Parties and Party Systems in Established and New Democracies
2.4 Interaction between Parties: The Party System
When we shift our attention from individual political parties to the nature of party interactions within a political system, our unit of analysis becomes the party system. This section reviews some of the key concepts and methods used in the party politics literature to distinguish between different types of party systems. This is useful not only for showing how scholars have tried to make sense of the wide variety of party systems, but also for gaining a better
108 This observation is shared by Basedau and Stroh, who argue that when it comes to measure
party institutionalization “it is very difficult to differentiate between preconditions, defining elements, and immediate consequences”. Basedau and Stroh, Measuring Party Institutionalization in Developing Countries: A New Research Instrument Applied to 28 African Political Parties.
understanding of the specific party system typology that will be used in the next chapter on post-war societies as well as the subsequent case study chapters.
Different Methods for Counting and Classifying
The variation between party systems across the world is substantial. Current systems range from the one-party regimes in China, Cuba and Vietnam to different forms of multi-party systems in Britain, the Netherlands and Lebanon. Generally speaking, party systems vary on a number of aspects, including the number of parties in elections and parliament, the relative size and strength of political parties, the ideological distance between parties, their willingness to cooperate in government formation and governance, the degree of openness of competition over government, and the degree of institutionalization.109
In order to analyze the ways in which parties interact, scholars have developed different classifications of party systems. A first well-known typology of party systems is that by Jean Blondel, whose main contribution is his distinction between different types of two- and multiparty systems. Focusing exclusively on the party systems of established Western democracies, Blondel distinguishes between two-party systems, two-and-a-half party systems, multiparty systems with a dominant party and multiparty systems without a dominant party.110 Although his classification seems a step forward in comparison with the older, more general typologies of Duverger and Neumann,
109 Wolinetz, “Party Systems and Party System Types”, p.53.
Blondel’s typology has been criticized for not being distinctive enough and putting together party systems whose dynamics can be very different.111
A more elaborate and comprehensive classification was subsequently developed by Giovanni Sartori. The significance of his work lies in the use of four criteria for classifying party systems: the number of parties, the relative size of parties, the ideological polarization of parties and the interaction pattern between parties. By using these criteria and distinguishing between the consolidated party systems of ‘formed states’ and the unconsolidated party systems of ‘unformed or fluid polities’, Sartori has developed two separate, but related typologies, each subdivided into different classes of party system (further explained below).112
Other, more recent classifications are those by Alan Ware, Peter Mair and Alan Siaroff, who in addition to numbers, relative size and ideological polarization of parties also take into account the extent to which parties penetrate societies, the closed or open structure of competition for government, and the size and relative strength of parties.113
It is important to recall here that most classifications of party systems focus on the numerical criterion, i.e. how many parties there are in a system. But among these classifications, there are different methods for counting the relevant number of parties. Some of the most well-known methods use quantitative, mathematical rules for establishing the number of relevant parties. These include
111 Sartori,
Parties and Party Systems, pp.105-107; and Wolinetz, “Party Systems and Party
System Types”.
112 Sartori,
Parties and Party Systems.
113 Ware,
Political Parties and Party Systems, pp.150-151; Mair, “Comparing Party Systems”;
Rae’s fractionalization index and Laakso and Taagepera’s method for measuring the relevant or ‘effective’ number of electoral or parliamentary parties.114 By measuring parties’ weighted vote or seat share in elections both methods provide a continuous measurement of the relative size of parties.
But there are also methods that use qualitative rules for assessing how many relevant parties there are. The simple counting method, for example, consists of nothing more than counting all political parties that are registered or represented in parliament within a particular country. A more widely used method is to include only those parties that obtain a certain number of votes or seats in parliament. Common thresholds in this regard are 3%, 5% or 10%. The problem with both methods, however, is that they say little about a party’s political strength or relevance in the overall party system. For example, small parties that are unable to obtain 3%, 5% or 10% of the seats in parliament can still play an important role in the creation of a coalition government with parliamentary majority. That is why a third qualitative method is more useful.
According to this method by Sartori, parties should not only be counted on the basis of their relative size (vote or seat share), but also on the basis of their potential to influence the formation of government coalitions and direction of party competition. In this view, only those parties are relevant that find
themselves “in a position to determine over time, and at some point in time, at least one of the possible governmental majorities” (coalition potential), or whose “existence, or appearance, affects the tactics of party competition and particularly when it alters the direction of the competition –by determining a switch from
114 Rae,
The Political Consequences of Electoral Laws; Laakso and Taagepera, “‘Effective’
centripetal to centrifugal competition either leftward, rightward, or in both directions– of the governing-oriented parties” (blackmail potential).115 In practice, the coalition potential of a party can easily be deduced from the fact whether it has taken part in a government coalition or whether a party has provided a government with the necessary political support for it to take or stay in office. The blackmail potential of a party can be inferred by its ability to veto legislation within parliament.
Sartori’s measure of relevance has the benefit that it draws attention to the interactions between parties and outlines the power configuration within party systems. This stands in sharp contrast with the other counting methods which focus mainly on the format (numbers of parties) and do not really capture the real essence of a party system, i.e. the workings and interactions between parties. This is important because a party system is more than the sum of its individual component parties. After all, it is “the forms and modes of their coexistence [that] define the ‘party system’”.116
The value of the various typologies and counting rules becomes noticeable when identifying party systems that are dominated by a single party. As is highlighted by Bogaards, definitions of these so-called ‘dominant party systems’ vary significantly in the party politics literature, depending on what classification
115 Sartori,
Parties and Party Systems, pp.108-109. In the case of centripetal competition,
positions taken by political parties gravitate towards the centre of the (‘left-right’) political spectrum; in the case of centrifugal competition, parties are less attracted to the centre and take more extreme positions. See Ware, Political Parties and Party Systems, p.170.
116 Duverger,
or counting rule is used.117 Coleman, for instance, applies a relatively high threshold and argues that in a ‘one-party dominant system’ the dominant party holds more than 70% of the seats and the opposition is fragmented.118 In Ware’s classification, a predominant party system is characterized by “more than one relevant party but only one party ever controls the legislature”.119 Blondel characterizes a multi-party system with a dominant party as having “one very large party (…) [obtaining] about 40 per cent of the electorate and generally [gaining] about twice as many votes as the second party”.120 Pempel, finally, argues that we can speak of dominant party rule if there is “electoral dominance for an uninterrupted and prolonged period, dominance in the formation of governments, and dominance in determining the public agenda”.121
The problem with these definitions is that the reasons for a particular vote or seat threshold remain vague and arbitrary. They also say little about the relation between threshold and the power balance between the different political parties. In addition, it is not always clear what the unit of analysis is; the individual party or the party system? Sartori’s definition, on the other hand, addresses most of these problems. According to him, in a dominant party system there is only one relevant party that has won an absolute majority of seats in parliament over at
117 See Bogaards, “Counting Parties and Identifying Dominant Party Systems in Africa”, pp.174-
179 for a comparative analysis of the different definitions of dominant parties and dominant party systems.
118 Coleman, “The Politics of Sub-Saharan Africa”. 119 Ware,
Political Parties and Party Systems, p.159 and p.162.
120 Blondel, “Party Systems and Patterns of Government in Western Democracies”, p.186. 121 Pempel,
Uncommon Democracies: The One-Party Dominant Regimes, quoted in Giliomee
least three consecutive elections.122 The absolute majority criterion indicates the dominant party’s power to govern alone, while the time requirement reflects its endurance. The main advantage of Sartori’s definition, however, is that it allows for a further specification of different types of dominant party system, including (consolidated) predominant and hegemonic party systems as well as (unconsolidated) dominant-authoritarian and dominant non-authoritarian systems.
Consolidated Party Systems
Recognizing that there are major differences in the nature of party systems across the world, Sartori proposes two different typologies: one for what he terms “formed states with consolidated party systems”, and another for “fluid polities with unconsolidated party systems”. Despite the differences in consolidation, the two typologies are closely related. Sartori even assumes that the various classes within the unconsolidated party system typology will crystallise into the corresponding party systems of the consolidated typology. For our purposes it is useful to briefly review both of them. We start with the typology for consolidated – i.e. moderately to highly institutionalized – party systems, which applies to
122 Sartori,
Parties and Party Systems, 171-177; Bogaards adds that Sartori’s counting rules
mainly apply “to parliamentary systems in which the government rests on parliamentary majorities”. According to him, in presidential systems “the parties that count are simply the ones that make a difference in helping (or obstructing) the president’s election, and that determine his having (or not having) a majority support in the legislative assemblies”. The latter addition is of particular relevance to post-conflict developing countries, many of which have presidential forms of government as we shall see in chapter 3. Bogaards, “Counting Parties and Identifying Dominant Party Systems in Africa”, p.175. and p.184, quoting Sartori, Comparative Constitutional Engineering: An Inquiry Into Structures, Incentives and Outcomes. p.34.
Western established democracies plus a few other countries with “solidly entrenched mass parties”.123
Based on the above-described rules for establishing the relevance (‘r’) of parties, Sartori proposes a classification with seven different ‘types’ of party system (see box 2.1 below). The first main distinction being made is whether the party system is competitive or not. In competitive systems, there are no restrictions on the contestation between different political parties, and “electoral contests are unrestrained and bitterly fought”. In non-competitive systems on the other hand, there is no real contestation for power between political parties. In Sartori’s words, “(…) competition ends, and non-competition begins, wherever contestants and opponents are deprived of equal rights, impeded, menaced, frightened, and eventually punished for daring to speak up”.124 In practice, the distinction between competitive and non-competitive systems largely corresponds with the distinction made between democratic and non-democratic or authoritarian regimes. This is one of the reasons why regime type is used by many scholars as an indicator of, or even one of the explanatory variables for the nature of a party system.
123 Sartori,
Parties and Party Systems, p.217.
124 Sartori,
Box 2.1. Classification for Consolidated Party Systems125
Non-competitive party systems
One-party (r = 1) Hegemonic party (r = 1)
Competitive party systems
Predominant party (r = 1) Two party (r = 2)
Limited (moderate) pluralism (r = 3 – 5) Extreme (polarized) pluralism (r = 6 – 8)
Atomised: (r = > 9)
In one-party, hegemonic party and predominant party systems, there is really only one relevant political party (r = 1). But because open party competition is prohibited in one-party systems and strictly controlled in hegemonic party systems, these first two systems are essentially non-competitive whereas predominant party systems are (slightly) more competitive. One-party systems,
where there is only one official party, include countries such as China, Vietnam, the former Soviet Union and former Yugoslavia.
In hegemonic party systems, there are usually other minor parties in addition
to the ruling or hegemonic party. However, the hegemonic party does usually not allow for an unrestrained electoral competition and actively makes it difficult for other parties to contest its dominant power position. In one-party and hegemonic party systems the level of party fragmentation is usually rather low due to the very limited number of relevant parties.126 Although it is impossible to characterize the ideological distance of a party system when it only has one
125 Sartori,
Parties and Party Systems, p.110. ‘r’ refers to the number of relevant parties, i.e. those
parties having either coalition or blackmail potential.
126 According to Sartori, a party system is fragmented “only when it has many parties, none of
political party, the ideological content of these party systems may vary.127 For the purposes of this study this subdivision is of lesser importance, however.
The category of competitive party systems is divided into five different types. In a predominant party system there are various parties that exist and
regularly compete for power, but there is one party that manages to win an absolute majority (more than 50% of the seats in parliament) during at least three consecutive elections. The key difference between a predominant party system and the earlier-mentioned hegemonic party system lies in the nature of the regime. In case dominance is achieved through competitive (democratic) elections, it is a predominant party system. However, if dominance is “achieved through political repression and the denial of civic rights to actual or potential competitors”, it clearly belongs to the hegemonic variant.128 Research has shown that hegemonic and/or predominant party systems are relatively rare in industrialized and semi-developed (industrializing) countries.129 Mexico (with the Institutional Revolutionary Party between 1929 and 2000), Japan (with the Liberal Democratic Party between 1955 and 1993) and Sweden (with the Social Democratic Labour Party between 1951-1993) are among the few countries that have had dominant party systems. This stands in sharp contrast with less developed, particularly post-war countries. As we will see in the next chapter, there dominant party systems are much more common.
127 Sartori distinguishes between one-party totalitarian, one-party authoritarian, one-party
pragmatic, ideological-hegemonic and pragmatic-hegemonic systems. See Sartori, Parties and Party Systems, p.197 and p.205.
128 Lewis, “Party States and State Parties”, p.477. 129 Pempel,
When there are two major relevant parties, with one having a parliamentary majority large enough to govern alone, we speak of a two-party system. The main
distinction between this and a predominant party system is that the margin in seat share between the two largest parties is so small that the expectation of governmental turnover is very real.130 In other words, in a two party system there are two major parties alternating in power. Only if one party rules for more than three consecutive elections, the system becomes predominant. In addition, it is important to realize that there are usually more than two political parties in a two-party system. However, the other parties are usually much smaller and not strong enough to rule by themselves. In case a third minor party is able to draw a significant share of the vote it is sometimes described as a two-and-a-half party system.131
In the next two types of party system there is no single party that dominates the party interactions or can control government because of its seat share. Instead, there is a multiplicity of smaller independent political parties that have to create coalitions in order to create a parliamentary majority or form a government. In a situation of limited pluralism there are between three and five
political parties that compete for power. This is the situation in countries like Belgium and Germany. Where there are more than five relevant political parties the system is usually described as extreme pluralism. The party systems of the
Netherlands, Switzerland, and Italy are cases in point. Factoring in the
130 Sartori,
Parties and Party Systems, p.165.
131 Blondel, “Party Systems and Patterns of Government in Western Democracies”. Classic
examples of two-party systems are the United States, New Zealand and Britain (until 1970). Because of the electoral strength of a third party, Britain now more resembles a two-and-a-half party system.
ideological distance between the multiple political parties, Sartori further refines this class by distinguishing between ideologically moderate pluralism (fragmented, but not polarized) and ideologically polarized pluralism (fragmented as well as polarized). The latter, polarized pluralist class is characterized by a number of aspects.
First, there are anti-system parties that contest the legitimacy of the regime and the system of government. Also, there are multiple opposition parties at different ends of the political spectrum, making a unified opposition joining forces against a governing party extremely difficult. Third, the system is multipolar, having not only a clear left and right-wing party, but also a party or a group of parties occupying the political centre. In addition, the system is polarized because of its large ideological distance. In other words, the two main parties are “literally two poles apart, and the difference between them covers a maximum spread of opinion”.132 Fifth, as the political centre is occupied by weak parties –or at least parties with a weak ideological message– while the parties at the extreme ends are much stronger, votes tend to go to one or both of the extremes (centrifugalism). As a result, the fundamental (ideological) differences between parties are so big that there is almost no room for