Sjoerd R.A. van Lierop (4092716)
Radboud University Nijmegen
Augustus 2013
A Comparative Study on
Party Family Change.
Have Party Families
Changed the System?
Chapter 0. Contents
I. Abstract 3
II. Acknowledgements 4
III. List of Figures 5
IV. List of Abbreviations 6
V. Contents of the Appendix 7
Chapter 1. Frozen or Not Frozen? 9
1.1 The “Freezing” claim 9
1.2 The Empirical problem 10
1.3 The Missing Element 12
1.3.1 Political Relevance 12
1.3.2 Scientific Relevance 13
1.4 Questions that Need to be Answered 14
1.5 The Structure of the Thesis 15
Chapter 2. Party System Change 17
2.1 What is a Party System and What is Party System Change? 17
2.2 Approaches to Party System Change 18
2.2.1 Number of Parties 19
2.2.2 Organizational Style 20
2.2.3 Competition Style 21
2.2.4 Cleavage Structure 22
2.2.5 Electoral Volatility 24
2.3 The Scientific Gap 25
Chapter 3. Party Family Change 29
3.1 What is a Party Family? 29
3.2.1 Social Democratic Party Family 33
3.2.2 Christian Democratic Party Family 36
3.2.3 Liberal Party Family 41
3.3 Hypotheses and Expectations 44
Chapter 4. Methodological Choices 47
4.1 Previous Research and Methods 47
4.2 Methodological Choices 51
4.3 Hypotheses Unraveled 52
4.4 Constructing Variables 55
Chapter 5. Tests and Results 59
V.1 Ideological Issue-exchange 59
5.1.1 The Social Democratic Party Family 59
5.1.2 The Christian Democratic Party Family 63
5.1.3 The Liberal Party Family 67
5.2 Individual Party Families or Intertwining Ideological Preferences 72
5.2.1 Issue Salience through Top-5 Comparison 72
5.2.2 Left-Right Scatterplots 86
Chapter 6. Conclusion and Discussion 93
6.1 Ideological Issue-exchange: the Conclusion 93
6.2 Individual Party Families or
Intertwining Ideological Preferences: the Conclusion 95
6.3 No Party Family Change? But What Then? 96
Chapter 7. References 99
I.
Abstract
In the 1960s Lipset and Rokkan established the now well-known argument that party systems have frozen. Even though there are sources that validate this statement, there are more that say party systems have not frozen. Looking into the phenomenon of party system change, this thesis explores a new method of measuring whether party system change has occurred. Based on the existing approaches to party system change and the method of measuring, one concept seems to be under research, that of party families.
In the following thesis an attempt has been made to link the concept of party family change to the concept of party system change. Using the Comparative Manifesto Project three different methods where used to examine whether party families have changed and how this is linked to the idea of party system change.
By focusing on three traditional party families, it has been possible to examine change over a wider time span. Examining the post-war period (1945-1960) has revealed that specific party families did exist during this period. Comparing the post-war period to the contemporary period (1985-2000) by using three different methods, it is possible to conclude the extent to which party families have changed and in what kind of change they have experienced.
II.
Acknowledgements
During the lengthy process of writing this thesis a small group of people have helped me through the hard times and the good times. Not only have tutors helped me, but also family and friends. All have helped in their own way and I would like to thank everyone around me that believed in me and my research. Of this group there are a few I would like to pay special attention to.
To start off with the one person who I could ask anything at any given moment and who has been there from start to finish. I would like to thank Andrej Zaslove for his contribution to my thesis, in word and in thought. Thanks to his approach to my work, but also my mindset I have had the privilege to work aside a renowned author and appreciated teacher.
Within my family I would like to thank two people. Both have contributed allot of time and effort to get me to think about what I have written. For their relentless work and their patience with me, I would like to thank both my parents, Marjon Verhees and Fred van Lierop. By editing, reading, criticizing and encouraging me and my work they let me believe in myself and the quality of my work. I would especially like to thank them for their patience and their belief they had in me. Only because of their support and guidance did I see the light at the end of an extremely long tunnel.
Someone else that helped me see that same light is my partner Susanne Crommentuijn. Even though she admitted she could not contribute to the thesis when I started writing, she has helped in her own (sometimes mysterious) ways. Lifting me up when I was down and by putting a smile on my face, she also made this entire process possible.
Thanks to these people, writing this thesis was not only possible but in the end also very gratifying. Writing a thesis is something you cannot do by yourself. The people around you and their individual manners of helping you define the quality of the work. For their contributions I would like to thank Andrej, Marjon, Fred and Susanne.
III.
List of Figures
Chapters 1-4, 6-8* do not include figures. Figur e 1 PWSD 1945-1960 Average Figur e 2 PWSD 1985-2000 Average Figur e 3 ConSD 1945-1960 Average Figur e 4 ConSD 1985-2000 Average Figur e
5 Change issue importance PWSD Figur
e
6 Change issue importance ConSD Figur e 7 PWCD 1945-1960 Average Figur e 8 PWCD 1985-2000 Average Figur e 9 ConCD 1945-1960 Average Figur e 10 ConCD 1985-2000 Average Figur e
11 Change issue importance PWCD Figur
e
12 Change issue importance ConCD Figur e 13 PWLib 1945-1960 Average Figur e 14 PWLib 1985-2000 Average Figur e 15 ConLib 1945-1960 Average Figur e 16 ConLib 1985-2000 Average Figur e
17 Change issue importance PWLib Figur
e
18 Change issue importance ConLib Figur
e
19 Issue preference per party family based on literature Figur
e
20 Issue preference Austria based on CMP Figur
e
21 Issue preference Denmark based on CMP Figur
e
22 Issue preference France based on CMP Figur
e
23 Issue preference Germany based on CMP Figur 24 Issue preference Great Britain based on CMP
e Figur e
25 Issue preference Italy based on CMP Figur
e
26 Issue preference The Netherlands based on CMP Figur
e
27 Issue preference Sweden based on CMP Figur
e
28 Left-Right score per election per party family Austria Figur
e
29 Left-Right score per election per party family Denmark Figur
e
30 Left-Right score per election per party family France Figur
e
31 Left-Right score per election per party family Germany Figur
e
32 Left-Right score per election per party family Great Britain Figur
e
33 Left-Right score per election per party family Italy Figur
e
34 Left-Right score per election per party family The Netherlands Figur
e
35 Left-Right score per election per party family Sweden
* = Figures included in the appendix have not been added to this list, these have been mentioned in the table of contents concerning the appendix in section V.
IV.
List of Abbreviations
CD Christian democrats / Christian democratic party family CDA Christian Democratic Appeal (Netherlands)
CDU Christian Democratic Union (Germany)
CMP Comparative Manifesto Project
Con Contemporary period / Contemporary variables CSU Christian Social Union (Germany)
DC Christian Democratic Party (Italy) KOK National Coalition Party (Norway) Lib Liberals / Liberal party family
L-R Left-Right
MRG Manifesto Research Group
OECD Organization for Economic Co-operation and Development PW Post-war period / Post-war variables
SD Social democrats / Social democratic party family
PWSD Post-war Social democratic issues / Post-war Social democratic variable (also possible with Christian democracy (CD) or liberal party families (Lib)) CONSD Contemporary Social democratic issues / Contemporary Social democratic
variable (also possible with Christian democracy (CD) or liberal party families (Lib))
RRRS French party Radical Republican and Radical Socialist UDF Union for French Democracy (France)
V.
Contents of the Appendix
Page
8.1 Social Democratic Party Family Figure 105
8.2 Christian Democratic Party Family Figure 106
8.3 Liberal Party Family Figure 107
8.4 Constructing the Post-war and Contemporary Social Democratic Variable 108 8.5 Constructing the Post-war and Contemporary Christian Democratic Variable 110 8.6 Constructing the Post-war and Contemporary Liberal Variable 112 8.7 Data Bar Charts Hypothesis 1 Social Democratic Party Family 114 8.8 Data Bar Charts Hypothesis 1 Christian Democratic Party Family 117
8.9 Data Bar Charts Hypothesis 1 Liberal Party Family 120
8.10 Bar Chart PWSD 1945-1960 123
8.11 Bar Chart ConSD 1945-1960 123
8.12 Bar Chart PWSD 1985-2000 124
8.13 Bar Chart ConSD 1985-2000 124
8.14 Bar Chart PWCD 1945-1960 125
8.15 Bar Chart ConCD 1945-1960 125
8.16 Bar Chart PWCD 1985-2000 126
8.17 Bar Chart ConCD 1985-2000 126
8.18 Bar Chart PWLib 1945-1960 127
8.19 Bar Chart ConLib 1945-1960 127
8.20 Bar Chart PWLib 1985-2000 128
8.21 Bar Chart ConLib 1985-2000 128
8.22 Data Top-5 Lists Hypothesis 2 Per Domain (Austria) 129
8.23 Data Top-5 Lists Hypothesis 2 Per Domain (Denmark) 130
8.24 Data Top-5 Lists Hypothesis 2 Per Domain (France) 131
8.25 Data Top-5 Lists Hypothesis 2 Per Domain (Germany) 132
8.26 Data Top-5 Lists Hypothesis 2 Per Domain (Great Britain) 133
8.27 Data Top-5 Lists Hypothesis 2 Per Domain (Italy) 134
8.28 Data Top-5 Lists Hypothesis 2 Per Domain (The Netherlands) 135
8.29 Data Top-5 Lists Hypothesis 2 Per Domain (Sweden) 136
Chapter 1. Frozen or Not Frozen?
Even though many believe we cannot speak of laws and structures in social sciences, the thesis on ‘the freezing of party systems’, which had first been promulgated, as it were, by S.M. Lipset and Stein Rokkan almost fifty years ago, has features of a scientific law. Even now it continues to be one of the most familiar and most frequently cited theses within the field of comparative (party) studies: ‘The party systems of the 1960s,’ wrote Lipset and Rokkan (1967: 50), ‘reflect, with few but significant exceptions, the cleavage structures of the 1920s…[T]he party alternatives, and in remarkably many cases the party organizations, are older than the majorities of the national electorates.’
Subsequent discussions of the freezing hypothesis have taken a lot for granted. Since no ‘freezing law’ had ever really been elaborated and since there was no actual theory “which could be formally tested and/or challenged” (Mair 1997: 4), much of the discussion of the original Lipset and Rokkan analysis has tended instead to revolve around questions of its continued empirical validity.
At first sight this thesis is also guilty of trying to explain change instead of examining the assumed stability. This is not a thesis about parties and party systems as such. Rather, it is about party system change with the focus resting primarily on how this question can be approached and interpreted in a new way. The thesis focuses on party families and party family change. Even though research on party system change has been done by many other scholars in the discipline of comparative politics, this thesis explores a new perspective on the matter of party system change and/or stability. Apart from the “freezing”-claim (Lipset and Rokkan 1967; Smith 1988; Mair 1974), authors have argued that party system change has occurred and is due to change in the number of parties (Mair 1997; Sartori 1976), change in the organizational style (Mair 1997; Kirchheimer 1966), change in the style of competition (Mair 1997; Dahl 1966; Sartori 1976) or is due to electoral volatility (Mair 2008; Bovens and Wille 2011). The entire phenomenon of party system change can still be debated, and for that reason it is interesting to look at other approaches measuring party system change.
1.1 The “Freezing” claim
The central question of this thesis is: have party systems changed? A sub question, perhaps more methodological, is how can we measure this change? According to Lipset and Rokkan (1967: 50), the freezing of the party systems was due to the freezing of the cleavage structures of the 1920s. This freezing-thesis has been acknowledged by many political scientists. Three years later, Rose and Urwin (1970: 295) found robust empirical support for these observations: “[T]he electoral strength of most parties in Western nations … has changed very little from election to election, from decade to decade, or within the lifespan of a generation”. More than 20 years after these observations, Mair (1993: 132) gave support to this by claiming that the electoral balance in the early 1990s was no different from that thirty years earlier. In general, Mair maintained, electorates were not more volatile than before.
Despite challenges from new parties, most of the early political parties in mature democracies had managed to remain in powerful positions. Quite recently, Golder (2003) confirmed that the pattern of stability, in general, still holds. Based on this, it is fair to conclude that there are mixed results when it comes to party system change. It is for that discrepancy that researching party system change and approaches of measuring it are so interesting.
1.2 The Empirical Problem
Already by the 1980s, the signs of what might be interpreted as change were apparent in the political world (Caljé and Hollander 1990: 324). Evidence came from the European periphery, with the gradual decay of Fianna Fáil predominance in Ireland, and with the sudden emergence of a new party, the Progressive Democrats which, in its first electoral outing in 1987, became the third largest party in the system. There was evidence from Britain (Caljé and Hollander 1990: 329), with the long-term erosion of Labor and the concomitant growth of third-party support in the form of, however temporarily, the Liberal/SDP Alliance. In France, the first left-wing government in the history of the Fifth Republic was elected in 1981 and later returned to office in 1988; the 1980s in France had also been characterized by the continued marginalization of the Communists and by the re-emergence of an extreme right party (Caljé and Hollander 1990: 332). Italy in the 1980s witnessed the appointment of the first non-Christian Democratic premiers since World War II and, more recently, recorded relatively major shift in support from the Communist Party to the Socialists and the Ecologists (Caljé and Hollander 1990: 336). Ecologists also managed to win parliamentary representation in Austria, with further change being evident in a doubling of the right-wing Liberal Party vote in the important election of 1986 (Caljé and Hollander 1990: 341). In West Germany, the Social Democrats continued to poll less than 40 percent of the vote and found themselves facing the increasingly vibrant challenge of the Greens, who now command almost 10 percent of the seats in the Bundestag (Caljé and Hollander 1990: 345). In Belgium the linguistic divisions in all three major parties quickly solidified, the ruptures in the Christian Social Party and in the Liberals being paralleled by the division in the Socialist Party in 1978. The late 1970s in the Netherlands also witnessed a major change, with the merger of the three traditional denominational parties into the more broadly-based Christian Democratic Appeal (Caljé and Hollander 1990: 349). In Denmark, despite the relative peace which has followed the ‘earthquake’ election of 1973, the 1980s witnessed substantial growth in support for both the Conservatives and the Socialist People’s Party (Caljé and Hollander 1990: 351). Conservative growth has also been evident in Norway and Finland with, in the latter case, the return of the National Coalition Party KOK to government in 1987 for the first time in more than twenty years. In Sweden, the 1980s have witnessed the continued challenge to the long-term dominance of the Social Democrats, which lost office in 1976, and which then faced a new challenge from the Green Party, which first entered parliament in 1988 (Caljé and Hollander 1990: 356).
One thread which can be identified in recent years involves the growth of ecology parties, which have now gained representation in the parliaments of Austria, Belgium, Finland, Germany, Italy, Switzerland, Sweden, and the Netherlands (Mair 1997: 51). A second thread is the apparent decline of the traditional class left, or at least, its re-equilibrium following the erosion of support for significant components of the left (Mair 1997: 53). A third thread, discerned with perhaps more difficulty and dispersed less evenly, is the challenge to the Christian parties (Mair 1997: 59). The pan-Christian merger in the Netherlands, for example, represented an unabashed attempt to restore some electoral credibility to the fading fortunes of the three constituent denominational parties.
Even on higher level, there are deep-rooted changes in mass worldviews effecting and reshaping economic, political and social life (Inglehart 1997: 4). Throughout advanced industrial society, freedom of expression and political participation are becoming increasingly important to a growing share of the public. Inglehart (1997) concludes that the most of the changes in economic, political and social life are based on the assumption that modernization of so many different aspects change world values into (Western) post-modernism. Modernization is, above all, a process that increases the economic and political capabilities of a society: it increases economic capabilities through industrialization, and political capabilities through bureaucratization.
In addition, drawing on the extensive country studies in their volume, Dalton and his colleagues (1984) find evidence of change in the weakening of alignments, increased fragmentation, and increased volatility. This volatility grew, according to Mair (2008), because of the relative openness of party systems. In his review of the Dutch party system, he concluded that it was the party system that made volatility one of the leading instabilities in the electoral base. Of course, it could be argued that the current wave of electoral instability in the Netherlands is far from exceptional and is simply part of a wider European trend. Indeed in Europe, or at least within the older European Union area, it can be argued that all polities and hence all electorates now share more or less the same political experiences, and that volatility in one political system is likely to quickly echo in another. Certainly, the established parties in the Netherlands are now less popular with the voters, and seem less able to engage their political interests or attention. But this is true throughout the older European democracies in the early 21st century, in that party organizational networks in almost all of these polities are now withering away.
The question is, however, do these changes amount to party system change? In part, the answer to this question depends on how change is measured: is it simply the increase in the number of parties, changing electoral allegiances, or perhaps is it when parties change their ideology?
1.3 The Missing Element
Examining the literature on party system and party system change it becomes clear that there is no clear answer: scholar come to different conclusions. The contemporary political arena is changing and has been changing from the moment that the freezing-claim was proclaimed. Many schools of thought have found different forms of measurement by which this change would have occurred. However, does this constitute true party system change?
1.3.1 Political Relevance
Analyzing the existing schools of thought it has become clear that there is still a missing element. Not just any element, but a very important element within comparative politics. As will be argued in the following section, even though the concept of party families is well research within comparative politics, it is not been readily linked with party system change and/or stability. In the following chapters it will become clear that ideology in the form of the party family is another possible method of measuring party system change.
This concept is of importance because political parties are often categorized in terms of their ideologies. Even though this characteristic is a very specific and personal characteristic for the party, it links parties on a cross-national level. Not only because, for example, France, the Netherlands and Germany all have a Christian-democratic party, but also because the international integration of parties has been forming an alliance on European level, where the European Parliament has its own Christian-democratic party, formed by national party members (Mudde and Mair: 1997).
On a national level the notions of research on party system change and party family change can show what parties can expect within the political arena. Political parties are formed around certain ideologies and these ideologies take form in the political arena through a political party. Once an ideology is covered we would expect that other parties would not be interested in these ideological issues. Because parties rise and fall not only in size but also in number it can be expected something happens to these parties and the entire system. If we were to establish a link between the issue coverage of some or all party families and the change in party systems, political actors can than map out or plan their actions. If party system stability is a fact and we see party families have either shifted to the left, to the center or to the right, this would mean parties can shift ideologically whatever they want but a party system will most likely influence their political actions instead of their ideological preference.
As example, let us take a two party system much like Great Britain. We have a leftist party family and a rightest party family. With two party families it will be most likely that this country we have in mind has two parties. These two parties have a certain coherent set of policies during a period, say 50 years ago. A new political issue arises and seems to become more important within society. A new political
party is trying to enter the political arena. The two older parties can two different things: (1) the older parties incorporate the new party within the party system which means it is no longer a two party system, or (2) the older parties incorporate the new issues into their own set of policies which means this new party is superfluous.
If parties would have any idea of how to cope with party family change and party system change, new parties would be able to calculate the perfect time to enter the political arena. For older parties it would become easier to predict when they would have to change their policy preferences or to stay on the track they are following already.
1.3.2 Scientific Relevance
When looking at party system change problems arise. One problem that needs to be indicated: the distinction between party change and party system change. With this the next problem rises, the allied problem of when precisely the former also implies to the latter. Many authors have struggled working on a clear demarcation (Mair 1997: 21). Since changes in specific aspects of parties are a permanent feature of the political landscape and since the process of party adaptation is also continuous it may be tempting to conclude that party change is so pervasive as to be almost irrelevant. Because some argue that change is a constant factor in party systems, it will be necessary to look at larger patterns. To do so, one needs to keep in mind that party families are in this case an important factor that need not be forgotten.
A party system clearly involves something more than the sum of its component (party) parts, and incorporates some element of understanding of the mode of interaction between these parties. Thus, Mair concludes (1997: 51), party system change occurs when a party system is transformed from one class or type of party system into another. Party system change might therefore occur when, as a result of ideological, strategic, or electoral shifts, there is a transformation of the direction of competition or the governing formula (Mair 1997).
This leads to the following problem. The moment we associate the freezing process with the stabilization of a particular social structure, for example, or with the consolidation of a particular type of organizational intervention, the risk of dissolving the problem of the freezing of party systems into that of the freezing of individual parties or groups of parties arises. In other words, we risk forgetting about the system as a whole and devote most of our effort instead to an understanding of its component parts. In this sense it is the system which makes for the parties as much as it is the parties which make for the system.
Still, in the perspectives of change listed above (ideological, strategic, or electoral shifts) there seems to be a missing concept or component that comparative scientists regularly use. This concept is the party family and can be introduced into the debate because it focuses on policy appeals, ideology and
party comparability. In more than one way it can be linked to the idea of cleavages, but it supersedes it by relating parties from different countries in a cross-national comparisons.
Where Lipset, Rokkan and Kitschelt regard the importance of cleavage and in some measure ideology, Mair argues that change must not be looked at from outside anything but the party. According to Mair parties are in a crisis and not ideology or class (Mair 1997: 31). By saying this, Mair comes to the conclusion that neither ideology nor class have had difficulties coping with any form of change. It is the party that is having trouble coping with change. Parties, says Mair, have stagnated in their development and have parted their relation with society and politics almost as a whole (Mair 1997: 33). Mair links this apparent crisis of the party with that of the party family. Due to the fact not ideology, but the party as such is in crisis it is unnecessary, if even impossible, to measure either party ideology or party family coherence. When looking at party families, Mair is incorrect. I argue that party family consistency and/or party family crystallization can give more insight into party system change. Looking at the different party families and trying to find out whether ideology has changed, it can become clear whether party families can and do influence party systems in such a severe manner we can speak of party system change.
1.4 Questions that Need to be Answered
Because of the size and relevance of the puzzle in this thesis it demands a multi-facetted research and therefore needs to be divided into different sub-puzzles or questions. This thesis will focus on three different questions:
I. To what extent was there party alignment between traditional party families and certain issues between 1945 and 1960?
Because Lipset and Rokkan argued that party systems where long frozen, it is necessary to evaluate the existence and consistency of party families over Western Europe. This thesis will proceed by analyzing parties between 1945 and 1960 to see if traditional party families existed during this period.
II. To what extent is there party alignment between traditional party families and certain issues between 1985 and 2000?
Many authors have argued that party systems have changed, and so it is necessary that we include a second moment in time to make a comparison possible between these two periods. By using a more recent point in time, changes that have occurred can clearly be noticed while compared. Therefore this thesis will proceed by analyzing party families between 1985 and 2000.
III. To what extent can we speak of traditional party families and consistency between 1945 to 2000, and how can this measure party system change?
Because a purely spatial research would be incomplete, a comparison of the two moments in time is needed. This enhances the power of the argument when we can conclude that party systems have indeed changed and that party families have in some way contributed to this process. In comparing both periods it will become clear not only if change has actually occurred but also in what direction. Answering this question will not only link the two periods of time looking into the party families and possible change, but it will also enable us to link the concept of party family and party family change to the concept of party system change.
1.5 The Structure of this Thesis
The next chapter will focus on the concepts of party systems, party system change and the different existing approaches to the measurement of party system change (Chapter 2). Chapter 3 will explain what concepts and definitions are needed for the argument of party families. In this chapter the focus lies on constructing sets of issues portraying the essence of the individual party families under investigation in this thesis covering each time period separately. After having completing this, hypotheses will be formulated. This chapter will be followed by a methodological chapter in which the methods used for the argument will be explained step by step (Chapter 4). After having explained how the methods will be used, the results of these methods will be set out. The chapter will focus on the Comparative Manifesto Project (CMP) database and results following from that analysis (Chapter 5). The final chapter will focus on these conclusions towards the argument and will focus on what scientist can do in the future with these results (Chapter 6).
Chapter 2. Party System Change
2.1 What is a Party System and What is Party System Change?
Political parties competing with each other for elective office and control of government form a party system (Mair 1997: 6). Party systems have been a key factor in the study of political parties and more broadly in comparative analysis. In this definition lies the relationship between party systems and parties. As such, the relationship between party systems and party families is not that far off. Party families, as will be portrayed in the next chapter, are formed once certain characteristics between parties, based on ideology, can be categorized. When looking at a party system it is not only interesting to look at political parties as individuals, but also looking at political party families. In doing so it might be stronger to define a party system as a system that consists of regular and recurring interactions among its component parts which can thus consist of parties or party families.
Although the term ‘party system’ had already been in use for a long time, one of the first ‘systemic’ uses of the term can be found in Duverger’s ‘Political Parties’. Duverger (1954: 203) argues that: “With the exception of the single-party states, several parties co-exist in each country: the forms and modes of their coexistence define the party system of the particular country being considered. In addition to characteristics of the parties, these include new elements that do not exist for each party community considered in isolation: numbers, respective sizes, alliances, geographical localization, political distribution and so on.” A party system is defined by a particular relationship amongst all these characteristics (Durverger 1954: 203).
Party systems have a number of distinct features which arise from electoral competition and parties’ relation to each other (Durverger 1954: 189). These include the number of parties contesting elections and winning legislative seats, their relative size and strength, the number of dimensions on which they compete, the distance which separates them on key issues, and their willingness to work with each other in government formation and the process of governing (Lijphart 1982). Voters, politicians, and political analysts often think of parties divided along a Left-Right spectrum, but it is not unusual for party systems, at least in their origins, to reflect multiple dimensions of conflict (Lijphart 1982). Party systems can be more or less polarized on any or all of these dimensions. Other features on which party systems may differ include the degree to which their competition for government is open to all parties or closed and the degree to which the party system itself is institutionalized or entrenched. However, this latter facet reflects not so much the ways in which parties relate to each other, as the degree to which parties, taken together, are able to enlist durable support and structure the electorate (Lijphart 1982).
In following Caramani (2007: 319), an important element of the competitive interaction between parties is the shape of party systems. Here, even though this does not look like it on first sight, the
focus lies on the competition between its parts. The two main elements of the morphology of party systems are: (1) the number of competing units, that is, parties, and (2) the size of these units. It is important to distinguish types of party systems. In using Caramani’s definition as well as Mair’s definition, party systems exist of political parties competing with each other for elective office and control of government, with regard to number of and size of the competing units.
This leads to the distinction of six party systems: (1) single-party systems, (2) hegemonic-party systems, (3) dominant-party systems, (4) two-party systems, (5) multi-party systems, and (6) bipolar-party systems (Caramani 2007: 321). Due to the lack of democracy in the first and second bipolar-party system types, these will be omitted from this thesis.
The key problem with the phenomenon of party system change is that it is seen as either happening all the time or scarcely happening at all. Even though this might be a debate surrounding the concept it is assumed change happens. The change that is interesting within party systems is the change that focuses on the core of the party system. Change that occurs on the margins of a party system, if it is even possible to be specified, is not of interest. For that reason, the terms of reference need to be made clear.
The core of any party system qua system is constituted by the structure of competition for control of the executive (Caramani 2007: 325). It then follows that a party system changes when there is a change in the structure of the system (Caramani 2007: 329). The term structure is defined as structure of competition by Mair, but this is widely debated. Still, this is an important notion. It can be broken down into three different components (Caramani 2007: 334). The first of these identifies change in the change of the prevailing pattern of alternation in government. The second component refers to the extent to which the governing alternatives in the system prove stable or consistent over time, or whether they involve innovative formulae. The third component refers to the question of who governs, and to the extent to which access to government is either open to a wide range of diverse parties or limited to a smaller subset of established governing parties.
2.2 Approaches to Party System Change
There are some hundred states that display – at least on paper – some kind of party arrangement. The variety of these arrangements is equally impressive. How are we to order the maze? This maze is made even more difficult by the perspective on party system change. Different approaches seem to argue that their answer – their tool of measurement – is the one that measures party system change best. In the following section these different perspectives will be looked at. Clearly these approaches are based on certain phenomena that are used as counter evidence for the freezing-claim. The question remains what evidence they use and which method they use. Even though these approaches seem to be the correct path to follow according to their author’s, critiques can be formulated.
2.2.1 Number of Parties
The first thing that we can observe in a party system is the number of parties that exist within the system. Even now, the most simple and applied way for categorizing party systems is by looking at the number of parties. The traditional division was between a single-party party system, a two-party party system and a multi-party party system. Because of developments within party systems it seems that this simple to observe variable is one of the basic ways in measuring party system change.
For quite a long time party systems have been classified by counting the number of parties – whether one, two or more than two. Some time now, however, there has been a mounting wave of dissatisfaction with this approach. LaPalombara and Wiener (1966) proposed for the competitive party systems a fourfold classification which drops altogether the numerical base, precisely “on the assumption that the traditional distinction between two-party and multi-party patterns has not led to sufficiently meaningful insight” (1966). This can be readily granted; yet one may be equally dissatisfied with the alternative taxonomies proposed so far.
The trouble with the numerical criterion of classification is that we resort to a counting system without having counting rules (Mair 1997: 34). Paradoxical as this may seem, we are not even able to decide when ‘two is two’, whether a system is a two-party system or not. It is hardly surprising, therefore, that the number-of-parties approach gives uninspiring results. It could be argued that the traditional distinction between one, two and more-than-two party systems is hardly a classification at all – if we require a classification to be derived from a precisely stated criterion.
For one thing, the number of parties is a highly visible element. It immediately indicates an important feature of political systems, namely, the extent to which power is fragmented or not fragmented. Intelligent counting can also detect, however roughly, the relative distribution of power among the parties. Clearly, party numbers cannot be taken at their face value. This means that if we resort to counting we should know how to count. However, before entering this discussion a few preliminary remarks are in order. First of all, there is a difference between saying ‘this system contains two parties’, and ‘this system displays two-party properties’. Furthermore, the requirement of a number of parties base is not that the entire scheme should be based on a numerical criterion alone (Mair 1997: 39).
There is no absolute yardstick for assessing the relevance of size. What really weighs in the balance is the extent to which a party may be required as a coalition partner for one or more of the possible governmental majorities. A party may be electorally weak but have a strong coalition-bargaining power.
The foregoing implication has a limitation, for it applies only to the parties that are governing oriented and, furthermore, ideologically acceptable to the other coalition partners. This leaves out, or may leave
out, some relatively large parties of permanent opposition – such as anti-system parties. In summary, we can discount the parties that neither have (1) coalition potential, nor (2) blackmail potential. So far with counting. But intelligent counting requires us to establish, in addition, a classification of party systems in which the number of the parties is significantly related to the distribution of power between them. A power configuration can be such that one party ‘counts more’ than all the other parties together. This is notably the case with the predominant party systems; and this is why the primary characteristic retained by the taxonomy is not how many parties – whether only one or more – oppose the predominant party, but the existence of a stabilized situation of predominance. With regard to the predominant party system the relevant indication is, then, that one party holds the absolute majority of seats without being subjected to alternation (Sartori 1976).
Turning to Sartori’s (1976) own approach, for example, Mair suggests that “the appearance or disappearance of a party leads to party system change when the number of relevant parties is altered to an extent which shifts the party system in question from, say, the limited pluralism class to the extreme pluralism class” (Mair 1997: 52).
Mair believes that change in number of parties is sufficient for analyzing party system change. He claims that “the conventional distinction involved here has also proved appealingly straightforward: that between a two-party system on the one hand, and a multiparty system on the other (…), it was believed to tap into a more fundamental distinction between more or less stable and consensual democracies, which where those normally associated with the two-party type, as opposed to more or less unstable and conflictual democracies, which were those associated with the multiparty type” (Mair 1997: 200).
Apart from its simplicity, this approach is very unsatisfactory. Even though it is the main goal of a scientist to simplify reality, concluding that the change in number of parties leads to party system change is an over simplification of reality. As a complex institution within party politics the party system forms the base, the structure of the arena in which parties interact.
2.2.2 Organizational Style
Another approach is to focus on organizational style. Mair (1997: 41) has sought to emphasize the link between party organization or style of organizational intervention and electoral (de)stabilization.
To the extent that there is a specific crisis of party, this is likely to result in and possibly also emerge from changes in party organization and in the forms of linkage which tie parties to the mass electorate (Mair 1997). Mair argues we need to understand the nature of and the basis of party organizational diversity and change in contemporary Western Europe. On the one hand, as individual parties have moved from being relatively closed communities to more open, catch-all, and professionally driven structures, they have proved better equipped, at least in principle, to respond to shifts within the wider
society. No longer obliged to listen almost exclusively to their own distinct clienteles, and no longer burdened with the particularistic demands and associated with these clienteles, the parties behave more flexible and more adaptable (Mair 1997: 52). Organizational change is therefore based on evolution of the party within society. Even though this might be the case, the strange thing is that the party has been criticized these last few decades of ever more drifting away from the people they represent. On the one hand, all people seem to be dissatisfied with political parties. On the other hand, parties are more interested in speaking to the general public instead of a certain group.
Kirchheimer (1996) suspects that the emergence of the catch-all party is the greatest factor in the changing of party organizations. He gives a list of five characteristics for this phenomena: “a drastic reduction of the parties ideological baggage; a further strengthening of the top leadership groups; a downgrading of the role of the individual party member; a de-emphasis of the class gardée, specific social-class, or denominational clientele, in favor of recruiting voters from the population at large; and a process of securing access to a variety of interest groups” (Kirchheimer 1966: 190).
The danger in following in Kirchheimer and Mair’s footsteps is that we lose the party system as a structure. While changes in a party can be little, Kirchheimer believes that the larger change from fairly closed to relatively open party organizations has influenced all parties over Western Europe.
2.2.3 Competition Style
The classification of party systems now only rests on mere numerical division or on the individual organization of the party. Still, one needs to keep the manner in which parties interact with one another in mind. It has always been possible to gain a reasonably valuable insight into the ways in which these polities differ from one another. In the conclusion to his classic volume, Robert Dahl (1966) sought to move away from an almost exclusive concern with simply the numbers of parties and built an alternative classification based around the competitive strategy adopted by the opposing parties, distinguishing between competitive, cooperative, and coalescent strategies, and distinguishing further between opposition in the electoral arena and opposition in the parliamentary arena. This led Dahl to elaborate a fourfold typology, distinguishing between strictly competitive systems, co-operative-competitive systems, coalescent-competitive systems, and strictly coalescent systems.
Sartori also focused on the interparty competition as one of the most important factors in a party system (1976). Unlike the alternative approaches, it helps to focus attention directly on what is perhaps the most important aspect of party systems according to Sartori, and on what distinguishes most clearly between different party systems, and that is the structure of interparty competition and especially the competition for government.
Building on Sartori, then, how might differential patterns in the competition for government be understood? Three related factors are relevant here. First, there is the question of the prevailing pattern
of alternation in government in any given party system, and the extent to which this is either wholesale, partial or even non-existent (Sartori 1976: 94). This refers to the degree of change within the governing parties. The first type is where all governing parties are replaced by non-governing parties. The second type refers to the partial replacement of governing parties by non-governing parties. The final type is where there is no alternation at all. Secondly, there is the question of the stability or consistency of the governing alternatives, and the extent to which innovative formulae are adopted (Sartori 1976: 96). This refers to the extent that if alternation takes place, let’s say wholesale, that the new parties are ideologically different from the leaving parties. Finally, there is the simple question of who governs and the extent to which access to government is either open to a wide range of diverse parties or limited to a smaller subset of parties (Sartori 1976: 101).
This holds that the “structure of competition [can be seen as] the establishment of a language of politics in which one particular conflict is prioritized and which any potentially alternative alignment of forces is either absorbed or marginalized” (Mair 1997: 15).
In general, scientists argue it is clear that an understanding of party system change must focus on that which defines the system in the first place: the pattern of interactions among the parties. Electoral developments, ideological change, organizational revitalization, and so on, are all important aspects of party change, but they are appropriate indicators of party system change according to Mair (1997: 75) only if they begin to have a bearing on the pattern of interactions which. On the other hand, there is the evidence of the mobilization and success of new parties, which suggests that mass politics is no longer simply dominated by old parties, or by parties older than their national electorates.
Even though Mair believes, it is not only just the sheer size of the electoral market and the extent of electoral availability which are relevant, but also the degree to which competition itself matters. The existence of an electoral market may therefore be seen as a necessary if not a sufficient condition for there being party competition, in much the same way that democracy itself is a necessary but not sufficient condition for there being competition more generally.
2.2.4 Cleavage Structure
Even though Lipset and Rokkan (1967) concluded that there is a freezing of party systems, it continues to be one of the most familiar and most frequently cited theses within the field of comparative party studies, and also within the literature on the changing of party systems.
Here, change in the cleavage system is seen as symptomatic of change in the party system. The difficulties with this line of reasoning are manifest. The first and most evident difficulty has been highlighted by Gordon Smith (1988: 3), who addresses the link between cleavages and party system change, and who questions whether cleavages should actually be considered of systemic relevance. As he suggests, from what he admits to being a heretical position: “The cleavage structure relates to the
social make-up of support according to individual parties – not to the system, not that it is if we follow a definition based on interaction. Social cleavages, and changes in them, naturally do have important consequences for the system as a whole, but those effects are registered through the other dimensions: affecting the number and relative size of parties, the extent and intensity of polarization, the volatility of the party system.”
According to Lipset and Rokkan, mass politics in modern Europe has been structured by four major cleavages. This complex of cleavages, emerging in different forms as it did across Europe, largely accounted for the variation among the party systems which developed in the wake of universal suffrage, and, more significantly, established a set of parameters which proved remarkably enduring. The introduction of universal male suffrage at the beginning of the twentieth century proved the crucial catalyst which, as it were, froze these cleavage structures into place, institutionalizing a language of politics which was to prove more or less immutable. Thus, even in the 1960’s, when Lipset and Rokkan were writing, the cleavage structures were still all too apparent. Their analysis may have stopped with the 1920’s, but, as they argued, there was little reason to continue beyond that point.
Even though this thesis has law-like features, Mair (1974: 4) argues that the freezing of cleavage structures, and therefore, the freezing of party systems is no longer the case. According to Mair Lipset and Rokkan’s argument has lost its strength. The argument projected by Lipset and Rokkan would need to be turned around in this thesis. Accordingly, the goal of this thesis is to explain why change occurs.
Over the past decades, a number of crucial factors have been identified as leading to electoral instability, and a particular emphasis has been put on the impact of changes in the social structure and the blurring of traditional class boundaries and on the impact of new value systems and the mobilization of a new post materialistic cleavage. This is an observation against the freezing of the cleavages stated by Lipset and Rokkan, but it still do not overthrow their theory. Clearly much of the existing differences reflect the institutionalization of past bases of opinion cleavage; once formalized in political parties these cleavages have survived the decline or disappearance of the original social conflicts which gave rise to party divisions. The blurring of these demarcations, do not per se lead to a new and greater question. The question remains how cleavages influence party systems.
Next to the fact that it is not clear how cleavages influence party systems, this approach seems to be the most appropriate taking social factors in account. The down side about these social features is that the organizational style and the competition between parties are missing. Even though there is competition between the cleavages, and the social groups linked to these, it is no more than social groups. There are three criteria that make a group of people part of a cleavage. These groups need to know there is a difference between ‘us’ and ‘them’. Secondly, these groups need to form an
organization around this different identity. Finally, the individuals in these groups need to act according to these differences. All these criteria don’t purely focus on parties, but can also lead to groups organizing in lobby groups.
Even though socio-economic changes can lead to cleavage structures changing, this is too vague. Socio-economic changes can lead to that many different phenomena, that it would not be clear in how far this would also lead to party systems changing. For that reason, this approach would need to focus more on parties and the structure they thrive in.
2.2.5 Electoral Volatility
The impact of socio-economic changes on voting behavior, and the way in which these can then feed through into electoral realignment or dealignment, also serves to emphasize how the long-term survival of individual parties may be dependent upon stable social relations (Mair 2008: 235). What needs to be underlined here is that, concurrent with the evident instability of contemporary European electorates, other phenomena also exist that, in themselves, represent fundamental forces for continuity, and which therefore suggest the presence of a substantially more complex situation than is often taken to be the case (Mair 2008: 241).
There is, for instance, the impact of social change on organization itself in that the development of a more knowledgeable, well-informed, and competent electorate, together with the undeniable increase in the individualization and atomization of modern society, have together undermined the sense of collective solidarity which once served as the prerequisite for the traditional mass party (Bovens and Wille 2011: 24). More specifically, an emphasis on electoral change has usually been extended to imply change in party systems in general, and change in the cleavage structures in particular. First, taking the direct argument, it is suggested that electoral change leads to party system change only when it brings about a shift from one type of party system to another. Secondly, there is the indirect argument that electoral change is seen as symptomatic of change in the cleavage system and this, in turn, is seen to imply a transformation of the party system (Bovens and Wille 2011: 31). This argument is based on the evidence of trends in aggregate electoral volatility – the net shifts in votes from one election to the next. This individual electoral choice in constrained by the encapsulation of the mass of the electorate into one of the sub-cultural groups that the parties represent, so that electoral politics is less about differential rates of conversion than it is about differential rates of mobilization.
As being one of the most poorly looked at approaches, the electoral volatility-approach can be seen as an additive to other demand-side approaches (Mair 2008: 249). The upside of this approach is that it is one of the more easy approaches to research with a large specimen. The down-side is that many different variables can lead toward electoral volatility. The next step that the electoral volatility would have to make is towards party system change. Unfortunately for this approach, there seems to be no direct step towards that phenomenon. Because electoral volatility is about the individual votes of
people for an individual party within that system, it would be farfetched to say that these votes explain party system change.
2.3 The Scientific Gap
Where Lipset and Rokkan regard the importance of cleavage structure and in some measure ideology as important factors, Mair argues that change must not be looked at from outside the party. According to Mair parties are in a crisis and not ideology or class, but the party itself (Mair 1997: 31). Apparently it suffices to say that party systems and parties change due to parties are in crisis. This does not make the problem or the answer to the problem any clearer. What needs to be said at this moment is that any shift in number of parties, organization within parties, and competition style between parties, cleavage structure between blocks of parties or volatility within the electorate don’t seem to grasp the entire problem at hand.
When looking into the concept of party systems and how we measure change within these systems, it is the concept of party systems and its definition that show the importance of the notion of party families and especially ideology. Only by taking a closer look at Duverger’s definition, it becomes clear ideology is an important component. It will enable us to compare and categorize parties on a higher level, above the nation-state. This can be seen as a critique to all the existing methods arguing they have a method for measuring party system change.
When looking into the method of counting the number of parties in a party system we should account not only for the mathematical possibilities but also for the ideological possibilities of coalition. This last possibility, unfortunately, is lost when examining the approach in counting the number of parties. It becomes clear that the notion of ideology is lost in this method of approach. Ideology is seen as a characteristic that can be neglected, while it is one of the founding features of a party. If we leave this factor the same in all parties, then we couldn’t speak of a fair party system, because it would not make a difference whether there would be two or twenty parties.
In examining the method focusing on the organizational style of parties within a system the entire concept of the system is lost. Parties apparently have changed and it is clear that all parties have become more open, but this is predominantly because of technological and sociological changes in the communities. Still, this does not coincide with the changes made between, par example, Christian, i.e. religious parties and liberal or social-liberal parties. The notion of ideology is lost in this method of approach in a way that the number-of-parties approach also loses it. Ideology is seen as a characteristic that can be neglected, while it is one of the founding features of a party. If we leave this factor the same in all parties, then we couldn’t speak of a fair party system, because it would not make a difference whether there are open or closed parties.
In looking at the party system and its component parts using the competition style approach, ideology is a factor that plays a small role. It is not only the case that the concept of ideology is missing, we can also criticize the competition-approach that it is incomplete. For the same reason as criticizing the previous approach, this approach is committing the ecological fallacy. In this case the fallacy is less apparent, but where the former approach only uses individual features to explain a higher phenomenon such as party system change, this approach does not accept individual features enough to explain lower phenomenon such as party ideology change. Apart from the fact that this approach does not look at individual features, this approach also does not incorporate any social or political features on which the political parties are based upon.
In entering the scientific world of Lipset and Rokkan, it becomes clear that the method of approach concerning the cleavage structure is the approach most dear to the idea of ideology. Where other approaches miss this concept entirely, it is the idea of cleavage structure change enhancing party system change that closes in on the concept of ideology the best. What does need to be criticized is the fact that Lipset and Rokkan acknowledge that this approach is country specific. The cleavages do not cross borders which makes it difficult to analyze party system change on a cross-national scale. Apart from that it is also the case that, even if we were able to forget about country boundaries, cleavages focus more on the society and the ideologies within this society than the ideologies incorporated into the parties. With this it is meant that not all ideologies that have a base within society have been translated into political parties and maybe not even into political parties winning in elections.
The final approach, that of electoral volatility, is one of the most difficult approaches to link to the changing of party systems as a whole. Even though Mair concludes that the impact of socio-economic changes on voting behavior, and the way in which these can then feed through into electoral realignment or dealignment, also serves to emphasize how the long-term survival of individual parties may be dependent upon stable social relations, it is rather bold to conclude that such a seemingly small factor can influence such a large system. The change enhanced by electoral volatility is small, as Mair concluded, and therefore it is uncertain what direct links can be found between electoral volatility and party system change. This factor may start a larger effect, but it is not the leading factor that enhances party system change.
In the forthcoming it will be argued that change in party families can also lead to change in party systems. Due to the shifting of political issue preferences on party family basis, parties have changed their main set of issues through history. This perspective deserves more attention within the profession, because it rights the wrongs done by the other perspectives.
The main objection towards these previous perspectives, and what Mair warned for, is that they are either macro-explanations or micro-explanations. Each of these explanations can be put either in a supply-side or in a demand-side category. None of the former explanations takes both the structure and
the society in mind. With the view on party system change through party family change presented in this thesis pitfalls may be avoided. Party families gather micro and macro features, because they keep the ideology, the party as a part of the system, the party as a whole and as an organization in mind. Party families combine ideology and party organization in much broader scope, and in doing so include ideological competition between party families due to distinct cleavage structures.
The essence of party families in the debate on party system change and the falsification of the freezing-claim can be explained through a thought experiment:
Take for instance a party system with more than two parties (two, or less is also possible), and imagine that everything is the same about these parties. Imagine that the organization of these parties is not different. All parties are either very open, very closed, or even a combination of both. Imagine then that these parties all focus around comparable cleavages. In this case imagine that, with each cleavage explained by Lipset and Rokkan, only one of these cleavages is represented within this fictional party system. Imagine that all of the electorate is volatile, but eventual elections don’t seem to shift on a large scale. Imagine that parties are very clear on if they prefer either consensus, conflict, or another form of competition. Imagine, if we looked at a longer period of time with multiple elections, that it becomes clear that parties seem to be very consistent. The number of parties do not vary much with each election.
What does this thought experiment tell us? Clearly, there is one important feature missing here. Even though this thought experiment is a mere thought, it still mirrors what real life looks like in the political arena. The diversity of these parties, and their categorization, become clearer when we give these parties an ideology. Parties in the contemporary political arena do not differ that much when we leave ideology out of the picture. Looking back at the scientific relevance, we can conclude that in this thought experiment there is not a feature that makes it relevant to research or even start looking at differentiation of parties or the party system. Only when we add the feature ideology to the bowl, does it become interesting to look at parties, party systems and party system change.
Chapter 3. Party Family Change
3.1 What is a Party Family?Although the notion of the party family underlies much of the standard work on comparative party politics, it nevertheless remains one of the most under theorized and least specified approaches to the general classification of parties.
In some ways, the metaphor of the party family is not a happy one, since the cross-party similarities it assumes are neither necessarily genetic nor formally structured (Mair and Mudde 1998: 23). Indeed, in some of its modes of operationalization, as will be seen later, the family connections that are asserted to exist may even be unrecognizable to a number of the family members concerned, who may be wholly unaware of their supposed kinship ties in other countries or in other times. Yet despite its drawbacks, the party family remains an attractive and easily grasped metaphor, and it has now become a standard reference point in much of the comparative party literature. It wins much more favor than the potentially more precise notion of functional equivalence, even though some of its versions imply this idea.
Even though the concept of party family is an under theorized concept, Mair and Mudde changed that when they analyzed the four principal approaches to identify party families. In his essay he argues that certain principles are better than others. The four principals elaborated by Mudde are (1) origins and sociology, (2) transnational links, (3) policy and ideology, and (4) party name.
The first of these criteria aims to group together parties that mobilized in similar historical circumstances or with the intention of representing similar interests. From Rokkan’s work, and later on the work of Von Beyme (1985: 64) 10 different ideological groups could be distinguished: liberals, conservatives, workers’ parties, agrarian parties, regional parties, Christian parties, communist parties, fascist parties, protest parties, and ecological movements. Seiler (1985) worked with the same cleavages as did Rokkan, but came up with less ‘political families’: bourgeois parties, workers’ parties, centralist parties, populist parties, Christian democrats, anticlerical parties, agrarian parties, and communist parties. The main notion all authors used was that parties could be divided into groups on one hand by their origins based on a pressing cleavage leading to the rise of political parties, and therefore based upon the leading conflict the party was based upon, and on the other hand according to the ties with the electorate and affiliated interest associations. The main problems specifically associated with this approach are that its geographic implication may be limited and that time may render it increasingly inadequate. Thus, while this approach might have suited analyses of the initial periods in which the now established parties began mobilizing support and contesting national elections, many of these actors have since changed in terms of their organizations, their ideologies, and even their electorates. Moreover, new parties have emerged in the latter part of the century that
cannot easily be integrated within the original cleavage model. Given that nearly a century of mass politics has elapsed in Western Europe since the initial freezing of cleavage structure, however, an approach that focuses exclusively on the origins of parties as the key to their contemporary classification risks neglecting more than it can offer (Mair and Mudde 1998: 33).
The second criterion that is increasingly adopted within the literature is derived from actions by the parties themselves. It involves the international links that political parties establish officially in the so-called transnational federations. These federations exist at the global level, such as the Liberal International; the continental level, such as the Asia-Pacific Socialist Organization; and the regional level, such as the Caribbean Democratic Union (Mudde 1998). This approach enjoys the advantage of being straightforward and easy to apply, since the lists of members of these transnational organizations are normally publicly accessible. Moreover, this standard follows the party’s own choices as to its cross-border links. Thus, the “growing inclination of political parties throughout the world to construct or join international organizations of like-minded formations” (Day 1988: ix) has made the applicability of this approach increasingly relevant with regard to both the variety of party groups and the number of countries involved.
Despite these advantages, however, the criterion is not as straightforward as is often assumed (Mair and Mudde 1998: 34). First and foremost, not all parties are yet members of transnational federations. Moreover, while agrarian, regional, and extreme right parties do sometimes build transnational federations at one level or another, their capacity to do so depends on the willingness of the individual parties involved, as well as on the size of their representation. A second major problem is that the classification of parties depends on which transnational federation is selected (Mair and Mudde 1998: 34). Some parties, for example, are members of different federations at the same level, as is the case with many parties. Finally, and perhaps inevitably, there is often little consistency over time in individual membership within the party groups, especially in the case of small and new parties.
The third criterion often adopted, based on the congruence of the policies and/or ideologies that are professed or even pursued by the parties, is probably the most difficult to delineate (Mair and Mudde 1998: 36).
The advantage of such an approach is that a variety of easily accessible and applicable datasets have recently been created and published, and these appear particularly well suited for quantitative cross-national analysis. This also constitutes one of the main problems, however (Mair and Mudde 1998: 36). The datasets are necessarily created through the use of similar scales for different political systems, thereby assuming, perhaps naively, that the same policy means the same thing in these different contexts.
Two major advantages were seen to accrue from this (Mair and Mudde 1998: 38). First, the classification and grouping of parties derived from what the parties themselves had chosen to
