CHAPTER 13 PARTY SYSTEMS 225 CHAPTER 13 Party systems Daniele Caramani Chapter contents Introduction 225 The genealogy of party systems 225 The morphology of party systems 230 The dynamics of party systems 237 Conclusion 242 Reader's guide This chapter looks at the competition between parties and how it leads to different party systems. First, the chapter looks at the origins of party systems. Historical cleavages between left and right, the liberal state and religious values or ethno-regional identities, agrarian and industrial sectors of the economy, led to socialist, liberal, religious, regional-ist, and other party families. Are they still the main actors today? Second, the chapter looks at the format of party systems, some of which include two large parties (two-party systems) while others are more fragmented (multiparty systems). What is the influence of the electoral system, and what are the consequences for governmental stability? Third, the chapter analyses the dynamics of party systems. To maximize votes parties tailor their programmes to voters' preferences and converge towards the centre of the] left-right axis. Is this why today new parties emerge at the extremes? Introduction v,onipr views parties in their connections within a This cnapitri r tern As in planet systems, the focus is not on single i ts but on the constellations they form: their num-the balance of size between them, and the distance that separates them. Parties can be ideologically near or distant there are systems with many small parties or few Ijrge ones or even—to pursue the analogy further—one large party with 'satellites' (as in some authoritarian sys-sj over time some systems change while others re-main stable. Thus, the variety of party 'constellations' is very large. Furthermore, the 'space' itself can change—either by expanding or shrinking (the extremes of left and right), or by acquiring more than one dimension {as for economic and cultural left-right positions). Whereas the dynamic principle of planets is gravity, the motor of political interactions is competition for power. In libera! democracies this competition is based on popular votes. The shape and dynamics of party-systems are determined by the electoral game in which parties are the main actors. Therefore, a party system is first and foremost the result of competitive interactions between parties. As in all 'games' there is a goal: the maximization of votes to control government. However, the set of interactions between parties is not exclusively composed of competition, but also of cooperation (for instance, when they build a coalition). Three main elements of party systems are important. | Which parties exist? Why do some parties exist in all party systems (e.g. socialists) whereas others only in some (e.g. regionalists or religious parties)? This relates to the origin, or genealogy, of party systems. 2. How many parties exist and how big are they? Why are some systems composed of two large parties and others of many small ones? This relates to the format, or morphology, of party systems. 3. How do parties behave? Why in some systems do parties converge towards the centre whereas in others they diverge to the extremes of the ideological 'space'? This relates to the dynamics of party systems. An obvious but important point is that party systems must be composed of several parties. There is no 'system' with one unit only. The competitive interaction between parties requires pluralism. If the goal is to get the most votes, there must be free elections and pluralism, without which competition cannot exist. Therefore this chapter fouises on democratic systems. KEY POINTS ^M*rty systems are sets of parties that compete and co-i °perate wth the aim of increasing their power in con-I trolling government. Interactions are determined by (l) which parties exist, (2) how many parties compose a system and how large they are, and (3) the way in which they maximize votes. It is appropriate to speak of a party system only in democratic contexts in which several parties compete for votes in open and plural elections. The genealogy of party systems The 'national' and 'industrial' revolutions Most contemporary parties and party families originated from the socio-economic and political changes between the mid-nineteenth century and the first two decades of the twentieth. Lipset and Rokkan (1967) distinguish two aspects of this transformation: (1) the Industrial Revolution refers to changes produced by industrialization and urbanization; (2) the National Revolution refers to the formation of nation-states (homogeneous and centralized political units) and liberal democracy (parliamentarism, individual civil and voting rights, rule of law, and secular institutions). The Industrial and National Revolutions created socio-economic and cultural divisions opposing different social groups, values, and interests. Lipset and Rokkan called these conflicts cleavages (see Box: What is a cleavage?, in the Online Resource Centréí, With the birth of modern parliaments and free elections, and with the extension of franchise, political parties developed and mirrored the socio-economic and cultural divisions created bv the two 'revolutions! Cleavages and their political translation Lipset and Rokkan distinguish four main cleavages created by the two 'revolutions' (see Table 13.1). These revolutions have each produced two main cleavages. Subsequent transformations have produced additional cleavages, namely the 'International Revolution! triggered by the Soviet Revolution of 1917, and the 'Post-Industrial Revolution' in the 1960s-70s, which led to a value cleavage between generations and globalization since the late 1990s. In the nineteenth century, socio-economic and cultural conflicts emerged simultaneously with democratic reforms. The fundamental features of today's party systems were set during the early phases of mobilization of, at first, restricted electorates (only very few people had the right to vote when liberals and conservatives dominated in the nineteenth century) and, later, of'mas-sifying' electorates when socialist parties mobilized the 226 DANIELE CARAMANI 1 Ö 9 mi Ii lists c pa c a c '3 Ep Mi St 2 c o o c 2 2 § j= er £ 'S Q P = ra <>. a 2 11 u s if 5 1 £ 3 g c .2 c S N ^ 1 II Ii 5 c o c -3 3i x a E fill .2 u 5 "ö § § IT o . If 2 C E 3 A Ii I 0 -s.»- i I f ' o — u s .S c t/i — S 3 £ 1 || 1 I < & ■ * £ S I o § 8 5od 11. ^ -I ft I §. I £ w 11 S i - M -i: =a v?. £ -§ ? 1 ft if L . c I E iss» ^ i o -u > ^ Im 0 X £ ,5 tt § £ • C 6 tL -ü. "Zj working class that emerged from the Industrial olution. lie National Revolution produced two cleavages. [ Centre-periphery cleavage This conflict emerged when nation-states formed in the nineteenth century, and political power, administrative structures, and taxation systems were centralized. This process also brought about national languages and sometimes religions. Most national territories were heterogeneous with different ethnicities and languages, and administration was fragmented. Nationalist and liberal elites carried out state formation and nation-building, facing resistance from subject populations in peripheral territories in two aspects. 1 Administrative: peripheries were incorporated in the bureaucratic and fiscal system of the new state (the central state controlled the territory of and extracted taxes), implying a loss of autonomy for regions. 2, Cultural: religious, ethnic, and linguistic identities in peripheral regions were replaced by the allegiance to the new nation-state fostered through compulsory schooling, military conscription, and other means of national socialization. As the first Italian prime minister said in 1870 after Italy unified, 'We have made Italy, let us make Italians'. Nation-building also took place in old-established states. In France in 1863, only 22 per cent of the communes spoke French, all located around Paris (Weber 1976:67). Resistance to administrative centralization and cultural standardization was and still is expressed in re-gionalist parti.es, such as the Scottish National Party, the various Basque and Catalan parties in Spain, the Bloc Quebequois in Canada, and so on. After democratization one finds many such ethnic parties in Africa. State-church cleavage Nation-states in the nineteenth century were not only centralized and homogeneous, but also based on liberal ideology and secular institutions (no church influence), individualism, and democracy (sometimes republicanism). Liberal reforms and the abolition of estates (clergy, aristocracy, bourgeoisie, peasantry) of pre-modern parliaments, as well as the individual vote and free elections, Put an end to clerical and aristocratic privilege. In this, liberals were opposed by conservatives in a conflict between the rising industrial bourgeoisie and the privilege of clergy and aristocracy. The new liberal secular state fought against the long-^tablistied role of the church in education. Compulsory education by the state was used to 'forge' new citizens. Pecially in Catholic countries this led to conflicts. The ch was also expropriated of real estate and, in Italy, CHAPTER 13 PARTY SYSTEMS it lost its temporal power and state (about a fourth of the Italian peninsula) when Italy unified as a nation-state in 1860-70. Conservatives wanted the return to the old pre-dem-ocratic regime. In some countries, Catholics took the place of conservatives, as in Belgium, Switzerland, and Germany. In other countries, Catholics were banned from being candidates and voting by papal decree (for this reason, Catholic parties did not appear in Italy and France until the 1910s). In fact, it was not until after the breakdown of democracy and the inter-war fascist period that the Catholic Church fully accepted democracy. 'Christian democracy' appears from this evolution after the Second World War. Today, some Islamic countries face a similar conflict, as in Turkey, Egypt, and Bangladesh. An interesting case is that of countries with mixed religious structures. In the Netherlands there was one unified Catholic party and a number of Reformed parties. In 1972 the religious parties merged into the Christian Democratic Appeal. An inter-confessional party also developed in Germany (the Christian Democratic Union). In Switzerland a major Catholic party emerged from the opposition to the Protestant Radicals/Liberals. The Industrial Revolution produced two additional cleavages. Rural-urban cleavage The first was the contrast between landed interests (agriculture) and the rising class of industrial and trading entrepreneurs. This cleavage focused on trade policies, with agrarians favouring trade barriers (protectionism) and urban industrialists favouring free market with low tariffs (liberalism). Weak sectors of the economy tend to be protectionist because of the threat of imports, whereas strong sectors favour the opening up of economic borders to increase exports (Rogowski 1989). Agriculture was threatened by technological progress and growth of productivity. The defence of agrarian interests—when peasant populations received the right to vote—was expressed through agrarian parties (also called peasants' or farmers' parties). Large or small agrarian parties existed everywhere in Europe, but were particularly strong in Eastern Europe and in Scandinavia. They also existed in Latin America. The period after the Second World War witnessed both the decline and the transformation of these parties. On the one hand, in most countries peasants' parties disappeared. On the other, the large agrarian parties of the north and east abandoned the agrarian platform and changed into centre parties. The recent reawakening of this cleavage is most notable in Latin America, where opposition to multinational companies, defence of raw materials and resources, and the threat of globalization has led to protectionist policies (e.g. gas and oil 227 228 : DANIELE CARAMANI nationalization in Bolivia and Venezuela). In the 1990s a number of peasant upheavals took place in the Chapas region in Mexico. This cleavage is also present in the European Union, where fanners' pressure groups lobby lor protectionist trade agreements and state subsidies. Workers-employers cleavage This is the cleavage between the industrial entrepreneurial bourgeoisie who started the Industrial Revolution and the working class that resulted from it. It is the opposition between 'capital' and 'labour' which, up to recently, characterized the left-right alignment. Left-right is the most common ideological dimension along which par-tics are placed, even in the US where a socialist party never developed (sec Box 13.1). Industrialization had a very deep impact on Western societies. It radically changed the production mode, caused unprecedented levels of geographical mobility through urbanization {from countryside to urban industrial centres), and transformed family structures. Living conditions in industrial centres were extremely poor. Therefore workers were easy to mobilize through trade unions, with socialism providing a unifying ideology. With the extension of voting rights, social democratic and labour parties gained parliamentary representation. Socialist parties campaigned for labour protection against the capitalist economy. They promoted social rights and welfare state provisions on top of civil and political rights, and a substantial equalization of living conditions in addition to formal legal equality (Marshall 1950; Kitschelt 1994). These claims concerned underage and female labour, wages, working hours, contract security, protection in the workplace and during unemployment or illness, progressive taxation, ac. cident insurance, and pension schemes. Socialists fa. 1 voured economic policies with a strong intervention I of the state in steering the economy and public invest- I ments (later Keynesianism) against the liberal free- I market ideology. They looked for state ownership 0f I infrastructure (transportation, energy), industries, and 1 sometimes finance. Many socialist and labour parlies originate from previously existing trade unions, the main organizations of the working class before universal suffrage. With restricted franchise most workers did not have the right to vote. | Unions responded to a number of needs of the working class, increased solidarity and cooperation within it, and I provided a wide range of services. The Soviet Revolution of 1917 produced a cleavage within the workers' movement. Communism-socialism cleavage In the aftermath of the Russian Revolution that led to the Soviet Union and the single-party regime controlled by the Communist Party, communist parties in all countries formed as splinters from the socialists. The main issue was the lead of the Soviet Communist Party in the international revolutionary movement and also ideological differences, namely whether a revolution would be necessary to take the proletariat to power, or if this goal could be achieved through elections. As a reaction against the radicalization of the working class and its powerful action through a new type of mass party organization, fascist parties emerged in a I number of European countries and, more or less directly, dominated government during the 1930s. Tlie.se parties ] favoured the nation over class and 'internationalism! I © BOX13.1 ZOOM-lM Why is there no socialism in the US? The main factors explaining the absence of a socialist ideology and workers' party in the most advanced capitalist country are: Open frontier Geographical and social mobility gave American workers the possibility to move on in search of better conditions. Party machines Dominance of Democrats and Republicans since the nineteenth century made the rise of third parties difficult. The free gift of the vote Working-class white men all had the right to vote, were integrated in the political system, and had a say in government's actions. Roast beef and apple pie The American working class was more affluent than the European and all socialist Utopias come to grief with a satisfied working class. No feudalism The absence of aristocracy in America made the working class very similar to the European bourgeoisie. Ask yourself: Do these conditions still apply in today's America? Can the popularity of Senator Sanders in the 2016 presidential primaries be seen as a proof of change, making socialism in the US possible? Lipset. S. M. {1977) 'Why No Socialism in the United States?', in S. Bialer and S. Sluzar (eds). Sources of Contemporary Radicalism (Boulder, CO: Westview Press), 13W9- _and Marks, C. (2000) ft Didn't Happen Here: Wfiy Socialism Failed in the United States (New Tork: Norton). 1 Sombart, W. (1976) Why is there No Socialism in the United States? (tondon: Macmillan}, translated from the German 1906 text. j private property over communism. Fascist parties Y/ete the product of the radicalization of the industrial bourgeoisie threatened by socialist policies, and of the ristocracy threatened by redistribution through land reforms- Finally, the 'Post-Industrial Revolution' (Bell 1973) created two more recent cleavages. fvlaterialism-post-materialism cleavage A cleavage between generations over sets of values emerged in the 1960s and 1970s as a consequence of the protracted period of international peace, economic wealth, and domestic security since the Second World War (lnglchart 1977). The younger cohort developed 'post-materialist values' focused on tolerance, equality, participation, expression, emancipation, respect for the environment, fair trade, peace, and Third World solidarity, as opposed to the 'materialist' values of the war generation centred around themes of national security, law and order, full employment, protection of private property, tradition, and authority (within the family and the state). These new values were primarily expressed in a number of new social movements (see Chapter 16): the civil rights movement in the US in the 1950s, pacifism from the Vietnam War in the 1960s, feminism in the 1970s, and environmcntalism in the 1980s. In the 1990s, new ant i-globalization movements developed against the globalization of the economy and the Americanization of culture (Delia Porta 1999). From a party politics perspective, however, there are only a few examples of a Uigntficant impact of these 'new left' movements, the main one being green parties (Miiller-Rommel and Pogmuke 2002). A more pervasive impact of the Post-Industrial Revolution is on the 'new right! The globalization cleavage Economic globalization has created a cleavage between sectors of the economy that profit from the blurring of national boundaries, and sectors that suffer from the competition from new markets and cheap labour from the East and Asia. 'Losers' of globalization and integration (Betz 1994) have reinforced support for populist protest parties who favour trade barriers to protect local manufacture and 'locals first' policies in the labour market. Ihese groups are the small and medium enterprises, unskilled workers, craftsmen, and agricultural producers. _ The economic defensive attitude of these groups is re-"tforced by cultural, anti-immigration, and xenophobic Prejudice, stressing religious and national values against "ethnic society and cosmopolitanism. Increasingly Swi Sent'ments finQl expression in referenda, as in Brjt ;criand °r in the case of the Brexit referendum in f ain *n 2016. The effect on parties systems is a vote CHAPTER 13 PARTY SYSTEMS for populist protest parties. Many of these parties rely upon an extreme-right heritage, such as the Austrian Liberal Party and the French National Front (Kitschelt 1995). New parties include the UK Independence Party, Alternative for Germany, Fidesz in Hungary, the Five-Star Movement in Italy, the Tea Party in the US, among many others. Others are sporadic parties, such as the One-Nation Party in Australia. In Southern Europe and Latin America populist tendencies have a left-wing (inclusive) character, as in Bolivia, Ecuador, and Venezuela (Mudde and Kaltwasser 2013, Burgess and Levitsky 2003). and in Spain with Podemos or Greece with Syriza. Populism is also a reaction to changing security conditions which, with terrorist attacks and the refugee crisis in Europe, have created a resurgenceollaw-and-ordervalues. Variations in cleavage constellations Cleavage constellations change across space and over time. Space Not all cleavages exist in all countries. There are a variety of constellations, and thus of party systems. Why do some cleavages exist in specific countries while not in others? To summarize the explicative part of the Lipset-Rokkan model, whereas the left-right cleavage exists everywhere and is a source of similarity, the state-church, rural-urban, and centre-periphery cleavages vary across countries and are a source of difference (Caramani 2015). Country-specific cleavage constellations are determined by the following. 1 Differences in objective factors such as diverse social structures: multiple ethnicities or religious groups, structure of the peasantry, class relations. The extent to which socio-economic and cultural divisions have been politicized by parties, i.e. by the action of elites (Rose 1976; Lijphart 1968&). ■ The relationship between cleavages: their existence and strength can prevent the development of new ones (e.g. agrarian claims have been incorporated by Catholic or conservative parties). There are homogeneous constellations where there is one predominant cleavage, namely the left-right cleavage (e.g. the US) and heterogeneous constellations in which various cleavages overlap or cut across one another in plural democracies, such as Belgium, Canada, India, the Netherlands, and Switzerland (Lijphart 1984). Time Until recently, party systems have remained extraordinarily stable since the 1920s. Even party labels have not changed (liberal, socialist, conservative), as a sort of 229 230 Í DANIELE CARAMANI political imprint. Lipset and Rokkan have formulated the so-called freezing hypothesis: [TJhe party systems of the 1960s reflect, with few but significant exceptions, the cleavage structures of the 1920s,.. [Tjhe party alternatives, and in remarkably many cases, the party organizations, are older than the majorities of the national electorates. {Lip-set and Rokkan 1967: 50; italics omitted) Today's party systems reflect the original conflicts from which they emerged (see Box: Party families, in the Online Resource Centre) despite a decline in cleavage politics with the blurring of social divisions (Franklin et al. 1992). In the 1920s the full mobilization of the electoral market through universal suffrage and PR caused its saturation. As in all markets, there are entry barriers in the electoral market. Little room was left for new parties. Thus, existing parties were able to maintain their control over electorates over generations. Empirical research debates the basic stability of electorates over time, with theses of dealignment and realignment of Western electorates (Dalton et al. 1985) based on survey data, or stabilization in a long-term perspective (Bartolini and Mair 1990) based on electoral volatility data (the change of votes from one election to the next). Over the last ten years, however, dramatic change has occurred in Western party systems with the rise of populist parties in many countries and the decline of established parties. KEY POINTS Party families originate from socio-economic and cultural cleavages created by industrialization, urbanization, and the formation of liberal states. I The centralized and liberal state creates conflicts with the church and with peripheral regions, leading to religious and regionalist parties. Industrialization opposes liberal economic interests to the rural world as well as to the working class, leading to agrarian and labour parties. Party constellations 'froze' and have remained stable until recently. Recent realignment along the generational value cleavage and globalization led to new party families: greens and, above all, populists. The morphology of party systems The competitive interaction between parties depends on the shape of party systems. The two main elements of their morphology are: (1) the number of competing parties, and (2) their size. How many players are there and how strong are they? The number and strength of act0 J can be observed at two levels: the votes parties get in e[ \ tions and the seats in parliament. Therefore a 'variable' 1 that must be considered is the electoral system, through which votes are translated into parliamentary seats Two types of party systems are not considered in this section because they do not fulfil the democratic condi tions that allow competition. 1. Single-party systems with only one legal party the authoritarian experiences of the Communist Party in the Soviet Union and today in China, the National So cialist Party in Germany in the 1930s, and the Baathist Party in Iraq until 1993 and in Syria until five years ago 2. Hegemonic party systems in which other parties are legal but as 'satellites', under the control of the hegemonic party: these are also totalitarian or authoritarian systems which existed in Egypt and Tunisia until the Arab Spring, and in many communist regimes in Central and Eastern Europe before 1989. There are four types of party systems. Dominant-party systems Dominant-party systems are characterized by one large party with a majority above the absolute majority of 50 per cent of seats for protracted periods of time (several decades). In these systems all parties are allowed to compete in free elections to challenge the dominant party. However, no other party receives enough votes to come close to 50 per cent. Therefore, there is no alternation in power and the dominant part}' does not need to enter coalitions to form a government. An example is India between 1947 and 1975. After Independence, the Congress Party received over 50 per cent of votes and was able to rule unchallenged until 1975-77 when the state of emergency' was declared. A more recent example of a dominant-party system is South Africa since the end of apartheid in the early 1990s. The African National Congress, initially led by Nelson Mandela, has been able to secure an absolute majority of votes because of the role it had in enfranchising the black population. In Europe, a case of a dominant-party system is Sweden. The Social Democratic Workers Party formed almost all governments from 1945 until 1998, with around 45 per cent of the votes on average. In Mexico, the Institutional Revolutionary Party was m power from the revolution of 1917 until 2000, Two-party systems A two-party system is one in which two equally balanced I large parties dominate the party system and alternate in I power. The two parties have comparable sizes and equ^ chances of winning elections. Even a small amount o j votes changing from one party to the other (electo I I . n cause a change of majority. Therefore alter-SW in power is frequent. These are very competitive nat'01\ Because both parties are large, the winning ^S is likely to receive an absolute majority of seats ^dform smgie-party governments without the need for ^The features of two-party systems are listed in bje 13 2. The two large parties have similar sizes j a^_45 ner cent of the votes each) which plu-(aroLina & i.> f . , rality electoral systems transform into absolute mr- joriti ies of seats for the largest party. A number of other mailer parties compete in the elections. However, they are margi inal as they are usually not necessary to form a government. In two-party systems single-party governments tend to alternate from one legislature to the next. This is, to a large extent, an effect of plurality electoral systems. Because the threshold in first-past-the-post (FPTP) systems is very high, the two main parties propose policies and programmes that are acceptable to a large part of the electorate. Plurality leads to ideological moderation and similarity of programmes. In turn, this similarity makes it easier for voters to switch from one party to the other, creating alternation. These systems are typical of the Anglo-Saxon world where plurality in single-member districts has been maintained, unlike continental Europe where around Table 13.2 Types of party system in democracies CHAPTER 13 PARTY SYSTEMS the First World War countries changed to PR systems. Today, only the US provides a 'perfect' example of a two-party system where Republicans and Democrats have dominated since I860.1 Australia maintains a strong two-party system with the Labour Party and the Liberals. Other examples include Costa Rica and Malta (where Labour and the Nationalist Party receive together close to 100 per cent of the votes). In Canada, Conservatives and Liberals dominated until 1993 (with a strong New Democratic Party), when the Bloc Quebequois and the Reform Party increased their support. Two-party systems can also be found in countries with PR electoral systems. Until recently, Austria and Israel were dominated by two parties. After the end of Franco's regime in 1977, Spain moved towards a two-party system which lasted until 2015 in spite of many (but small) regionalist parties. Germany was named a 'two-and-a-half system' with two large parties together (the Christian-Democratic Union and the Social Democratic Party) collecting more than 80 per cent of the votes and a smaller Liberal Party (around 5 per cent) with a pivotal position which enabled it to decide— through alliance—which of the larger parties would be in charge of government. With the rise of populist parties and the decline of established parties these systems are changing, in particular in Austria, Germany, Spain, and the UK. 231 Type of party system Features Cases Dominant-parr,' One large party with more than absolute majority of votes and seats No other party approaching 50% No alternation One-party government India until 1975, lapan between 1955 and 1993, Mexico until 2000, South Africa since 1994 Two-party Two large parties sharing together around 80% of votes and seats Balanced (35-45% each) with one of die two reaching 50% of seats Alternation between parties One-party government Austria, Israel, and UK until roughly 2000, Costa Rica, Malta, and New Zealand until 1998, Spain until 2015, South Africa until 1989, Turkey, US Multiparty Several or many parties, with none approaching 50% of votes and seats Parties of different sizes Parties run for elections individually and form coalitions after elections Alternation through coalition changes Coalition government Belgium, Canada, Colombia, Czech Republic, Denmark, Finland, Germany until 1989, Hungary, Italy before 1994, Netherlands, Poland, Russia, Switzerland Bipolar Two Large coalitions composed of several parties sharing together around 80% of votes and seats Coalitions are balanced (40-50% each) Coalitions are stable over time and run elections as electoral alliances Alternation between coalitions Coalition government France in the Fifth Republic, Germany since 1990, Italy since 1994, Portugal 232 DANIELE CARAMANI CHAPTER 13 PARTY SYSTEMS 233 Multiparty systems Multiparty systems are the most frequent and also the most complex type of party system, The number of parties ranges from three to double-digit figures. None of the parties in a multiparty system is majoritarian (with 50 per cent of the votes or seats). Furthermore, parties that compose a multiparty system are of different sizes: some are large (say, 30 per cent of the votes) and some are small (less than 5 per cent). Because no single party has an overall majority parties form coalitions to support a government. In parliamentary systems (see Chapters 5 and 7) the vote of confidence requires a 50 per cent majority of seats. Parties run individually in elections (contrary to bipolar systems) and governmental coalitions are negotiated after elections. PR does not hinder niche parties from addressing small segments of the electorate and does not lead to ideological moderation, which, in turn, makes it more difficult for voters to switch from one party to the ottaJ and cause a government change. As a consequen government change rarely takes place through electors change, but rather by swaps of coalition partners While multiparty systems are considered to nm resent better pluralism in countries with religi0 territorial, and ethno-linguistic cleavages, they ar' considered less stable, subject to frequent coaliti 'crises', with no single party clearly accountable (PoWeg 2000). Other positive aspects of multiparty systenj have been stressed since analyses in the 1960s and 1970s included small countries such as Belgium, the Netherlands, Switzerland, and the Scandinavian countries. In plural societies, PR and multiparty systems are a viable way to involve minorities in decision-making processes and reach consensus.2 As Chapter 5 shows, consociational or consensus democracies represent a different model of democracy from the majoritarianor 'Westminster' model. Both have advantages and disadvantages (see Box 13.2). BOX 13.2 FOR AND AGAINST A normative debate: advantages and disadvantages of party systems Two-party systems Multiparty systems Historically positive connotation Historically negative connotation Two-party systems are the main cases that resisted the breakdown of democracy between the First and Second World Wars: UK and US. After the First World War in Italy, Weimar Germany, the Spanish Second Republic, and the French Fourth Republic (1946-56) instability led to a crisis of democracy. Effective Ineffective Produces governments immediately after elections. Governments are stable because they are formed by a single party. Governments take a long time to form after elections because of negotiations between parties. Coalitions lead to unstable governments. Accountable Non-accountable Because there is only one party in government responsibility is clearly identifiable by the electorate. Because governments are formed by many parties, responsibility is obfuscated. Alternation No alternation Two main parties alternate in power. Voters directly influence the formation of government, and a small shift can cause government change. Coalition negotiations are out of the reach of voters' influence and shifts of votes are not necessarily followed by changes of government. Distortive Representative FPTP under-represcnts minorities and over-represents large mainstream parties of left-right. PR fairly represents minorities in societies with ethno-linguistic and religious parties. Moderation Radicalization All main parties have a chance to govern and thus avoid extreme claims. Need to gather votes from large moderate segments of the electorate. Multiparty systems allow representation of extreme parties. Some do not have any government prospect and do not hesitate to radicalize their claims. Discontinuity Continuity _ Decisions are made by majority and subsequent cabinets often reverse legislation. Decisions are made by consensus through consultation. More continuitv in legislation. ^Htg way in which multiparty systems function largely „ds on the degree to which parties are ideologically ^e A Sartori (1976) has distinguished two main fll of multiparty systems. Moderate rr ultiparty systems dynamic is similar to that of two-party systems. The umber of parlies is small and the direction of the com-0 ffion is centripetal, i.e. the main parties tend to con-e towards the centre of the left-right scale to attract the support of Hie moderate electorate. At the centre are one or more small parlies with whom the two large ones n either side may form a coalition. The ideological distance between parties is limited so that all coalitions are Polarized m Jtiparty systems These have three main features. First, there is a large ideological distance between parties with a strong dose of radicalism. Anti-system parties aim to change not only government but also the system of government (the regime). These parties do not share the principles of the political system and aim to change its institutions {Capoccia 2002). Thus, not all coalitions are viable, with some parties continuously excluded and in constant opposition. They become irresponsible and radicalize with promises they will never be called to put into practice. Second, there is one main party placed at the centre of the left-right axis which represents the 'system' against which extreme anti-system parties are opposed. Being always in power it also becomes irresponsible and unaccountable. This parry is not punished elcctorally because of the absence of viable alternatives. Third, the occupied centre discourages a centripetal move on the part of other parties. As a consequence, there is divergence and competition is centrifugal. Examples of polarized systems are the Weimar Republic in Germany from 1919 until 1933, and Italy between 1946 and 1992. Bipolar systems Bipolar party systems combine elements of multi- and two-party systems. As in multiparty systems there are many parties, none of which has a majorit}'. Again, coalition governments are the rule. These, however, form before elections and run as electoral alliances. They remain stable over time. There are usually two large coalitions ■ evenly balanced size alternating in power, Therefore, ^Petition resembles that of two-party systems. 'n ^rauce, let i and tight have alternated in power since Ihc left includes Socialists, Radicals, Communists, Greens, whereas the right includes GauIIists and ^ Wals (they merged in 2003 as the Union for a Popular 0Vemeilt)- In Italy since 1994 the centre-left coalition is composed of Social Democrats, Communists, Greens, and Catholics, whereas the centre-right coalition includes Silvio Berlusconi's party (which merged with the post-fascist party) and the Northern League. The coalitions have alternated in power in 1996, 2001, 2006, 2008, and partly in 2013. The number of parties As we have seen, the number of parties is important. But how, exactly, should parties be counted? If all parties that run in an election are counted (or even only those that get some votes) the number would be extremely large and useless for building a typology. In every election there are dozens of parties and candidates that get no votes or very few. Therefore it is necessary to have reasonable rules to decide how to count. There are two ways to count parties: (1) numerical, with indices based on the size of parties; (2) qualitative, with rules based on the role of parties in the system. Numerical rules These rules represent quantitative attempts to classify party systems on the basis of the number and size of parties that compose them. Various indices have been devised to summarize this basic information: are there many small parties {a fragmented party system) or a few large parties (a concentrated party system)? The most frequently used indices are Rae's fractional-ization index (Rae 1971) and the effective, number of parties (Laakso and Taagepera 1979). The fractionalization index (F) varies from zero (full concentration of seats or votes in one party) to one (total fragmentation with each seat going to a different party). The effective number of parties (E) indicates the number of parties in a system and does not have an upper limit. The two formulas are as follows: F = l-Zp] t-UZp] where p is the percentage of votes or seats for party iand E represents the sum for all parties. The percentages for all parties are squared to weight parties by their size. If there are two parties, A and IS, each receiving 50 per cent of the seats, first calculate the squares for party A (0.50 x O50 = 0.25) and party B (0.50 x 0.50 - 0.25) and then add them together (0.25 + 0.25 = 0.50), Thus; F = 1 - 0.50 = 0.50 E - 1/0.50 * 2 In this example, F is exactly mid-way between zero and one (0.50) and E counts that there are two parties. Table 13.3 lists the effective number of parties (based on seats) contesting recent elections in a number of countries.4 As can be seen, there is a wide variation between countries. The less fragmented countries are those using plurality/majoritarian or transferable vote 234 DANIELE CARAMANI Table 13.3 Rae's parliamentary fractionalization index (F), effective number of parliamentary parties (E), and Gallagher's index of disproportionality (LSq) r Country Election F E Australia 2013 0.51 2.0 1Z6 14.9 j 5.4 I Brazil 2.3 Canada 2.5 12.0 0.52 2.1 8.1 j ■ France 2012 0.80 2.8 17.8 ' 1 Germany 2013 0.72 3.5 7.8 J ■ Greece 2015 0.69 3.2 10.0 Hungary 2014 0.50 2.0 17.6 < 1 India 2014 0.54 2.2 19.9 Israel 2015 0.86 6.9 3.6 i 9 Italy 2013 0.81 3.5 17.7 9 "jM Japan 2014 0.59 2.4 22.4 Malta 2013 0.49 1.9 2.4 '1 ■ Mexico 2015 0.76 4.1 10.3 Netherlands 2012 0.82 5.7 1.0 New Zealand 2014 0.66 3.0 4.5 JH Portugal 2015 0.65 2.9 6.1 • Russia 2011 0.64 2.8 3.5 llH Spain 2015 0.76 4.1 6.8 j Sweden 2014 0.80 5.0 2.7 ] Switzerland 2015 0.80 4.9 4.4 iiH Turkey 2015 0.59 2.5 6.9 1 ■ UK 2015 0.60 2.5 15.0 US 2014 0.49 1.9 5.5 Notes: For calculations alliances have been considered (Brazil, Chile, Hungary, India). In mixed electoral systems, PR votes have been used (Hungary, Japan, Mexico). For Germany Zweitstimmen have been used and in France first-ballot votes. For the US, 2014. ] are mid-term elections. As a general rule for including parties in the calculation, all parties/alliances polling at least i%, or securing at least one seat, have been taken into account. Source: See sources in 'Country Profiles' (see also the Online Resource Centre). systems in single-member districts (Australia, France, UK, Hungary, Malta, the US), whereas the most fragmented countries are those with PR and many religious and ethno-linguistic parties (Belgium, Finland, the Netherlands, New Zealand, Norway, Switzerland). Qualitative rules In many cases it is not appropriate to consider numerical criteria only to decide whether or not a party is relevant. Often small parties—which quantitative rules would weight lightly—have far-reaching consequences for coalitions, influencing important decisions, mo- | bilizing people, and so on. Sometimes small parties are much more important than their sheer size would suggest. Sartori (1976) has developed two criteria to decide which parties really 'count' and should be 'counted' I. Coalition potential: a small party is irrelevant if over I a period of time it is not necessary for any type *J governmental coalition. On the contrary, a party must be counted if, disregarding its size, it is pivotal and determines whether or not a coalition is going to exist. Blackmail potential: a small party must be considered relevant when it is able to exercise pressure on gov-^Bwpimental decisions through threats or veto power. The influence of electoral laws given the impact of party system fragmentation on stability accountability, and representation, research in Mmparative politics has been concerned with the causes for varying numbers and size of parties. With the number of cleavages, the main cause is the electoral system. Electoral systems are mechanisms for the translation of preferences into votes, and votes into parliamentary seats. Chapter 10 shows that there are two main 'families' of electoral systems: (1) majoritarian systems in single-member constituencies, and (2) PR systems in multi-member constituencies. The first and best-known formulation of the causal relationship between electoral and party systems is Duverger's laws from his classic book Political Parties (1954). As can be seen in Box 13.3, the two laws are simple: plurality or majoritarian electoral systems favour two-party systems, whereas PR leads to multiparty systems. This relationship between electoral and party systems is due to two effects. Mechanical effects refer to the formula used to translate votes into seats. In single-member constituencies winning the seat is difficult. Only the part)' with the most votes gets the single seat. The second, third, fourth, etc., do not get any seat (first-past-the-post). This means that the threshold is high and all parties but the first one are eliminated. With PR, on the contrary, in each CHAPTER 13 PARTY SYSTEMS multi-member constituency many seats are allocated in proportion to the votes. Small parties are not excluded (a party with 5 per cent of votes gets roughly 5 per cent of seats) and therefore the overall number of parties making it into parliament is high. Psychological effects refer to the behaviour of voters and parties. 1. On the demand side (voters), in electoral systems in which only large parties have a chance to win seats, voters tend to vote strategically (not necessarily their first party preference) to avoid wasting votes on small parties with no chance of getting seats. Converging votes on large parties reduces their overall number. On the contrary, with PR in which small parties can win seats, voters vote sincerely (their first preference) because their vote is not wasted. This increases the vote for small parties and thus their overall number. 2. On the supply side (parties), with plurality small parties have an incentive to merge with others to i ncrease their chances of passing the threshold, thus reducing the number of parties. On the contrary, with PR parties have no incentive to merge: they can survive on their own and small splinter parties are not penalized. This increases the overall number of parties. Rae (1971), Riker (1982), and Sartori (1986) have questioned these laws by asking whether the reductive effect of majoritarian electoral systems works at the constituency level or at the national level. At the constituency level the high threshold reduces the number of parties. But does this always translate into a reduction at the national level? 235 )f BOX 13.3 DEFINITION The influence of electoral systems on party systems Duverger's 'laws' (1954) First Law 'The majority [plurality] single-ballot system tends to party dualism.' Second Law The second biillot [majority] system or proportional representation tend to multiparty ism.' Rae/Riker's 'proposition' {1971,1982} Plurality formulae are always associated with two-party compe-j tition except where strong local minority parties exist.' Sartori's 'tendency laws' (1986) Law 1 I Given systemic structuring and cross-constituency disper-(as joint necessary conditions), plurality systems cause : ta|,ea sufficient condition of) a two-party format.' 'PR formulas facilitate muitipartyisrr conducive to two-partyism.' und are, conversely, hardly Cox's 'coordination argument' (1997} 'Why ... would the same two parties necessarily compete in all districts [cross-constituency coordination or nationalization]?' Local candidates link together to compete more effectively. 'If a system (i) elects legislators by plurality rule in single-member districts; (2) elects its chief executive by something like nationwide plurality rule; and {3) holds executive and legislative elections concurrently, then it will tend to ... have a national two-party or one-party-dorninant system.' 236 DANIELE CARAMANI CHAPTER 13 PARTY SYSTEMS 237 Suppose that a parliament has 100 scats from 100 single-member constituencies. If in each constituency a different party wins the seat, we would end up with a fragmented parliament. Thus, the question is: under what conditions does the reductive effect of FPTP at the constituency level also reduce the number of parties at the national level? The answer is: majoritarian systems produce two-party systems at the national level only if parties are 'nationalized^ i.e. receive homogeneous sup-port in all constituencies (see Sartori's argument in Box 13.3). If there are parties with territorially concentrated support, this leads to fragmentation in the national party system. A party which is small nationwide can nonetheless be strong in specific regions and thus win seats and create fragmentation in the national parliament. In most countries, parly systems nationalized at the beginning of competitive elections in the mid-nineteenth century, so the support parties receive is increasingly homogeneous across regions and territorialized support has declined. This can be observed not only in Europe and North America, but also in India and Latin America (Caramani 2004; Chhibber and Koliman 2004; )ones and Mainwaring 2003) due to the development of national party organizations and increasing candidate coordination (Cox 1997). Therefore, where plurality systems exist, reduction of the number of parties has taken place. Plurality systems distort, party votes when they translate them into seats. They over-represent large parties (the share of seats for big parties is larger than their share of votes). They wtfder-re pre sent small parties. Howr can we measure the {dis)proportionality between votes and seats:" The most frequently used measure is the least squares index of disproportionality (LSq) (Gallagher 1991; Gallagher and Mitchell 2008: appendix B): LSq = -Jv 2L{v i) where v is the percentage of votes for party i, s is the percentage of seats for party i, and £ represents the sum for all parties. This index varies between zero (full propor- I tionality) and 100 (total disproportionality)- Take, as an I example, the results of the 2015 UK election in Table 13.4 1 If the total of the squared differences is halved (450.4/2 I = 225.2) and then the square root is taken, the result is 1 15.0, i.e. a high level of disproportionality between vote ] and seat distributions compared to other countries. ] The values of the LSq index are given in the last column of Table 13.3. In countries with plurality systems 1 (Canada, the UK, India) there is a stronger distortion of the popular vote. The same applies for other systems 1 based on single-member constituencies, such as France I with a two-ballot majoritarian system. On the contrary, disproportionality is lower for countries with PR systems. However, PR systems also have a reductive effect on the number of parties if the magnitude of constituencies is small, as in Spain. The magnitude refers to the number of seats allocated in a given constituency. Ihe larger the magnitude, the higher the proportionality between votes and seats. If the magnitude is small, the few seats go to the few largest parties.5 KEY POINTS The morphology of party systems is important for the competition between parties it concerns the number of players and their size. The main types are dominant- j party, two-party, multiparty, and bipolar systems. In two-party systems, moderate multiparty systems, and bipolar systems competition is centripetal and there is alternation in power In dominant-party sys- J terns and polarized multiparty systems there is no alfl ternation and competition is centrifugal. Measures of fragmentation ate b.^eo 01". the number and size of parties. However, small parties can also be important if they have coalition or blackmail potentia^H The format of party systems i=. inl'ut'nced by electoral systems. Through mechanical if rid psychological er- t| fects plurality tends towards two-party systems (large parties are over-represented) and PR tends towards multiparty systems. Table 13.4 Results of the 2015 UKelection and Gallagher's LSq index of disproportionality Party Votes (%) Seats (N) Scats (%) Difference (% seats-% votes) Squarec Conservatives Labour 36.9 30.5 331 232 50.9 35.7 14.0 5.2.......... 196.0^ 77S\ 43.6 Others Total 3.6 100.0 21 650 3.2 100.0 450_4_ -- The dynamics of party systems Table 13.5 The analogy between economic and electoral competition jji the wake of Joseph Schumpetcr's (1943) definition of democracy—a set of rules for selecting political lead-and making decisions by means of competition for votes—authors have developed analogies between electoral competition and market competition. In the electoral market, parties and candidates compete for 'shares' 0f the electorate, as happens in the economic world where firms compete for shares of the market. Parties are organizations whose main motive is the maximization of votes, and the exchange between represented and representatives is similar to that between demand and supply in the economy (see Table 13.5). The market analogy Anthony Downs's An Economic Theory of Democracy (1957) is a pioneering book in which the basic elements of these models were spelled out for the first time. In this model, actors (parties and voters) are rational. Parties calculate their strategies by formulating platforms with the goal of maximizing votes and being elected or re-elected. Parties are coalitions of individuals seeking to control institutions and act to gain office. To maximize votes parties offer programmes that appeal to many voters. Voters face alternatives which they order from most to least preferred and choose the alternative that ranks highest. Voters make a rational choice by voting for parties whose programmes are closest to their policy preferences, their interests, or their values and moral orientations. Voters vote on the basis of the proximity between parties' positions and their preferences. For that, they must know what the alternative proposals by different parties are, i.e. they are informed about their choices. Once elected, parties seek re-election through policies appealing large segments of the electorate. The goal of parties is to maximize utility in terms of votes; the voter, .oal is to maximize utility through Policies satisfying their interests and values. As in economic theory, the search for individual advantages Produces comma: goods, namely responsiveness and accountability. Rational choice competition models were first doused for two-pur: y systems—mainly the US. However, temsIm'ZaUOn uf lcs is also the main motive in sys- ^ >n which governments are coalitions. The more S' *e better the chances to enter a coalition, control Dimensions Economy Elections Market Actors Firms Parties Consumers Voters Profit Money Votes Supply Goods, services Programmes, policies Demand Product preferences Policy preferences Communication Advertising Campaigns governmental institutions, and place individuals in key official positions. The spatial analogy The idea of proximity/distance between individual preferences and parties' policies indicates that players move within a space. The second element that Downs 'imported' from economic models of competition is their spatial representation. In particular, Downs adapted models of the dynamics of competition between firms, i.e. where firms locate premises according to the physical distribution of the population. Let us take the simple case of a village in which there is only one street (the example is from Hbfelling (1929)). On each side of the street there are evenly spaced houses (the square dots in Figure 13.1). What are the dynamics between two competitors, say two bakeries A and B? Assuming that both bakers offer the same quality of bread for the same price and that consumers will rationally try to reduce their 'costs' by buying bread in the nearest shop (proximity), if A and B are located as they arc in Figure 13.1, B will have a larger share of the market. Ihe share of B's market goes from the right-side end of the street to the M-point, which is the mid-point between the locations of A and B. Residents on the right of the M-point will buy bread in bakery B and residents on the left of the M-point will buy bread in bakery A. Ihe dynamic element in this model consists of As move to increase its share of the market. By relocating the bakery at AA, the baker is able to gain the share of the market indicated by the dashed area. Obviously, B can also move towards the centre (BB) and win back part of the lost share of the market. Both bakers seek to optimize their location. gure 13.1 Hotelling's model (1929) 238 j DANIELE CARAMANI An additional element introduced by Smithies (1941) concerns the elasticity of demand. The further away they are from the bakery, the higher the 'costs' for buyers. Incentives for a new bakery at the edges of the village increase as people feel that AA and BB are too distant. The risk of strategies of relocation towards the centre is that a new bakery C appears, taking away part of B's share of the market (the dark shaded area). Therefore there are two dynamic elements in these models: (1) the movement caused by the search for the optimal location and (2) the appearance of new competitors. Downs's model Through the analogy between physical and ideological space, Downs imports these elements into the analysis of the dynamics of parly systems. Most elements are maintained: (1) the one-dimensionality of the space, (2) the principle according to which costs are reduced by choosing the closest option (proximity), and (3) competitors' search for the optimal location through a convergence towards the centre. Downs represented the ideological space by a zero to 100 scale, ranging from left to right. As will be seen, one-dimensionality is maintained, even if it is not ahvayi a realistic assumption, because it summarizes other dj] mensions and is the most important one (in terms of si of parties that define themselves according to this mension), and because it is present in all party systems. Both Hotelling and Smithies had previously appty spatial models to politics and were able to predict that' parties tend to converge towards one another in the effort to win the middle-of-the-road voters, and to present increasingly similar programmes. Downs adds one crucial element to the models: the variable distribution of voters along the left-right continuum. Voters are not distributed regularly along the scale but concentrate in particular ideological positions, namely around the centre. For Downs, this is the crucial explanatory and predictive element of the dynamics of party systems: 'if we know something about the distribution of voters' preferences, we can make specific predictions about how ideologies change in content as parties maneuver to gain power' (Downs 1957: 114). If one assumes a normal (or 'bell-shaped') distribution of the electorate with many voters at the centre of the scale and fewer at the extremes (see type A in Figure 13.2), the prediction of the model is that parties will converge towards the centre.6 Type A: Downs's basic model (1957): the bell-shape (or normal) distribution of the electorate: centripetal competition CHAPTER 13 PARTY SYSTEMS i 239 Type C: A skewed distribution of electors: enfranchisement in the nineteenth century and new parties Number of electors /S—---- \ 1 Type D: Polymodal distribution in multi-party systems Figure 13.2 Types of voter distribution (Continued) Figure 13.2 Types of voter distribution ;t, these models predict the convergence towards : centre and the increasing similarity of platforms 1 policy actions. The centripetal competition is determined by the parties' aim to win the median voter (see Box 13.4) Examples are the progressive convergence of previously radical left-wing workers' parties towards the centre to attract moderate voters (the German Social Democrats in 1959, the New Labour Party under Tony Blair, or the US Democrats under Bill Clinton). Second, centripetal competition arises also because there are more voters in the centre. Party strategy does not depend only on the logic of the model (the assumption of proximity voting), but also on the empirical dis-X™im °fthe electorate. The potential.loss of voters extremes c oes not deter parties from converging tJ^ause there ar-- few voters at the extremes. This is not Case if the d'stribution of the electorate is differ-in tv°r examPk a two-modal distribution as depicted JrPe B in n i3'2- This's a case of ideological ■fcation within a political system. Therefore, the gfc Ution of Lhe electorate determines the direction ^Ptpetition (centrifugal or centripetal). BOX 13.4 DEFINITION The median voter The median voter is the voter who divides a distribution of voters placed on a left-right scale into two equal halves. In a distribution from zero to 100 in which for each point there is a voter (including position zero}, the median voter is on position 50 (with fifty voters on each side). Suppose, however, that there are fifty voters on position 100, and the remaining voters are distributed regularly between positions 49 and 99 (one voter on each position). In this case the median voter is on position 99. Third, voters in the middle of the left-right axis are more flexible than at the extremes, where they are firmly encapsulated in strict ideologies and/or party organizations. Available' voters (Bartolini and Mair 1990), located in the middle, are less ideologized and have weak party identifications. These voters are ready to change their minds and therefore are very appealing to parties seeking to 'seduce' them. 240 DANIELE CARAMANI The broader application of rational choice models What are the links of these models with other aspects of parties and party systems? First, rational choice models help to interpret the transformation of party organizations from mass parties to catch-all parties (see Chapter 12). This transformation can be seen as organizational and ideological adaptation to competition. Second, they also help to interpret patterns oidealign-ment, i.e. the loosening of the relationship between parties and specific segments of society (workers for social democrats, for example). Centripetal competition and the maximization of votes lead parties to make their programmes and ideologies more vague to attract support from other groups. This blurs the connection between groups and parties, and causes a higher propensity to change vote from one election to the next. Third, these models can be applied historically to processes of enfranchisement and democratization. In both type A and type B the distributions are symmetrical. In type C, on the contrary, we have a skewed distribution. The solid curve represents an electorate that is skewed towards the right of the axis. Here the median voter is around position 65 rather than 50, and accordingly parties A and B (liberals and conservatives of 'internal origin') would converge towards this point. This is typical of restricted electorates in the nineteenth century when lower classes were excluded from the franchise. Enfranchisement and democratization extended the space and changed the shape of voters' distribution as represented by the dashed curve. This explains the emergence of new parties C and D of external origin' (social democrats and agrarians).' The dream of reformists {as opposed to revolutionary socialists) was that socialism and the proletariat could come to power through votes ('paper stones') and the ex-1 tension of the franchise, rather than through revoluUOn | (real stones!). The development of the industrial society I would lead workers to power through sheer numbers ' Yet the numbers of industrial workers did not grow—jn I fact, they declined—and socialist parties faced a dilemma between moving towards the centre to maximize their appeal to the middle classes—thus relaxing their pro- 1 gramme—and losing voters from workers (Przeworski and Sprague 1986). Fourth, one needs to consider these models under PR and multiparty systems. Convergence is likely under FPTP because the threat of other parties appearing at the extremes is low, given the high threshold required 1 to win a seat. Rather than new parties, under these systems the model predicts high abstention levels, as is the case in the US. Multiparty systems, on the other hand, develop when the distribution of the electorate is polymodal, with more than one or two peaks (type D). The dynamics of the competition is not centripetal. Existing parties have no incentive to converge towards the centre as PR is no hindrance to new parties on the extremes. On the contrary, ideological spaces are elastic, with extremes widening and increasing ideological distance between parties. Parties may adopt such strategies to distinguish themselves from moderate parties. This leads to radicalization to maintain a distinctive character. Despite critiques (see Box 13.5), spatial models are extremely useful. In all countries a number of voters j are ready to change their vote. This is an available electorate around which competition turns and on which these models focus. Rational choice models apply less to I segments that are encapsulated through identification (cthno-linguistic and religious dimensions with strong identities and non-available voters) and primarily to the left-right dimension along which voters are available j BOX 13.5 FOR AND AGAINST Critiques of rational choice models Assumption Critique Rationality As social psychology, experimental politics, and behavioural economics show, voters motivations are not always rational. .............. Full information Voters are not fully informed about programmes and are unable to evaluate the extent to which they correspond to their own interests. With technical issues this often proves unlikely. .................... Vote maximization Parties are not only 'vote-seeking' (Müller and Strom 2000). As 'office-seeking', parties do not require to maximize votes but to get just enough (Riker 1962). As 'policy-seeking; parties seek to influence public policy rather than aim for office (De Swaan 1973). One-dimensionality Not all parties compete along the left-right dimension. In most cases the space of competition is multidimensional._ Even if applicable to parts of the electorate only, they are crucial as they determine the direction of competition.8 A more fundamental question is how to interpret the convergence of parties and the increasing similarity of their programmes. It is difficult to separate the impact of competition from other factors, such as the development of a large and homogeneous middle class and the disappearance of class distinctions through the reduction of social inequalities and the secularization of society, the integration of societies, and the disappearance of ethno-Hnguistic particularities through nationalization and globalization. E5tablishing the extent of convergence is what empirical spatial analysis attempts to do. Empirical spatial analysis Knowing what the voters' distribution looks like and to know where parties are located is a matter of empirical research. How can one measure the ideological space? First, the distribution of voters can be measured empirically through surveys in which, through questions and scales, respondents are asked to position themselves, for example, along the left-right axis. Examples are the World Value Survey, Eurobarometers, European Social Survey, Latinobarometers, etc. Second, the position of parties can be measured empirically through two main instruments. The most important one is based on text data, whereby party manifestos are analysed using special software able to identify the salience and favourable (or unfavourable) mentions of a large number of issues, from taxation and the free market to development aid and the environment. Such items are often combined to build a more general scale such as the left-right scale. The most important data are those of the Comparative Manifesto Project (Budge etal 2001; Bingemann et al. 2006; Volkens et al 2013). Others CHAPTER 13 PARTY SYSTEMS analyse the position of parties based on press releases, newspaper articles, social media, websites, or parliamentary debates (see, for an important study based on the coding of newspapers, Kriesi et al. 2012). The alternative possibility is expert surveys, whereby the position of parties on various issues is established by asking a sample of scholars. The problem with these data is that they rarely allow an estimation of parries' past positioning. Two main projects exist: the Benoit, Hunt, and Laver project (Benoit and Laver 2006) and the Chapel Hill project (Steenbergen and Marks 2007). A number of empirical spatial analyses were able to establish that the space of competition is in fact composed of two dimensions. The first is dre traditional left-right economic/instrumental dimension on the degree of redistribution between rich and poor, planning of the economy vs free market, which Kitschelt (1994) summarized as the socialism-capitalism axis (Figure 13.3). The second dimension has acquired salience more recently in the wake of the post-materialist and globalization transformations. Sometimes this axis is considered a cultural axis (Kriesi et al. 2012) insofar as it includes, at one extreme, post-materialist values (Green-Libertarian-Alternative) and, at the other extreme, materialist values (Traditional-Authoritarian-Nationalist). This has been labelled the GAL-TAN dimension (Hooghe et al. 2002). This dimension, however, has also a class base insofar as the former is supported by the highly qualified 'winners' of globalization and economic openness (as well as by those in 'sheltered' state employment) while the latter is supported by the 'losers' of.the globalized society who favour protectionism and controlled borders (the less-qualified blue-collar workers working in economic sectors 'exposed' to the threat of open borders) (Betz 1994). Therefore, it includes more vs less favourable attitudes towards supra-national integration 241 Capitalism Figure 13-3 The economic and cultural dimensions of left-right 242 DANIELE CARAMANI (integration-demarcation cleavage). The integration-demarcation comprises also the willingness to integrate immigrants (another consequence of globalization and the openness of borders) or, conversely, to demarcate more strongly cultural and national identities. This dimension affects party systems but also direct votes, as in the case of the Brent referendum in Britain in 2016, as well as presidential elections, as in the US with Donald Trump's victory in 2016. With the decline of industry and the working class, the strategy of established parties had been to mobilize voters on the cultural axis rather than on the economic one. This has tilted the left-right axis diagonally, creating the opportunity for new competitors (populists in particular) to mobilize precisely the 'losers' who would have traditionally voted for the left. This tendency has increased with the economic and financial crisis since 2008. In terms of spatial analysis, the emergence of populist challengers can be seen as a direct consequence of KEY POINTS In the electoral market, parties (the supply side) pres ent platforms to appeal to many voters whose vote is determined by the proximity of their preferences (the demand side) to the parties' offer. Voters are assumed to be rational, informed about alternative proposals, and able to choose the alternative closest to their top preferences. The dynamics of party systems is determined by parties' search for the optimal location on the left-right axis. Depending on the distribution of the electorate along the scale, parties move to a position where the support is largest. The prediction of competition models is that parties converge towards the centre of the left-right axis, as the point where most votes concentrate, and as the point where voters are less rigidly ideologized. the convergence (often also as great coalitions) betwgj established parties on issues like immigration and supra-1 national integration which left uncovered vast sectors of 1 the electorate. Long-term data show clearly how social 1 democrats and centre-right parties have given radical. 1 right populists the chance to emerge on issues such »1 economic and cultural threat (Caramani 2015). Conclusion Understanding party systems requires the combination 1 of the various perspectives presented in this chapter. 1 The macro-sociological approach must be combined with institutional and actor-oriented models. They 1 complement each other and are not mutually exclusive. I We cannot understand party systems without reference to the social cleavages from which parties emerge. However, we must also take into account parties' capacity to act independently from social conditions—in fact, to shape them through ideology and policy. The | motivations of parties are not entirely determined by their origins. Parties' strategies, in turn, must take intol account the rules of the game—electoral laws being the most important ones—influencing the number and size of players. Both descriptive, and explanatory research is needed The ultimate goal is to account for the shape and dy-1 namics of party systems. However, before searching fori causes, part)' systems should be described carefully. As seen with counting parties, this is often more compli- j cated than it appears at first sight. Finally, we cannot understand party systems in iso-1 lation. We need comparison to assess whether or not! they are fragmented or unstable, as well as a /ti^-fen»| perspective rather than a myopic focus on just the! most recent elections. This is Lhe only way of assessing! how exceptional a given party system or a given change! really is. Questions Knowledge based 1. What are the National and Industrial Revolutions? 2. What are Stein Rokkan's four main social cleavages and which party families emerged from them?1 3. How should the number of parties in a system be counted? 4. What are the characteristic features of a two-party system? 5. What does 'effective number of parties' mean? Critical thinking 1. What is the effect of electoral systems on the shape of p«fl systems? 2. What does it mean when we say that parties are mizers'? 3. Describe centripetal and centrifugal party competitio | Downs's model. 4. Are voters really rational? 5. Can the space of competition be reduced to one 1« -*•■ dimension? er reading Classical texts on party systems f C w 0997' ^a^'nS V°tei Count: Strategic Coordination ;„the World's sectoral Systems (Cambridge: Cambridge I University Press) i rjowns A (1957) 'w ^°nomic Theory of Democracy (New York: [ narperColli'-i:-; Ouvergei, M- ('934) Po^'^ Parties (New York: John Wiley). Kitschelt. H. (iQv- 1 Transformation of European Social Democracy (C.i.iondge: Cambridge University Press). Upset S- M., an-.' f.okkan, S. {1967) 'Cleavage Structures, Party Systems, -.-.nd Voter Alignments: An Introduction', in S M. Lipset and s Rokkan (eds). Party Systems and Voter Alignments (Nc,v York: Free Press), 1-64. CHAPTER 13 PARTY SYSTEMS Sartori, C. (1976) Parties and Party Systems: A Frameworkfor Analysis (Cambridge: Cambridge University Press). For a more recent bibliography, see the works cited throughout in this chapter. Up-to-date reports on party systems can be found in journals. Detailed country-by-country developments from 1991 onwards are reported in the Political Data Yearbook. Students may also find useful material in journals such as the American Political Science Review, Comparative Politics, Comparative Political Studies, Electoral Studies, Party Politics, and West European Politics. 243 •? Web links http://www.electiondataarchive.ong TheConstituency-i.ovel Elections Archive (CLEA) is the largest databank on elections at the level of single constituencies since the beginning of -. nnetitive elections covering the entire world. http://www.manifesto-projed.wzb.eu Hfcbsite of the rvUnnesto Project Database hosted by the WZ8 I in Berlin with dat? un ideological and programmatic position of political parties http://www.psephos.adam-carr.net Adam Carr's websi:e with election results for 182 countries for legislative and presidential elections. http://www.nsd.uib.no/european-election_database Searchable datab...: of the Norwegian Social Science Data Services (NSD) w,!, results for national legislatures in Europe "ndfortheeleaiu : . the European Parliament. Includes information on parties and party systems. For additional material and resources, please visit the Online Resource Centre at: nttp://www.oxfordtextbooks.co.uk/orc/caramani4e/ http://www.electionguide.org Information on elections and electoral systems from the Consortium for Elections and Political Process Strengthening (CEPPS). Includes information on parties and party systems worldwide. http://www.eri.wikipedia.org/wiki/List_oi_.politicaLparties_by_ country Information about names, ideologies, mergers, and splits of political parties worldwide, and results for legislative and presidential elections with different levels of quality depending on countries. http://www.idea.int Website of the International Institute for Democracy and Electoral Assistance (IDEA). Endnotes f OK US, two-party systems exist largely because of the •rming _ party which make it difficult for third 0 present candidates. For this reason there is not "^Proportionality between votes and seats in Table - The literature on the positive sides of multiparty systems insists on a different decision-making mode based on consensus. This literature includes Ruslow (1955), Daalder (1966), Lorwin (1966„,_). Lijphart(19G8„,RMcRae (1974), and Steiner(1974), all stressing accommodation, agreement, and compromise. 244 : DANIELE CARAMANI J 1958 marks the beginning of the Fifth Republic in France with the new 'Gaullisť constitution and a two-ballot majority' system in single-member constituencies. 4 tithe effective number of parties is calculated on votes, this is usually referred to as 'effective number of elective parties' (ENEP), whereas if it is calculated on seats it is called effective number of parliamentary parties' (ENPP). :' Additional causes of distortion between votes and seats are the rules to form a party and present candidates, the size oi parliamentary groups, the barrier clauses, the type of quota (Hare, Droop. Imperiáli, etc.). and the number of tiers (i.e. various levels of constituencies for the allocation of seats}. 6 A further assumption is that if a voter prefers LdeologjJ position 60 over 70, then the voter will also prefer 60 over 80,%) etc. (transitivity). This is an assumption of single-peakedness J voter preferences, i.e. if a voter prefers 60 then the further aw« the position is from 60 the less it is liked. 7 Such a modification of the left-right distribution through new voters is unique. Later'waves' of enfranchisement—narneW when women and younger generations were enfranchised wift the lowering of the voting age—did not have a similar effect. 8 As a response to the criticism about full information, ration nal choice theorists argue that it is not rational to spend a loi of time gathering political information (the costs outweigh be» efits). This is a free-rider attitude. Chapter contents Introduction What are intr.resl groups? Interest associations in theory Interest associations in practice Conclusion 246 246 247 252 257 CHAPTER 14 Interest groups Roland Erne Reader's guide Political scientists should not just compare political institutions, but also assess the role of associations that seek to advance particular socio-economic and political interests. Interest groups play a crucial role in all political systems. But the forms in which interests are articulated depend on the particular context. Accordingly, this chapter begins with a review of different definitions of interest groups that have been used across time and space. Scholars of interest politics have also been inspired by different theoretical paradigms. Hence, the chapter discusses the legacies of competing theoretical traditions in the field, namely republicanism, pluralism, and neocorporatism. The final sections of the chapter assess the role of interest associations in practice, distinguishing different types of action that are available to different interest associations, namely direct lobbying, political exchange, contentious politics, and private interest government.