blindness towards the parties contributing to it (Sikk 2005, p. 408). Thus, an accurate estimation of vote shifts between existing parties is captured best by party volatility.

Analysis of volatility at any of the three levels has a unique departure point: the electoral performance of a political party across time. The stability of electoral performance over time can be assessed within the broadest framework in which the party competes (i.e., the party system), alongside competition partners on various dimensions (party block), or at the individual party level. In this respect, the political party is the most important unit of analysis. This observation is also valid on theoretical grounds. In a functional sense, political parties are the main tools of citizen representation. Parties are the visible actors that run in elections and are held accountable by voters for their actions and policies. A party’s electoral performance is a function of the interaction between this party and voters, of the exchange of votes for policies. In this respect, it makes the most sense to analyze volatility where it occurs in practice: at the level of the individual party.

Party system and block volatilities have analytical value, but are constructed on the basis of behaviors observed at other levels. In this sense, Bartolini and Mair (1990, p. 4) explain that “electoral instability is a phenomenon originating at the level of individual behaviour, but which acquires political relevance only by reference to the changes which it produces in the structure of party systems.” Block volatility is an intermediary between the party system and individual party levels. It aggregates electoral swings on certain dimensions: ideological, governmental, etc. Although it is necessary to understanding the dynamics within a political system, block volatility does not provide the basis upon which substantial conclusions can be drawn about parties’ electoral volatility, as the configuration of these blocks often changes. In conceptual and empirical terms, the substantive characteristics of party blocks are rarely consistent across countries, thereby limiting our ability to compare these units across different settings. This is why this concept has been left out of my analysis.

Three empirical arguments – all specific to CEE – lend support to the party level approach in explaining electoral volatility. First, the post-communist party systems are structurally unstable. They are characterized by a large number of entries and exits, thereby generating episodic appearances in the political space. In consecutive elections, the format and composition of the party system are rarely similar, and methodological problems related to comparability arise. For example, Bulgarian politics has displayed continuous change over the past two decades, with only three parties that have successfully gained parliamentary representation on a continuous basis.

As a result of the fact that the same parties have not participated in all elections, the party system measurement cannot distinguish between volatility among stable parties and volatility created by entry and exits (Neff Powell and Tucker 2008, pp. 2–3). As these two types of volatility have different causes and effects on political processes, it is crucial to isolate and calculate them separately. By lumping them into a common measurement, we create confusion and provide results that are not only difficult to interpret but also far removed from reality. Moreover, previous studies (Sikk 2005) have illustrated that in CEE an index of volatility calculated at party system level ignores the parties contributing to it. Consequently, looking at the components of the party system and evaluating their performance from a cross-sectional and a cross-time perspective allows for a deeper understanding of the political dynamics in the region.

Second, only an investigation conducted at the party level can answer questions related to splits, mergers, coalitions, and label modifications. Such developments are essential to tracing the histories of political parties which are relevant in terms of electoral performance. These decisions are elite driven and usually rely on beliefs about voter behavior, being assumed either to influence it or to be an effect of it. Sometimes the two rationales are combined. For example, electoral coalitions are both proactive (as the parties that decide to join forces wish to maximize their performance in the upcoming elections) and reactive (as they are built on favorable signals received from the electorate). Empirically, this interplay between political actors and their supporters can only be fully captured if developments within parties are closely examined. The following hypothetical example illustrates this argument: Party A, running in the elections t₀ and t₁, splits halfway between the two elections. The splinter group, party B, does not compete alone in elections but merges with another party – party C – in election t₁. Assuming that party B has some electoral support, parties A and C will both end up with different degrees of electoral volatility. Thus, if party dynamics are not taken into account, the relationship between voters and representatives is misleading.

Finally, size is a relevant feature of political parties when considering their societal support. Apart from the visibility of the party dimension in terms of government functioning or in occupying positions in the political system, this feature also differentiates parties with respect to electoral volatility. A small party that appeals to 5–6 percent of the electorate and loses 2 percent of its usual vote share suffers much more than a large party that gathers, on average, 20–22 percent of the votes and loses the same percentage. Thus, not taking into account the differences and conflating the volatility scores into a single measurement at the party system level means that the gains and losses of small parties are underestimated, whereas for large parties they are overestimated. A methodological shortcoming such as this cannot be overcome when taking into account the effective number of parties. The score obtained indicates a relative division of power and importance between parties within the system at a given moment; it still does not say much about the extent to which parties perform relative to their own position in the system. Put simply, calculating absolute vote transfers does not provide an accurate image of voting patterns within a party system; it provides only comparability between countries. The longitudinal performance of parties is best accounted for in terms of relative gains and losses.

As a consequence of all of these features, electoral volatility at the party level can best reflect the dynamism of relevant competitors within the system. The determinants of party system volatility are unlikely to explain variation in volatility at the party level. They are constant for all of the components of a party system at a particular moment in time. For example, political parties within a country develop in the same democratic environment, face similar economic crises, and fight according to the same electoral rules. However, there is an undeniable linkage between parties and the party system that is visible in many areas, from policy drafting to electoral results. Accordingly, it is expected that a linkage will exist between the party and party system volatility. The electoral volatility of the party system is the sum of the vote shifts among its components. Conversely, the volatility of individual parties can be shaped by the environment in which they develop. That is why some explanations provided by previous studies were primarily identified at the level of individual parties and afterwards extrapolated to the party system. Moreover, as they are not always easy to separate in practice, the effect of some determinants is visible at both levels. For this reason, a separation of the layers of volatility and their sources is useful. The following section reviews the literature discussing the determinants visible solely at party system level.

Determinants of party system volatility

Four general factors have been considered to influence electoral volatility in the party system exclusively: the electoral system (and changes to it); the level/degree of democratization;¹ the party system format; and the ideological polarization of political parties. This section sheds light on the way in which these factors function and how they (may) apply to the CEE context.

The electoral system

The components of the electoral system are the first to make a difference with respect to volatility. In general, the nexus between the stability of institutional rules and that of electoral choice is uncertain (Bielasiak 2002). In particular, there are a few observable linkages. For example, in their longitudinal study Bartolini and Mair (1990, p. 39) explain volatility in terms of several changes in electoral institutions (including regulations), franchise enlargement, introduction or abandonment of compulsory voting, provisions of electoral laws, and the disproportionality of different electoral systems. In contemporary times, most of these elements either do not vary (i.e., the franchise) or are no longer on the political agenda (i.e., compulsory voting). Consequently, only the proportionality of the system continues to be a variable that can explain variation in electoral support. The mechanisms at work here are driven in large part by the formula employed and the electoral laws at work specifying modifications to the electoral threshold or district magnitude.

Electoral setting can, thus, indirectly influence voter choice and the stability of preferences by way of the “psychological” (complementary to the “mechanical”²) effects of electoral systems (Duverger 1954). These effects imply the presence of strategic incentives both for voters (who seek to avoid wasting their votes) and for politicians under different electoral rules (Blais et al. 2001; Gschwend 2007). These incentives influence the number and nature of political choices that voters confront. For example, high explicit electoral thresholds discourage political parties from contesting seats (Rae 1971; Lijphart 1994; van Biezen 2003; Bakke and Sitter 2005). In a situation such as this, voters face limited choices and are encouraged to cast strategic votes for larger parties. By discouraging the expansion of alternative vote choices, the level of electoral volatility is likely to be reduced.

Yet none of these theoretical arguments is supported by evidence from the CEE countries. To start with, the electoral threshold, the average total electoral volatility calculated at the party system level (Enyedi and Casal Bertoa 2011, p. 134), indicates that countries with similar thresholds display stark differences. For example, Romania since 2000 and Poland have similar thresholds for political parties (5 percent) and coalitions (8 percent),³ but the average volatility scores are extreme in both cases: 19.1 for Romania and 34.3 for Poland. As these are also the only countries in CEE that did not use electoral thresholds in their first free elections – those of 1990 in Romania and 1991 in Poland – the likelihood that similar volatility trends and values will be experienced in both cases increases. Similarly, the Czech Republic and Slovakia have the same 5 percent threshold for political parties, but the average volatility of the former is around 21 percent, whereas in Slovakia it is 6 percent higher.

Moreover, the theoretical expectation according to which countries with higher thresholds should have lower volatility is not supported by evidence in CEE. During the time period under investigation, 176 out of the 386 seats in the Hungarian Parliament were allocated on the basis of competition in single-member districts. These districts have very high implicit thresholds. The average volatility of Hungary’s party system is higher than that of Romania and that of the Czech Republic, both of which have lower electoral thresholds: 19.1 percent and 21 percent respectively. In addition, a longitudinal inve...

Table of contents

• Cover
• Half Title
• Series Title
• Title Page
• Copyright Page
• Table of Contents
• List of figures
• List of tables
• Acknowledgments
• List of abbreviations for the political parties included in the analysis
• List of abbreviations for other political parties, coalitions, or alliances referred to in the book
• Introduction: institutional determinants of voter choice
• 1 Layers and sources of electoral volatility
• 2 Party organization and electoral volatility: an analytical model
• 3 Mapping electoral volatility in Central and Eastern Europe
• 4 The benefits of decentralized candidate selection
• 5 Loose mobilizing networks
• 6 The continuity of representation
• 7 A multivariate analysis of electoral volatility
• Conclusions
• Appendix: methodological choices
• References
• Index