OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13

Party systems
Daniele Caramani

Chapter contents Reader’s guide

Introduction 238 This chapter looks at the interactions between parties and
The genealogy of party systems 238 how party competition leads to different party systems.
The morphology of party systems 244 First, the chapter looks at the origins of party systems. His-
The dynamics of party systems 252 torical cleavages between left and right, the liberal state and
Conclusion 257 religious values or ethno-regional identities, agrarian and
industrial sectors of the economy, led to socialist, liberal, re-
ligious, regionalist, and other party families. Why are these
old party families still the main actors in our time? 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 chap-
ter 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 parties propose increasingly similar policies and
programmes?
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

238 DANIELE CARAMANI

Introduction KEY POINTS

This chapter views parties in their connections and re- ● Party systems are sets of parties that compete and
lationships within a system. As in planet systems the cooperate with the aim of increasing their power in
focus is not on single planets but on the constellations controlling government.
they form: their number, the balance of size between
● What determine interactions are (1) which parties
them, and the distance that separates them. Parties can
exist, (2) how many parties compose a system and
be ideologically near or distant, there are systems with
how large they are, and (3) the way in which they
many small parties or few large ones or even—to pursue maximize votes.
the analogy further—one large party with ‘satellites’ (as
● It is appropriate to speak of a party system only in
in some authoritarian systems). Over time some systems
democratic contexts in which several parties compete
change while others remain stable. The variety of party
for votes in open and plural elections.
‘constellations’ is, thus, very large.
Whereas the dynamic principle of planets is grav-
ity, the motor of political interactions is competition
for power. In liberal democracies this competition is
The genealogy of party systems
based on popular votes. The shape and dynamics of
party systems are determined by the electoral game
The ‘national’ and ‘industrial’
in which parties are the main actors. A party system is revolutions
therefore first and foremost the result of competitive in- Most contemporary parties and party families originated
teractions between parties. As in all `games´ there is a from the radical socio-economic and political changes
goal: the maximization of votes to control government. between the mid-nineteenth century and the first two
In this sense, party systems are much more dynamic decades of the twentieth. Lipset and Rokkan (1967)
(changeable) than star systems. The set of interactions distinguish two aspects of this transformation: (1) the
between parties, however, is not exclusively composed Industrial Revolution refers to changes produced by in-
of competition, but also of cooperation. Parties, for ex- dustrialization and urbanization; (2) the National Revo-
ample, cooperate when they build a coalition to support lution refers to the formation of nation-states (culturally
a government. homogeneous and centralized political units), and liberal
Three main elements of party systems are important: democracy (parliamentarism, individual civil and voting
1. Which parties exist? Why do some parties exist in all rights, rule of law, and secular institutions).
party systems (e.g. socialists) whereas others only in The Industrial and National Revolutions created
some (e.g. regionalists, agrarians, or confessional par- socio-economic and cultural divisions opposing different
ties)? This relates to the origin, or genealogy, of party social groups, elites, sets of values, and interests. Lipset
systems. and Rokkan name these conflicts cleavages (see Box:
2. How many parties exist and how big are they? Why What is a cleavage?, in the Online Resource Centre).
are some systems composed of two large parties and With the birth of modern parliaments and free elections,
others of many small ones? This relates to the format, and with the progressive extension of franchise, political
or morphology, of party systems. parties developed and reflected the socio-economic and
cultural divisions created by the two ‘revolutions’. Mod-
3. How do parties behave? Why in some systems do
ern party families appeared as the ‘political translation’
parties converge towards the centre whereas in
of social divisions in systems in which conflict is increas-
others they diverge to the extremes of the ideologi-
ingly settled through vote.
cal ‘space’? This relates to the dynamics of party
systems.
An obvious but important point is that party systems Cleavages and their political
must be composed of several parties. There is no ‘sys- translation
tem’ with one unit only. The competitive interaction
between parties requires pluralism. If the goal is to get Lipset and Rokkan distinguish four main cleavages cre-
the most votes, there must be free elections, some de- ated by the two ‘revolutions’ (see Table 13.1). These
gree of enfranchisement, and pluralism without which revolutions have each produced two main cleavages.
competition cannot exist. This chapter therefore focuses Subsequent transformations have produced addi-
on democratic systems and excludes totalitarian or au- tional cleavages, namely the ‘International Revolution’,
thoritarian regimes with single parties (such as China or triggered by the Soviet Revolution of 1917, and the ‘Post-
Syria). Industrial Revolution’ in the 1960s–70s, which led to
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 239

Table 13.1 Stein Rokkan’s cleavages and their partisan expression

Revolution Timing Cleavage Divisive issue(s) Party families Examples

National Early 19th Centre– Liberals and conserva- Regionalists, Scottish
century periphery tives face resistance to ethnic parties, National
(restricted state/administrative linguistic par- Party, Bloc
electorates) centralization and cul- ties, minorities. Québéquois,
tural standardization Partido
(language/religion). Nacionalista
Vasco.
State–church Conflict between Conservative Austrian
liberal and secularized and religious People’s Party,
state against clerical parties (Catho- Christian-
and aristocratic privi- lic mainly), Democratic
lege, and over religious Christian Union, Swiss
education, influence democracy. Catholic
of church in politics, Party, Partido
democratic institu- Popular.
tions.
Industrial Late 19th Rural–urban Conflict between Agrarian and Finnish Centre
century industrial and agri- peasant parties. Party, Austral-
(suffrage cultural sectors of the ian Country
extension) economy on trade Party, Polish
policies: agrarian Peasant Peo-
protectionism vs. in- ple’s Party.
dustrial liberalism (free
trade vs. tariffs).
Workers– Employers vs. the ris- Workers’ par- British Labour
employers ing working class on ties, socialists Party, Argen-
job security, pensions, and social tinian Socialist
social protection, democrats, Party, Swedish
degree of state inter- labour parties. Social-Demo-
vention in economy. cratic Workers’
Party, Spanish
PSOE.
International Early 20th Communists– Division within the Communists. Partito Comu-
century socialists ‘left’ (workers’ move- nista Italiano,
(mass elec- ment) over centrality Izquierda
torates) of the Soviet Union Unida, Parti
Communist Party Communiste
and its international Français,
leadership, and over Japan’s Com-
reformism vs. revolu- munist Party.
tion.
Post-industrial Late 20th Materialist– Generational cleavage Green and Die Grünen,
century (de- post- over policy priorities: ecologist Austrian
mobilized materialist new values of civic parties. Grünen/Grüne
electorates) values rights, pacifism, femi- Alternative,
nism, environment. Democrats
’66, Women’s
Party.
Open–closed Globalization of the Protest parties, FPÖ, Front Na-
societies economy, opening nationalist tional, Danish
up of labour markets, parties, Progress Party,
competition from extreme right- Fifth Republic
cheap Asian labour, wing parties, Movement
fiscal and monetary neo-populist (Hugo Chávez),
integration in Europe, parties. Movement for
and anti-Americaniza- Socialism (Evo
tion of culture. Morales).
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

240 DANIELE CARAMANI

a value cleavage between generations and globalization such as the Scottish National Party, the Swedish Party in
since the late 1990s. Finland, the various Basque and Catalan parties in Spain,
In the transformation of the nineteenth century socio- the parties of the German- and French-speaking minori-
economic and cultural conflicts emerged simultaneously ties in Italy, the Bloc Québéquois in Canada, and so on,
with democratic reforms: the creation of modern par- opposing nationalist/liberal parties.
liaments, free competitive elections, and the extension
of civil and political rights. Conflicts of that time were
expressed in organizations that were typical of this new
State–church cleavage
regime. Political parties are the product of the parliamen- Nation-states in the nineteenth century were not only
tary and electoral game, and party systems reflect the centralized and homogeneous, but also based on the lib-
social oppositions that characterize society when parties eral ideology promoting secular institutions (no church
first appear. The fundamental features of today’s party influence), individualism, and democracy (sometimes
systems were set during the early phases of mobilization republicanism). Liberal reforms and the abolition of es-
of, at first, restricted electorates (only very few people had tates (clergy, aristocracy, bourgeoisie, peasantry) of pre-
the right to vote when liberals and conservatives domi- modern parliaments, as well as individual vote and free
nated in the nineteenth century) and, later, of ‘massifying’ elections, put an end to clerical and aristocratic privilege.
electorates when socialist parties mobilized the vast work- In this, liberals were opposed by conservatives. To a
ing class that emerged from the Industrial Revolution. large extent, this was a conflict between the rising indus-
The National Revolution produced two cleavages. trial bourgeoisie and the corporate privilege of clergy and
aristocracy.
The new liberal secular state fought against the long-
Centre–periphery cleavage established role of the church in education. Compulsory
This conflict emerged when nation-states formed and in- education by the state was used to ‘forge’ citizens with
tegrated in the nineteenth century, and political power, new (non-religious) values. Especially in Catholic coun-
administrative structures, and taxation systems were tries this led to strong conflicts, whereas in Protestant
centralized. It also brought about national languages countries—where churches belong to the state—the
and the adoption of a national religion. Some states were cleavage focused on moral principles. The church was
old, some formed through unification and some through also expropriated of land and buildings and, in Italy, it
secession. Most national territories were heterogeneous lost its temporal power and state (about a fourth of the
with different ethnicities and languages, and adminis- Italian peninsula) when Italy unified from the previous
tration was fragmented. Nationalist and liberal elites state mosaic in 1860–70.
carried out state formation and nation-building, facing The conflict characterized the opposition to liber-
resistance from subject populations in peripheral territo- als against the conservatives, who believed in a return
ries in two aspects. to the old pre-democratic regime. In some countries,
1. Administrative: peripheries were increasingly incor- Catholics took the place of conservatives, as in Belgium,
porated in the bureaucratic and fiscal system of the Switzerland, and Germany. In other countries, Catholics
new state (for example, with the creation of provinces were banned through papal decree from participating
or departments through which the central state con- in the political life of the liberal nation-state (by being
trolled the territory of and extracted taxes), implying candidates, voting, or creating a party). For this reason
a loss of autonomy for regions. Catholic parties did not appear in Italy and France until
the early 1920s. In fact, it was not until after the break-
2. Cultural: religious, ethnic, and linguistic identities in
down of democracy and the inter-war fascist period that
peripheral regions were replaced by the allegiance to
the Catholic Church fully accepted democracy. ‘Chris-
the new nation-state fostered through compulsory
tian democracy’—in Italy, France, Germany, Austria—is
schooling, military conscription, and other means of
the family that appears from this evolution after the
national socialization. As the first Italian prime minis-
Second World War.
ter said in 1870 after Italy unified, ‘we have made Italy,
An interesting case is that of countries with mixed
let us make Italians’. Nation-building took place also
religious structures. In the Netherlands there was one
in old established states. In France in 1863, according
unified Catholic party and a number of Reformed and
to official figures, only 22 per cent of the communes
Calvinist parties reflecting the fragmentation of Prot-
spoke French, all located around the Paris region
estantism. Religious parties merged in 1972 into the
(Weber 1976: 67).
Christian Democratic Appeal. In Germany, too, an
Resistance to administrative centralization and cultural inter-confessional party developed (the Christian Dem-
standardization was expressed in regionalist parties ocratic Union). In Switzerland a major Catholic party
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 241

emerged from the opposition to the Protestant Radicals/ as this split is present in all countries, it is the most im-
Liberals. portant one. Left–right is the most common ideological
The Industrial Revolution produced two additional dimension along which parties are placed (even in the US
cleavages. where a socialist party never developed: see Box 13.1).
Industrialization had a very deep impact on West-
ern societies. It radically changed the production
Rural–urban cleavage mode, it caused unprecedented levels of geographi-
The first was the contrast between landed rural interests cal mobility through urbanization (the dislocation of
(agriculture) and the rising class of industrial and trading people from countryside to urban industrial centres),
entrepreneurs. This cleavage focused on trade policies, it transformed family structures. Workers moved to
with agrarians favouring trade barriers for the protection new industrial centres where living conditions were
of agricultural products (protectionism) and industrial- extremely poor. They were therefore easy to mobilize
ists favouring free market and trade liberalization with through trade unions, with socialism providing a unify-
low tariffs (liberalism). This cleavage was reinforced by ing ideology. With the extension of voting rights social
cultural differences between countryside and urban cen- democratic and labour parties gained parliamentary
tres where industries concentrated. Cultural openness/ representation.
closure added to the division between primary and sec- Socialist parties campaigned for labour protection
ondary sections of the economy. against the capitalist economy. They promoted social
Weak sectors of the economy tend to be protectionist rights and welfare state provisions on top of civil and
because of the threat of imports, whereas strong sectors political rights, and a substantial equalization of living
favour the opening up of economic borders which favour conditions besides formal legal equality (Marshall 1950;
exports (Rogowski 1989). Agriculture was threatened Kitschelt 1994). These claims concerned under-age and
by technological progress and acceleration of produc- female labour, wages, working hours, contract security,
tivity. The defence of agrarian interests—when peasant protection in the workplace and during periods of un-
populations received the right to vote—was expressed employment or illness, progressive taxation, abolition
from the end of the nineteenth century through agrarian of heritage, accident insurance, and pension schemes.
parties (also called peasants’ or farmers’ parties). Large Socialists favoured economic policies with a strong
or small agrarian parties existed everywhere in Europe intervention of the state in steering the economy and
but were particularly strong in Eastern Europe and in public investments (later Keynesianism) against the lib-
Scandinavia. They also existed in Latin America. eral free-market ideology. They looked for state owner-
The period after the Second World War witnessed ship of infrastructure (transportation, energy), indus-
both the decline and transformation of these parties. tries, and sometimes financial institutions.
On the one hand, in most countries peasants’ parties Many socialist and labour parties originate from previ-
disappeared. On the other, the large agrarian parties ously existing trade unions, the main organizations of the
of the north and east abandoned the agrarian platform working class before universal suffrage. With restricted
and changed into centre parties. The recent reawaken- franchise most workers did not have the right to vote.
ing of this cleavage is most notable in Latin America The state was therefore controlled for most of the nine-
where opposition to multinational companies, defence teenth century by liberals and conservatives who were
of raw materials and resources, and the threat of global- able to impose their policies. Unions responded to a
ization has led to protectionist policies (e.g. gas and oil number of needs of the working class, increased solidar-
nationalization in Bolivia and Venezuela). In the 1990s ity and cooperation within it, and provided a wide range
a number of upheavals of peasants took place in the of ‘services’. With enfranchisement, workers’ parties de-
Chapas region in Mexico. This cleavage is also present veloped as an ‘electoral branch’ of trade unions.
in the European Union where farmers’ pressure groups The Soviet Revolution of 1917 produced a cleavage
lobby for protectionist trade agreements and for state within the workers’ movement.
subsidies.
Communism–socialism cleavage
Workers–employers cleavage In the aftermath of the First World War and the Russian
This is the cleavage between the industrial entrepre- Revolution that led to the Soviet Union and the single-
neurial bourgeoisie who started the Industrial Revolu- party regime controlled by the Communist Party, in all
tion and the working class that resulted from it. It is the countries communist parties formed as splinters from
opposition between ‘capital’ and ‘labour’ which, up to the the socialists. The main issue was the acceptance of the
present, characterizes the left–right alignment. In so far lead of the Soviet Communist Party in the international
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

242 DANIELE CARAMANI

revolutionary movement and also ideological differences, Materialism–post-materialism cleavage

namely whether a revolution would be necessary to take
the proletariat to power, or if this goal would be achieved A cleavage between generations over sets of socio-political
through electoral means. values emerged in the 1960s and 1970s as a consequence
As a reaction against the radicalization of the working of the protracted period of international peace, eco-
class and its powerful action through a new type of mass nomic wealth, and domestic security since the end of the
party organization, fascist parties emerged in a number of Second World War (Inglehart 1977). The younger cohort
European countries and, more or less directly, dominated developed ‘post-materialist values’ focused on tolerance,
government during the 1930s. These parties favoured the equality, participation, freedom of expression, respect
nation over class and ‘internationalism’, and private proper- for the environment, fair international trade, peace, and
ty against communism. Fascist parties were the product of Third World aid; as opposed to the ‘materialist’ values
the radicalization of the industrial upper bourgeoisie threat- of the war generation centred around themes of national
ened by socialist policies, and of the aristocracy threatened security, law and order, full employment, protection of
by the redistribution of land and agrarian reforms. private property, tradition, and authority (within the
Finally, the ‘Post-Industrial Revolution’ (Bell 1973) family and the state).
created two more recent cleavages. These new values were primarily expressed in a num-
ber 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
Box 13.1 Why is there no socialism claiming equality in the labour market and family, and
in the US? environmentalism in the 1980s. In the 1990s, new anti-
globalization movements developed against the glo-
A number of classical studies have addressed this ques- balization of the economy and the Americanization of
tion. The main factors explaining the absence of a so- culture (Della Porta et al. 1999). From a party politics
cialist ideology and workers’ party in the most advanced perspective, however, there are only a few examples of
capitalist country are: a significant impact of these ‘new left’ movements, the
● Open frontier geographical and social mobility gave main being green parties (Müller-Rommel and Poguntke
American workers the possibility to move on in 2002). A more pervasive impact of the Post-Industrial
search of better conditions. Revolution is on the ‘new right’.
● Party machines dominance of Democrats and Re-
publicans in the nineteenth century made the rise of
The globalization cleavage
third parties difficult.
● The free gift of the vote working-class white men all
Economic globalization has created a further post-
had the right to vote, were integrated in the political
industrial cleavage between sectors of the economy
system, and had a say in government’s actions.
that profit from the blurring of economic boundaries,
and sectors that are negatively affected by the competi-
● Roast beef and apple pie the American working class
tion from new markets and cheap labour from the East
was more affluent than the European and all socialist
and Asia. ‘Losers’ in globalization and—in the European
utopias come to grief with a satisfied working class.
Union—integration (Betz 1994) have reinforced support
● No feudalism the absence of aristocracy in America for neo-populist protest parties who favour trade barri-
made the working class very similar to the European ers to protect local manufacture and ‘locals-first’ policies
bourgeoisie. in the labour market. These groups are the small and
Read: medium enterprises, unskilled workers, craftsmen, and
agricultural producers.
● Lipset, S. M. (1977) ‘Why No Socialism in the United
The economic defensive attitude of these groups is
States?’, in S. Bialer and S. Sluzar (eds), Sources of
reinforced by cultural, anti-immigration, and xeno-
Contemporary Radicalism (Boulder, Colo.: Westview
phobic prejudice stressing religious and national val-
Press), 131–49.
ues against multi-ethnic society and cosmopolitanism.
● and Marks, G. (2000) It Didn’t Happen Here:
Many of these parties rely upon an extreme right-wing
Why Socialism Failed in the United States (New York:
heritage, such as the Austrian Liberal Party, the French
Norton).
and Belgian National Fronts, the former Italian Nation-
● Sombart, W. (1976) Why is there No Socialism in the al Alliance (Kitschelt 1995). Others are more sporadic
United States? (London: Macmillan), translated from parties, such as the One-Nation Party in Australia. In
the German 1906 text. Latin America neo-populist tendencies have a left-wing
‘Bolivarian’ character as the movements led by Morales
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 243

in Bolivia and Chavez in Venezuela (Burgess and Levitsky cleavage constellations and party systems have re-
2003). Neo-populism is also a reaction to changing se- mained extraordinarily stable. Up to the present even
curity conditions which—since the terrorist attacks in party labels have not changed (liberal, socialist, conser-
the early 2000s—have created a resurgence of materialist vative), as a sort of political imprint that crystallized.
values, the need for internal police and external control. Lipset and Rokkan have formulated the so-called freez-
ing hypothesis:

Variations in cleavage [T]he party systems of the 1960s reflect, with few
constellations but significant exceptions, the cleavage structures
of the 1920s… [T]he party alternatives, and in re-
Cleavage constellations change through space (from markably many cases, the party organizations, are
country to country) and over time. older than the majorities of the national electorates.
(1967: 50; italics omitted)
Space
Today’s party systems reflect the original conflicts from
Not all cleavages exist in all countries. There is a vari- which they emerged (see Box: Party families, in the On-
ety of constellations, and thus of party systems. Why do line Resource Centre) in spite of a decline in cleavage
some cleavages exist in specific countries while not in politics with the blurring of social divisions (Franklin
others? It is difficult to summarize the explicative part of 1992). In the 1920s the full mobilization of the elector-
the Lipset–Rokkan model here. Whereas the left–right al market through universal suffrage and PR caused its
cleavage exists everywhere and is a source of similar- saturation. With the extension of suffrage citizens were
ity, the state–church cleavage developed especially in incorporated in the political system. Voters acquired
Catholic countries in Europe and Latin America. The strong political identities through partisan identification
rural–urban cleavage was strong in regions with small and socialization processes that proved stable over time.
farming and independent units, where farmers were not As in all markets, in the electoral market too there are
under the control of landlords. The centre–periphery cleav- entry barriers. Little room was left for new parties. Exist-
age appears where there are ethno-linguistic minorities. ing parties were thus able to maintain their control over
Country-specific cleavage constellations are therefore electorates through the generations.
determined by: Empirical research has confirmed the basic stability
● Differences in objective factors such as diverse social of electoral patterns over time, rejecting the thesis of in-
structures: multiple ethnicities or religious groups, creasing dealignment and realignment of Western elector-
structure of the peasantry, class relations. ates (Dalton et al. 1985). Rose and Urwin (1970) and, in
a long-term perspective, Bartolini and Mair (1990) have
● The extent to which socio-economic and cultural divi-
analysed trends of electoral volatility (the change of votes
sions have been politicized by parties, that is, by the
from one election to the next) from 1885 to 1985. First,
action of elites (Rose 1974; Lijphart 1968b).
they found that general levels of volatility are stable sup-
● The relationship between cleavages: their existence porting the freezing hypothesis. Second, they found that
and strength can prevent the development of new ones volatility between left and right declines, confirming the
(agrarian claims have been incorporated by Catholic stabilizing of ideological identities. Third, however, they
parties or by conservative parties where, as in England, found that volatility within the left and within the right in-
agriculture had been commercialized early). creases which means that, whereas left and right identities
Generally, two types of constellations are distinguished: persist, the identification with a specific party declines (see
(1) homogeneous constellations where there is one pre- also Kriesi 1998). Therefore, in spite of secularization and
dominant cleavage, namely the left–right cleavage on the post-industrial economies—and in spite of some degree of
distribution of resources between classes (for example, change within Western electorates—a dramatic realign-
Britain), and (2) heterogeneous constellations in which ment along new cleavages does not seem to have taken
various cleavages—economic, ethno-linguistic, religious, place, maintaining the validity of the freezing hypothesis.
territorial—overlap or cut across one other in plural de-
mocracies such as Belgium, Canada, India, the Nether-
KEY POINTS
lands, and Switzerland (Lijphart 1984).
● Modern party families originate from socio-economic
Time and cultural cleavages created by industrialization,
urbanization, and the formation of centralized liberal
Lipset and Rokkan do not take into account develop- states.
ments that took place after the 1920s as, over time,
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

244 DANIELE CARAMANI

● The centralized and democratic liberal state creates

of parliamentary seats) over protracted periods of time
conflicts with the church and with peripheral regions,
(several decades). In these systems all parties are legal
leading to religious and regionalist parties. Industrial-
and allowed to compete in free elections with universal
ization opposes liberal economic interests to the rural
suffrage to challenge the dominant party. However, no
world as well as to the working class, leading to agrar-
other party receives enough votes to come close to 50
ian and labour parties.
per cent. Electors vote massively for the dominant par-
ty. There is, therefore, no alternation in power and the
● The introduction of universal suffrage and PR after the
dominant party does not need to build coalitions to form
First World War ‘freezes’ the party constellations that
a government.
remain stable until the present.
An example of a dominant-party system is In-
● The most significant examples of realignment in re- dia between 1947 and 1975. After Independence the
cent time are the generational cleavage over (post-) Congress Party received votes above 50 per cent and
materialist values and the economic changes trig- was able to rule unchallenged until 1975–77 when
gered by globalization that led to new party families: the ‘state of emergency’ was declared. Over the long
the greens and the neo-populist parties. period of uncontested rule forms of patronage devel-
oped and in 1977 the Congress Party was eventually
defeated. A more recent example of a dominant-party
The morphology of party systems system is South Africa since the end of apartheid in
the early 1990s. The African National Congress, ini-
An important element of the competitive interaction be- tially led by Nelson Mandela, has been able to secure
tween parties is the shape of party systems. The two main the absolute majority of the votes because of the role it
elements of the morphology of party systems are: (1) had in enfranchising the black population. In Europe a
the number of competing units, that is, parties, and (2) case of a dominant-party system is Sweden. The Social
the size of these units. How many are the players and Democratic Workers’ Party formed almost all govern-
how strong are they? The number and strength of actors ments from 1945 until 1998, with around 45 per cent of
can be observed at two levels: the votes parties get in elec- the votes on average. Only in a few cases did it have to
tions and the seats in parliament. A ‘variable’ that must form a minority government or rely on small coalition
be considered is, therefore, the electoral system through partners such as the formerly communist Left Party. In
which votes are translated into parliamentary seats. Mexico, the Institutional Revolutionary Party was in
It is important to distinguish types of party systems. power from the revolution of 1917 until the 2000 elec-
Two types of party systems are not considered in this tion when it was defeated for the first time.
section because they do not fulfil the democratic condi- Between 1946 and 1994 the Italian Christian
tions that allow competition: Democracy was uninterrupted in power. However, only
1. Single-party systems in which one party only is legal: until 1953 did this party receive an absolute majority of
these are the totalitarian and authoritarian experienc- the seats and was able to form a single-party government.
es of the Communist Party in the Soviet Union, the After 1953 it relied on small coalition partners. A simi-
Nationalist-Socialist Party in Germany in the 1930s, lar case is the Liberal Democratic Party in Japan between
or the Baathist Party in Iraq until 1993 and in Syria. 1953 and 1993. Factions developed within both parties
2. Hegemonic-party systems in which other parties are and, because of the lack of alternation, they became less
legal but are ‘satellites’, under the strict control of the responsive to the demands from the electorate with
hegemonic party with whom they cannot compete to forms of patronage (Shiratori 2004: 105).
control government: these are also totalitarian or au-
thoritarian systems existing in Egypt or Algeria today,
and in many former communist regimes before 1989
Two-party systems
in Central and Eastern Europe. A two-party system is one in which two fairly equally
balanced large parties dominate the party system and al-
The other four types are: (1) dominant-party system, (2)
ternate in power. The two parties have comparable sizes
two-party system, (3) multiparty system, and (4) bipolar
and equal chances of winning elections. Even a small
system.
amount of votes changing from one party to the other
(electoral swing) can cause a change of majority. Alter-
Dominant-party systems nation in power is therefore frequent. These are very
competitive systems. Because both parties are large, the
Dominant-party systems are characterized by one very winning party is likely to receive the absolute majority
large party that dominates all others with a large ma- of seats and form single-party governments without the
jority (well above the absolute majority of 50 per cent need for partners.
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 245

The features of two-party systems are those listed For years Germany came close to a two-party system
in Table 13.2. The two large parties have similar sizes and was named a ‘two-and-a-half party system’ with two
around 35–45 per cent of the votes each, that plural- large parties collecting together more than 80 per cent
ity electoral systems transform into absolute majorities of the votes (the Christian–Democratic Union and the
of seats for the largest party. This does not mean that Social Democratic Party) and a smaller Liberal Party
these are the only parties. A number of other smaller (around 5 per cent) with a pivotal position and able to
parties compete in the elections. However, they are mar- decide—through alliance—which of the larger parties
ginal as they are not necessary to form a government. In would be in charge of government. With the rise of the
the United Kingdom, the Liberal Democratic Party did Greens the system turned towards a bipolar system. Is-
not have a strong impact for long on the party system; rael has used a PR electoral system since the creation of
nor did the Scottish National Party or Plaid Cymru (the the state in 1948. Yet until the end of the 1990s the sys-
Welsh national party). tem was structured around two main parties: the Likud
In two-party systems single-party governments tend and the Labour Party.
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) Multiparty systems
systems is very high, the two main parties have a ma- Multiparty systems are the most frequent type of party
joritarian vocation. Parties propose policies and pro- system. These are also the most complex types. In a mul-
grammes that are acceptable to a large part of society. tiparty system the number of parties ranges from three to
Plurality leads to ideological moderation and similarity double-digit figures. Three to five parties exist in Canada,
of programmes. In turn, this similarity makes it easier Ireland, Japan, and Norway. Party systems in which the
for voters to switch from one party to the other and number of parties approaches ten (or even more) are
creates alternation. Belgium, the Netherlands, and Switzerland. None of the
There are not many cases of two-party systems. These parties in a multiparty system is majoritarian (with 50
systems are typical of the Anglo-Saxon world where— per cent of the votes or seats). Furthermore, the parties
unlike continental Europe where around the First World that compose a multiparty system are of different sizes:
War all countries changed from majoritarian to PR elec- some are large (say, 30 per cent of the votes) some small
toral systems—plurality in single-member districts has (less than 5 per cent).
been maintained. In addition, the trend seems to be de- Because in multiparty systems no single party has an
clining. Only the US provides today a ‘perfect’ example overall majority the result is that parties must form co-
of a two-party system where Republicans and Democrats alitions in order to support a government. In parliamen-
have dominated since 1860.1 Australia maintains a strong tary systems (see Chapters 5 and 7) the vote of confidence
two-party system with the Australian Labour Party and requires a 50 per cent majority of seats. Parties run indi-
the Liberals. In Great Britain the Conservatives and the vidually in elections (contrary to bipolar systems) and gov-
Labour Party have been increasingly challenged by the ernmental coalitions are negotiated after the results are in.
Liberal Democratic Party and in 2010, for the first time Unlike plurality in single-member constituencies, PR
since the second World War, a coalition government does not hinder small parties from addressing small seg-
was formed. Other examples include Costa Rica (Na- ments of the electorate, sometimes through extreme
tional Liberation Party and Citizens’ Action Party) and ideologies and programmes. PR therefore does not lead
Malta (where the Labour Party and the Nationalist Party to ideological moderation which, in turn, makes it more
receive together close to 100 per cent of the votes). In difficult for voters to switch from one party to the other
Canada Conservatives and Liberals dominated until 1993 and cause a government change. In addition, PR does
(with a strong New Democratic Party), since when the not provide the ‘amplification’ effect of electoral swings
Bloc Québéquois and the Reform Party have been in- as does plurality. As a consequence, government change
creasing their support. rarely takes place through electoral change but rather by
Two-party systems can be found also in countries with swaps of coalition partners.
PR electoral systems. Austria, since the Second World While multiparty systems are considered to repre-
War, has been dominated by two parties—the Austrian sent better socio-political pluralism in countries with
People’s Party and Austrian Socialist Party. After the religious, regional, and ethno-linguistic cleavages, their
transition from Franco’s regime to democracy in 1977, negative aspects have been at the forefront since the
the party system of Spain moved towards a two-party Second World War. Multiparty systems were held re-
system. In spite of many (but small) regionalist parties, sponsible for instability, frequent coalition ‘crises’, and
the party system in Spain presents two large parties of a poor responsiveness, with no single party clearly ac-
similar size: the Spanish Socialist Workers’ Party and the countable. Classical political scientists such as Finer
People’s Party. (1932), Hermens (1941), Duverger (1954), and Almond
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

246 DANIELE CARAMANI

Table 13.2 Types of party systems in democracies

Type of party system Features Cases

Dominant-Party One large party with more than India until 1975, Japan between
absolute majority of votes and 1955 and 1993, Mexico until
seats. 2000, South Africa since 1994.
No other party approaching 50%.
No alternation.
One-party government.
Two-party Two large parties sharing Austria, Britain, Costa Rica,
together around 80% of votes Malta, New Zealand until 1998,
and seats. Spain, South Africa until 1989,
Balanced (35–45% each) with one US.
of the two reaching 50% of seats.
Alternation between parties.
One-party government.
Multiparty Several or many parties, no one Belgium, Canada, Colombia,
approaching 50% of votes and Czech Republic, Denmark,
seats. Finland, Germany until 1989,
Parties of different sizes. Hungary, Italy before 1994,
Parties run for elections indi- Netherlands, Poland, Russia,
vidually and form coalitions after Switzerland, Turkey.
elections.
Alternation through coalition
changes.
Coalition government.
Bipolar Two large coalitions composed of France in the Fifth Republic,
several parties sharing together Germany since 1990, Italy since
around 80% of votes and seats. 1994, Portugal.
Coalitions are balanced (40–50%
each).
Coalitions are stable over time
and run elections as electoral
alliances.
Alternation between coalitions.
Coalition government.

(1956) blamed PR and multiparty systems for the lack of Moderate multiparty systems
ideological moderation in the 1920s and 1930s which
eventually led to the breakdown of democracy in most The logic is similar to that of two-party systems. First,
continental European countries. the number of parties is limited (below five) and, second,
Positive aspects of PR and multiparty systems the direction of the competition is centripetal, that is,
have been stressed since analysis in the 1960s and the main parties tend to converge toward the centre of
1970s including small countries such as Belgium, the the left–right scale to attract the support of the moder-
Netherlands, Switzerland, and the Scandinavian coun- ate electorate. At the centre are one or more small par-
tries. Studies on ‘consensus democracies’ showed that ties with whom the two big ones on either side may form
multiparty systems are stable, functioning, and peace- a coalition. The role of these small parties is ‘pivotal’ in
ful. In plural societies PR and multiparty systems are the that they can decide whether the coalition is going to be
only viable ways to involve minorities in decision-mak- centre-left or centre-right. The ideological distance be-
ing processes and reach consensus.2 As Chapter 5 shows, tween parties is limited so that all coalitions are possible.
consociational or consensus democracies represent a This type of party system is named ‘moderate’ because of
different model of democracy from the majoritarian or the absence of extreme parties.3
‘Westminster’ model. Both have advantages and disad-
vantages (see Box 13.2). Polarized multiparty systems
The way in which multiparty systems function largely
depends on the degree to which parties are ideologically There are three main features in polarized systems.
polarized. Sartori (1976) has distinguished two main 1. Polarization. There is a large ideological distance
types of multiparty systems. between parties with a strong dose of dogmatic
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 247

Box 13.2 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 After the First World War in Italy, Weimar Germany,
breakdown of democracy between First and Second Spanish Second Republic, and in the French Fourth
World Wars: Britain and US. Republic (1946–56) instability led to crisis of democracy.
Effective Ineffective
Produces governments immediately after elections. Governments take long to form after elections because of
Governments are stable because they are formed by a negotiations between parties. Coalitions lead to unstable
single party. governments.
Accountable Non-accountable
Because there is only one party in government respon- Because governments are formed by many parties ,re-
sibility is clearly identifiable by the electorate. sponsibility is obfuscated.
Alternation No alternation
Two main parties alternate in power. Voters influence Coalition negotiations are out of the reach of voters’
directly the formation of goverment: a small shift can influence and shifts of votes are not necessarily followed
cause government change. by changes of government.
Non-representative Representative
FPTP under-represents minorities and over- PR fairly represents minorities in societies with ethno-
represents large mainstream parties of left–right. linguistic and religious minorities.
Moderate politics Extreme politics
All main parties have a chance to govern and thus Multiparty systems allow representation of extreme
avoid extreme claims. Need to gather votes from large (anti-system) parties. Some do not have any government
moderate segments of the electorate. prospect and do not hesitate to radicalize their claims.
Discontinuity Continuity
Decisions are made by majority with a clear strategy Decisions are made by consensus through consultation.
but there can be discontinuity between subsequent More difficult to find a clear strategy but more continuity
goverments. Legislation is often in legislation.
reversed.

radicalism. Extreme anti-system parties aim not only occupied. As a consequence, there is divergence. The
to change government but also the system of govern- competition is centrifugal.
ment (the regime). These parties do not share the
Examples of polarized multiparty systems are the
principles of the political system and aim to change
Weimar Republic in Germany, from 1919 until 1933,
its institutions (Capoccia 2002). Given the ideological
and Italy between 1946 and 1992. Italy in particular
distance between parties not all coalitions are viable.
has often been taken as an ideal type. A strong centre
Some parties are continuously excluded. Such parties
party, Christian Democracy, was opposed on both sides
are in constant opposition, become irresponsible, and
by unreformed anti-system parties: the Italian Commu-
radicalize their discourse with promises they know
nist Party and Italian Social Movement (a post-fascist
they will never be called to put into practice.
party). Coalitions between Christian Democracy and
2. An occupied centre. There is one main party placed at either anti-system parties were not viable and Christian
the centre of the left–right axis which represents the Democracy ruled uninterrupted until 1992, although the
‘system’ against which extreme anti-system parties Communists had a large share of votes (35 per cent).
are opposed. The centre party is always in power and
becomes irresponsible and unaccountable. This party
is not punished electorally because of the absence of Bipolar systems
viable alternatives. Bipolar party systems combine elements of both multi
3. Centrifugal competition. The occupation of the centre and two-party systems. As in multiparty systems there
discourages a centripetal move on the part of other are many parties, none of which has a majority. And,
parties because, ideologically, the centre is already again, coalition governments are the rule. However,
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

248 DANIELE CARAMANI

coalitions—rather than single parties—are the important all parties are squared to weight parties through their
players. These form before elections and run as electoral size. If there are two parties A and B, receiving each 50
alliances. They remain stable over time. There are usu- per cent of the seats, one calculates first the square for
ally two large ones of evenly balanced size alternating party A (.50 × .50 = .25) and for party B (.50 × .50 = .25)
in power. Competition therefore resembles that of two- and then adds them together (.25 + .25 = .50). Thus:
party systems.
F = 1 − .50 = .50 E = 1/.50 = 2
In France left and right have alternated in pow-
er since 1958.4 The left includes Socialists, Radicals, In this example, F is exactly between zero and one (.50)
Communists, and Greens, whereas the right includes and E counts perfectly that there are two parties only. In
Gaullists and Liberals (they merged in 2003 as the Union the real world, the distribution of power among parties is
for a Popular Movement). In Italy since 1994 the centre- obviously more complex.
left coalition is composed of Social Democrats, Commu- Table 13.3 lists the effective number of parties (based
nists, Greens, and Catholics, whereas the centre-right on seat distributions) in a number of countries for recent
coalition is of Silvio Berlusconi’s party (that recently elections.5 As one can see there is a great variation be-
merged with the post-fascist party) and the Northern tween countries. The less fragmented countries are those
League. The coalitions have alternated in power in 1996, using plurality/majoritarian or transferable vote systems
2001, 2006, and 2008. In Germany, finally, two coalitions in single-member districts (Australia, France, Great
oppose each other: Social Democrats and Greens on Britain, Hungary, Malta, the US), whereas the most frag-
the one hand, Christian–Democratic Union, Christian– mented countries are those with PR and many religious
Social Union, and Liberals on the other. and ethno-linguistic parties (such as Belgium, Finland,
the Netherlands, New Zealand, Norway, Switzerland).

The number of parties Qualitative rules

So the number of parties is important, but how, exactly, In many cases it is not appropriate to consider numerical
should parties be counted? If all parties that run in an criteria only to decide whether or not a party is relevant.
election are counted (or even only those that get some Often small parties—that quantitative rules would not
votes) their number would be extremely large and useless count—have far-reaching consequences for coalitions,
in building a typology. In every election there are dozens influencing important decisions, mobilizing people in
of parties and candidates that do not get any votes or very demonstrations, and so on. In many cases, small parties
few. It is therefore necessary to have reasonable rules to are much more important than their sheer size would
decide what and how to count. There are two ways to suggest. Sartori (1976) has developed two criteria to de-
count parties: (1) numerical with indices based on the cide which parties really ‘count’ and should be ‘counted’:
size of parties; (2) qualitative with rules based on the role
1. Coalition potential: a small party is irrelevant if over a
of parties in the system.
period of time it is not necessary for any type of gov-
ernmental coalition. On the contrary, a party must be
Numerical rules counted if, disregarding its size, it is pivotal and deter-
mines whether or not a coalition is going to exist and
These rules represent quantitative attempts to classify
which.
party systems on the basis of the number and size of
parties that compose them. Various indices have been 2. Blackmail potential: a small party must be considered
devised to summarize this basic information: are there relevant when it is able to exercise pressure on gov-
many small parties (a fragmented party system) or few ernmental decisions through threats or veto power
large parties (a concentrated party system)? and by doing so alter the direction of competition.
The most used indices are Rae’s fractionalization
index (Rae 1971) and the effective number of parties
(Laakso and Taagepera 1979). The fractionalization in- The influence of electoral laws on
dex (F) varies from zero (full concentration of seats or the format of party systems
votes in one party) to one (total fragmentation with each
seat or vote going to a different party). The effective Given the impact of party system fragmentation on gov-
number of parties (E) indicates the number of parties in ernment stability, accountability, and responsiveness, as
a system and does not have an upper limit. well as on the type of consensus vs. majoritarian deci-
The two formulas are the following: sion-making, a large amount of comparative politics has
been concerned with establishing the causes for varying
F = 1 − ∑pi2 E = 1/∑pi2
numbers of parties and their size. Two sets of causes
where p is the percentage of votes or seats for party i and have been identified: (1) the electoral system and (2) the
∑ represents the sum for all parties. The percentages for number of cleavages in the society.
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 249

Table 13.3 Rae’s parliamentary fractionalization index (F), effective number of parliamentary
parties (E), and Gallagher’s index of disproportionality (LSq)

Country Election F E LSq

Australia 2007 0.60 2.4 8.7
Austria 2008 0.79 4.3 3.3
Argentina 2009 0.81 5.3 13.5
Belgium 2007 0.89 7.9 3.8
Brazil 2002 0.88 8.5 3.7
Canada 2006 0.69 3.2 8.7
Chile 2009 0.82 5.6 6.8
Czech Republic 2006 0.74 3.1 6.3
Finland 2007 0.83 5.1 3.4
France 2007 0.76 2.5 14.7
Germany 2009 0.78 4.1 2.7
Greece 2007 0.67 2.6 7.1
Hungary 2006 0.63 2.4 4.4
India 2009 0.86 7.1 3.9
Ireland 2002 0.70 3.3 6.6
Israel 2009 0.87 7.8 3.2
Italy 2008 0.74 3.1 6.5
Japan 2009 0.56 2.3 19.0
Malta 2008 0.52 2.0 1.5
Mexico 2009 0.67 3.0 6.4
Netherlands 2006 0.83 5.5 1.3
New Zealand 2005 0.66 3.0 1.4
Norway 2005 0.78 4.6 2.9
Poland 2007 0.70 2.8 4.7
Portugal 2005 0.61 2.6 5.9
Russia 2007 0.70 3.4 10.7
Spain 2008 0.63 2.3 5.1
Sweden 2006 0.78 4.2 3.5
Switzerland 2007 0.82 5.0 3.3
Turkey 2007 0.46 1.9 27.4
United Kingdom 2010 0.59 2.5 16.8
United States 2008 0.50 2.0 3.9
Mean 0.69 3.9 7.3

Notes: For calculations parties rather than alliances have been considered (France, Chile, Italy). Seat figures al-
ways based on final allocation. For mixed electoral systems, PR votes have been taken (Japan, Mexico, Hungary).
For Germany Zweitstimmen have been used and in France first-ballot figures for votes. As a general rule for
including parties in the calculation, all parties polling at least 1 per cent, or securing at least one seat, have been
taken into account.
Source: Author’s calculations based on results in ‘Country Profiles’ (see also the Online Resource Centre).

Electoral systems majoritarian systems in single-member constituencies;

(2) PR systems in multimember constituencies. The
Electoral systems are mechanisms for the translation of first and best-known formulation of the causal relation-
votes into parliamentary seats. Chapter 10 shows that ship between electoral and party systems is Duverger’s
there are two main ‘families’ of electoral systems: (1) Laws from his classic book Les Partis Politiques
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

250 DANIELE CARAMANI

(1951, translated in 1954). As can be seen in Box 13.3, the Psychological effects refer to the awareness of voters
two laws are simple: plurality or majoritarian electoral and parties of mechanical effects:
systems favour two-party systems whereas PR leads to
multiparty systems. This causal relationship between 1. On the demand side (voters), in electoral systems in
electoral and party systems is due to both mechanical which only large parties have a chance to win seats,
and psychological effects. voters tend to vote strategically (not necessarily their
Mechanical effects refer to the formula used to trans- first party preference) to avoid wasting votes on small
late votes into seats. In single-member constituencies parties with no chance of getting seats. Converging
to win the one seat is difficult. One party with the most votes on large parties reduces their overall number.
votes gets the single seat. The second, third, fourth, and On the contrary, with PR in which small parties can
so on, do not get any seat (first-past-the-post). If in a con- win seats, voters vote sincerely (their first preference)
stituency Party A receives 29.4 per cent of votes, Party B because their vote is not wasted. This increases the
29.3 per cent, and all other parties even less, only Party vote for small parties and thus their overall number.
A is represented (winner-takes-all). This means that the 2. On the supply side (parties), with plurality small
threshold is high and all parties but the first one are elim- parties have an incentive to merge with others to
inated. With PR, on the contrary, in each multimember increase their chances to pass the threshold, reducing
constituency many seats are allocated in proportion to the number of parties. On the contrary, with PR par-
the votes. If Party A receives 32.4 per cent of votes, it has ties have no incentive to merge: they can survive on
a right—more or less—to a third of the seats allocated in their own and small splinter parties are not penalized.
that constituency. Small parties are not excluded (a party This increases the overall number of parties.
with 5 per cent of votes gets roughly 5 per cent of seats)
and the overall number of parties that end up in parlia- Rae (1971), Riker (1982), and Sartori (1986) have ques-
ment is much higher than under single-member plurality tioned these laws by asking whether the reductive effect
systems. of majoritarian electoral systems works at the constituency

Box 13.3 The influence of electoral systems on party systems

Duverger’s ‘laws’ (1954) Sartori’s ‘tendency laws’ (1986)

First Law Law 1

‘The majority [plurality] single-ballot system tends to party ‘Given systemic structuring and cross-constituency disper-
dualism.’ sion (as joint necessary conditions), plurality systems cause
(are a sufficient condition of) a two-party format.’
Second Law
‘The second ballot [majority] system or proportional repre- Law 2
sentation tend to multipartyism.’ ‘PR formulas facilitate multipartyism and are, conversely,
hardly conducive to two-partyism.’
Mechanical effects
Electoral systems with high thresholds of representation Cox’s ‘coordination argument’ (1997)
(first-past-the-post) exclude small parties from parliament
‘Why … would the same two parties necessarily compete
whereas PR allows small parties to win seats.
in all districts [cross-constituency dispersion or nation-
Psychological effects alization]?’ Local candidates link together because of the
Under plurality systems voters vote strategically avoiding need to coordinate in order to compete more effectively
small parties; parties have an incentive to merge to pass for (1) seats to implement policies, (2) support presidential
high thresholds of representation; under PR voters vote candidates, (3) elect the prime minister, (4) obtain more
sincerely for small parties which are not penalized and upper-tier seats, and (5) obtain more campain finances.
have no incentive to merge. ‘If a system (1) elects legislators by plurality rule in sin-
gle-member districts; (2) elects its chief executive by some-
Rae/Riker’s ‘proposition’ (1971, 1982) thing like nationwide plurality rule; and (3) holds executive
and legislative elections concurrently, then it will tend
‘Plurality formulae are always associated with two-party
to … have a national two-party or one-party-dominant
competition exept where strong local minority parties
system.’
exist.’
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 251

level or at the national level. At the constituency level the on single-member constituencies such as France with a
high threshold reduces the number of parties. But does two-ballot majoritarian system. On the contrary, dispro-
this always translate into a reduction at the national level? portionality is lower for countries with PR systems.
Suppose a parliament has 100 seats from 100 sin- The number of parties in parliament is always smaller
gle-member constituencies. If in each constituency a than the number of parties that run for elections. All
different party wins the seat, we would end up with a electoral systems—also PR systems—have a reductive
fragmented parliament. The question thus is: under what effect on the number of parties. Differences among PR
conditions does the reductive effect of FPTP at the con- systems are large. Spain’s system is more disproportional
stituency level also reduce the number of parties at the than others. This is because in Spain the magnitude of
national level? The answer is: majoritarian systems pro- constituencies is small. The magnitude refers to the num-
duce two-party systems at the national level only if par- ber of seats allocated in a given constituency. The larger
ties are ‘nationalized’, i.e. receive homogeneous support the magnitude, the higher the proportionality between
in all constituencies (see Cox in Box 13.3). If there are votes and seats. If the magnitude is small, the few seats
parties with territorially concentrated support, this leads go to few parties and are harder for small parties to win.
to fragmentation in the national party system. Under Some PR systems artificially increase the number of seats
plurality, a nationally small party can be strong in specific (those using the Droop quota or the Imperiali quota) in
regions and thus win seats and create fragmentation in order to make it easier for small parties to get represent-
the national parliament. If many parties are territorially ed. In these systems the correspondence between votes
concentrated the national fragmentation is larger. and seats is greater.6
In most countries party systems nationalized with the
beginning of competitive elections in the mid-nineteenth
century, so the support parties receive is increasingly ho-
Cleavages
mogeneous across regions and territorialized support Large numbers of parties are also the result of social and
has declined. This can be observed in Europe, North cultural pluralism. The presence of numerous cleavages
America, but also in India and Latin America (Caramani leads to more parties than in culturally homogeneous
2004; Chhibber and Kollman 2004; Jones and Mainwaring countries. PR electoral systems were introduced in plural
2003) due to the development of national party orga- societies to incorporate minorities in the representation
nizations and increasing candidate coordination (Cox circuit and in decision-making processes. Plurality systems
1997). Where plurality systems exist, therefore, the would have excluded large segments of the society which
reduction of the number of parties did take place. Plu- then—dangerously—may have looked for other (non-in-
rality systems distort party votes when they translate stitutional) channels of action. As Colomer (2005) argues,
them into seats: PR electoral systems are the result of an already existing
political fragmentation, and not the other way round.
● They over-represent large parties (the share of seats
for big parties is larger than their share of votes).
● They under-represent small parties. KEY POINTS
How can we measure the empirical level of (dis)propor- ● The morphology of party systems is important for the
tionality between votes and seats? Various indices have competition between parties: it concerns the number
been devised: the most used one is the Least Square of players and their size. The main types are dominant-
index of disproportionality or LSq (Gallagher 1991; party, two-party, multiparty, and bipolar systems.
Gallagher and Mitchell 2005: appendix B):
● In two-party systems, moderate multiparty systems,
LSq = √1/2∑(vi − si)2 and bipolar systems competition is centripetal and
where v is the percentage of votes for party i, s is the per- alternation between parties or coalitions takes place.
centage of seats for party I, and ∑ represents the sum for In dominant-party systems and polarized multiparty
all parties. This index varies between zero (full propor- systems there is no alternation and competition is
tionality) and 100 (total disproportionality). Take, as an centrifugal.
example, the results of the 2005 New Zealand election ● Measures of fragmentation are based on the number
in Table 13.4. If the total of the squared differences is and size of parties. However, small parties, too, can be
halved (3.9/2 = 1.9) and then the square root is taken, important if they have coalition or blackmail potential.
the result is 1.4, that is, an almost perfect proportionality ● The format of party systems is influenced by electoral
between votes and seats. systems. Through mechanical and psychological ef-
In the last column of Table 13.3 values of the LSq index fects plurality tends towards two-party systems (large
are given. In countries with plurality systems (Canada, parties are over-represented) and PR to multiparty sys-
Great Britain, India) there is a stronger distortion of the tems (which are less disproportional).
popular vote. The same applies for other systems based
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

252 DANIELE CARAMANI

Table 13.4 Results of the 2005 New Zealand election and Gallagher’s LSq index of disproportionality

Party Votes (%) Seats (N) Seats (%) Difference Squared

(% seats–%
votes)
Labour Party 41.1 50 41.3 0.2 0.0
National Party 39.1 48 39.7 0.6 0.3
New Zealand 5.7 7 5.8 0.1 0.0
First
Green Party 5.3 6 5.0 −0.3 0.1
Māori Party 2.1 4 3.3 1.2 1.4
United Future 2.7 3 2.5 −0.2 0.1
ACT New 1.5 2 1.7 0.2 0.0
Zealand
J A’s Progres- 1.2 1 0.8 −0.4 0.1
sive
Others 1.3 0 0.0 −1.3 1.7
Total 100.0 121 100.0 3.9

Note: Votes refer to party lists votes whereas seats include both party list seats and electorate seats.

The dynamics of party systems Voters, like consumers, face alternatives which they
order from most to least preferred and choose the alter-
In the wake of Joseph Schumpeter’s (1943) defini- native that ranks highest. Voters make a rational choice
tion of democracy—a set of rules for selecting political by voting for parties whose programmes are closest to
leaders and making decisions by means of competition their policy preferences, because they are close to their
for votes—authors have developed analogies between interests or to their values and moral orientations. Vot-
electoral competition and market competition. In the ers vote on the basis of the proximity between parties’
electoral market, parties and candidates compete for positions and their preferences and so they know what
‘shares’ of the electorate as happens in the economic the alternative proposals by different parties are, that is,
world where firms compete for shares of the market. Par- they are informed about their possible choices.
ties are organizations whose main motive is the maximi- Rational citizens vote on the basis of a self-interested
zation of votes, and the exchange between represented calculation, like consumers who calculate the benefit
and representatives is similar to that between demand between ‘packages’. On the other hand, parties are like
and supply in the economy (see Table 13.5). businesses competing for customers. They establish
what people ‘like’ so they can sell more. Following a logic
of supply and demand, parties offer policies that voters
The market analogy can either choose to ‘buy’ or not. Once elected, parties
Anthony Downs’s An Economic Theory of Democracy seek re-election through policies appealing to large seg-
(1957) is a pioneering book in which the basic elements ments of the electorate. The parties’ goal is to make a
of these models were spelled out for the first time. In ‘profit’ in terms of votes; the voters’ goal is to maximize
this model, actors (parties and voters) are rational. Par- utility by buying a product that increases their satisfac-
ties calculate their strategies by formulating platforms tion. As in economic theory, the search for individual
with the goal of maximizing votes and being elected or advantages produces common goods, namely responsive-
re-elected. Parties are coalitions of individuals seeking to ness and accountability.
control institutions rather than the implementation of Rational choice competition models were first de-
programmes. Parties act self-interestedly to gain office. vised for two-party systems— mainly the US. However,
Like firms in the economic market, they are indifferent to maximize votes is the main motive also in systems in
to the ‘product’ they offer but interested in making profit which governments are coalitions. The more votes the
(monetary in the economic market and votes in the elec- better the chances to enter a coalition, control govern-
toral one). To maximize votes parties offer programmes mental institutions, and place individuals in key official
that appeal to a large part of the electorate. positions.
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 253

Table 13.5 The analogy between economic and electoral competition

Dimensions Economy Elections

Market Economic Electoral
Actors Firms Parties
Consumers Voters
Profit Money Votes
Supply Goods, services Programmes, policies
Demand Product preferences Policy preferences
Communication Advertising Campaigns

The spatial analogy they are from the bakery, the higher the ‘costs’ for buyers.
To what extent is relocation towards the centre tolerated
The idea of proximity/distance between individual pref- by residents of the extremes? Incentives for a new bakery
erences and parties’ policies indicates that players move at the edges of the village increase as people feel that AA
in a space of competition. The second element that and BB are too distant. The risk of strategies of relocations
Downs ‘imported’ from economic models of competi- toward the centre is that a new bakery C appears taking
tion is their spatial representation. In particular, Downs away part of B’s share of the market (the dark shaded area).
adapted models of the dynamics of competition between In these models there are therefore two dynamic elements:
firms, that is, where firms locate premises according to (1) the movement caused by the search for the optimal lo-
the physical distribution of the population. cation and (2) the appearance of new competitors in spaces
Let us take the simple case of a village in which there left uncovered. Equilibrium is reached when no competi-
is only one street (the example is from Hotteling 1929). tor has an interest in changing its position along the axis.
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? As- Downs’s model
suming that both bakers offer the same quality of bread Through the spatial analogy between physical and
for the same price and that consumers will rationally try ideological space, Downs imports these analyses into
to reduce their ‘costs’ by buying bread in the nearest shop electoral studies. Most elements are maintained: (1) the
(proximity), if A and B are located as they are in the fig- one-dimensionality of the space, (2) the principle accord-
ure, B will have a larger share of the market. The share of ing to which costs are reduced by choosing the nearest
B’s market goes from the right-side end of the street to option (proximity), (3) competitors’ search for the opti-
the M-point which is the middle between the locations of mal location through a convergence toward the centre.
A and B. Residents on the right of the M-point will buy Downs represented the ideological space through a
bread in bakery B and residents on the left of the M-point zero to 100 scale ranging from left to right. As will be
will buy bread in bakery A. The dynamic element in this seen, one-dimensionality is maintained, even if it is not
model consists of A’s move to increase its share of the always a realistic assumption, because it summarizes
market. By relocating the bakery in AA, the baker is able other dimensions and is the most important one (in
to gain the share of the market indicated by the dashed terms of size of parties that define themselves according
area. Obviously, B too can move toward the centre (BB) to this dimension), and because it is present in all party
and win back part of the lost share of the market. Both systems (thus allowing comparisons).
bakers seek to optimize their location. Both Hotteling and Smithies had previously applied
An additional element introduced by Smithies (1941) spatial models to politics through analogies with the ide-
concerns the elasticity of the demand. The further away ological space and were able to predict that parties tend

M N

A AA BB B C

Figure 13.1 Hotteling’s model (1929)

 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

254 DANIELE CARAMANI

to converge towards one another in the effort to win the Party under Tony Blair, or the US Democrats under Bill
middle-of-the-road voters, and to present increasingly Clinton).
similar programmes and policies. Downs adds one cru- The second dynamic element consists of centripetal
cial element to the models: the variable distribution of competition arising not only because of the proxim-
voters along the left–right continuum. Voters are not ity principle, but also because there are more voters in
distributed regularly along the scale but concentrate in the centre. Parties’ strategy does not only depend on the
particular ideological positions, namely around the cen- logic of the model (the assumption of proximity voting)
tre. For Downs this is the crucial explanatory and pre- but also on the empirical distribution of the electorate.
dictive element of party systems’ dynamics: ‘if we know The potential loss of voters at the extremes does not de-
something about the distribution of voters’ preferences, ter parties from converging because there are few voters
we can make specific predictions about how ideologies at the extremes. This is not the case if the distribution
change in content as parties maneuver to gain power’ of the electorate is different, a two-model distribution as
(1957: 114). If one assumes a normal (or ‘bell-shaped’) depicted in type B. This is a case of ideological polariza-
distribution of the electorate with many voters at the tion within a political system (for example, the Weimar
centre of the scale and fewer at the extremes (see type A Republic and Italy during the ‘first republic’). The distri-
in Figure 13.2), the prediction of the model is again that bution of the electorate therefore determines the direc-
parties will converge toward the centre.7 tion of competition (centrifugal or centripetal).
The first dynamic element of these models is that The third element of the dynamics of party systems is
they predict the convergence toward the centre and that in the middle of the left–right axis voters are more
the increasing similarity of platforms and policy ac- flexible than at the extremes where they are firmly encap-
tions. This centripetal competition is determined by sulated in strict ideologies and/or party organizations.
the parties’ aim to win the median voter (see Box 13.4). ‘Available’ voters (Bartolini and Mair 1990), located in
Examples are the progressive convergence of previously the middle, are less ideologized and have weak party
radical leftwing workers’ parties toward the centre to at- identifications. These voters are ready to change their
tract moderate voters (the German Social Democrats in minds and, therefore, are very appealing to parties seek-
1959, the French Socialists in the 1970s, the New Labour ing to ‘seduce’ them.

Type A: Downs’s basic model (1957): the bell-shape (or normal) distribution of the
electorate: centripetal competition

Number of
electors

0 25 50 75 100
A B

Type B: A two-modal distribution of electors: centrifugal competition

Number of
electors

0 25 50 75 100
A B
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 255

Type C: A skewed distribution of electors: enfranchisement in the nineteenth century

and new parties

Number of
electors

0 25 50 75 100

C D A B

Type D: Polymodal distribution in multi-party systems

Number of
electors

0 25 50 75 100
A B C D

Figure 13.2 Types of voter distributions

Dealignment
Box 13.4 The median voter
These models also help to interpret patterns of dealign-
The median voter is the voter who divides a distribu-
ment, that is, the loosening of the relationship between
tion of voters placed on a left–right scale into two equal
parties and specific segments of the society (workers for
halves. In a distribution from zero to 100 in which for
social democrats, peasants for agrarians, middle classes
each point there is a voter (including position zero), the
for conservatives). Centripetal competition and the max-
median voter is on position 50 (with fifty voters on each
imization of votes lead parties to make their programmes
side). Suppose, however, that there are fifty voters on
and ideologies more vague to attract support from other
position 100, and the remaining voters distribute regu-
groups. This blurs the connection between groups and
larly between positions 49 and 99 (one voter on each
parties and causes a higher propensity to change vote
position). In this case the median voter is on position 99.
from one election to the next.

The wider application of rational Enfranchisement and democratization

choice models In both types A and B the distributions are symmetri-
cal. In type C, on the contrary, we have a skewed distri-
What are the links of these models with other aspects of bution. The solid curve represents an electorate that is
parties and party systems? Four are particularly impor- skewed toward the right of the axis. Here the median
tant as they show the range of their potential application. voter is around position 65 rather than 50, and parties
A and B would accordingly converge toward this point.
Party organization This is a situation typical of restricted electorates in the
Rational choice models help to interpret the transforma- nineteenth century when lower classes were excluded
tion of parties from mass parties to catch-all parties (see from the franchise. A and B would therefore be the typi-
Chapter 12). This transformation can be seen as the or- cal parties of the periods of restricted electorates, namely
ganizational and ideological adaptation to competition. Liberals and Conservatives as the parties of ‘internal
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

256 DANIELE CARAMANI

origin’ (Duverger 1954) of the bourgeoisie and aristoc- PR and multiparty systems
racy. Enfranchisement and democratization processes in
the second half of the nineteenth century and first two Under FPTP convergence is likely because the threat of
decades of the twentieth radically changed the shape of other parties appearing at the extremes is low, given the
the distribution of electors as represented by the dashed high threshold required to win a seat. Rather than new
curve, making it more similar to a normal curve. This parties, under these systems, the model predicts high ab-
new distribution explains the emergence of new parties stention levels as is the case in the US. Is it different in PR
C and/or D of ‘external origin’ (such as social democrats electoral systems?
and agrarians).8 First, multiparty systems occur when PR electoral sys-
The dream of reformists (as against revolutionary tems allow a lower threshold of representation. Second,
socialists) was that socialism and the proletariat could multiparty systems develop when the distribution of
come to power through votes (‘paper stones’) and the the electorate is polymodal, with more than one or two
extension of the franchise rather than through revolu- peaks (type D). With electors’ distributions of this type
tion (real stones!). For analytical Marxists the develop- the dynamics of the competition is not centripetal. PR
ment of the industrial society would naturally lead work- is no hindrance to new parties. Existing parties have no
ers to power through sheer numbers. Since, however, incentive to converge toward the centre since they would
numbers of industrial workers did not grow—in fact, lose part of their support to ‘neighbouring’ parties and
they declined—socialist parties faced a dilemma between because the position they would be moving to is already
moving toward the centre to maximize their appeal to busy. It is therefore less likely than in two-party systems
the middle classes—thus relaxing their programme—and that parties will look like each other ideologically.
giving up their ideology and losing voters from workers Downs assumed that the ideological space was
(Przeworski and Sprague 1986). fixed and not elastic, that is, ranged from zero to 100.

Box 13.5 Critiques to rational choice models

Assumption Critique
Rationality The relationship between parties and voters is determined by a number of ‘non-ra-
tional’ or ‘irrational’ factors: socio-economic conditions, party identification, political
socialization, influence of the media. Empirical research shows that most voters vote
according to these factors rather than rational ones (Budge et al. 1976).
Full information Voters are not fully informed about the proposal parties present in their platforms and
programmes, and are unable to evaluate the extent to which they correspond to their
own preferences. Also, do voters know what their preferences are? With technical is-
sues this often proves unlikely.
Vote maximization Parties are not vote maximizers:
● Parties as office seekers do not require to maximize votes but to get just
enough. There is the need to win but the magnitude of the victory is not im-
portant. Parties ‘seek to maximize only up to the point of subjective certainty
of winning’ (Riker 1962: 33).
● Parties may just seek to influence public policy rather than aim for office. As de
Swann notes, ‘[c]onsiderations of policy are foremost in the minds of the ac-
tors … [T]he parliamentary game is, in fact, about the determination of major
government policy’ (1973: 88).
●Vote maximization faces resistances within parties to keep a less ‘cynical’ at-
titude and a more coherent ideology. The influence of militants, activists, and
‘rank and files’ should not be underestimated.
Müller and Strøm (2000) find that only half of the parties they examine follow a strat-
egy of vote maximization.
One-dimensionality It is not realistic that all parties compete along the left–right dimension. This may be
true in two-party systems in which the electoral system reduces the variety of parties.
In multiparty systems, however, the number of dimensions is larger. Several empirical
analyses show that the space of competition is in most cases pluri-dimensional, as
genetic models show.
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

CHAPTER 13 PARTY SYSTEMS 257

Observing multiparty systems, Sartori (1976) was on less available. Even if these models apply to parts of the
the contrary able to establish that ideological spaces are electorate only, they are crucial as they determine the di-
elastic, that is, they can stretch, with extremes becoming rection of competition.9
more extreme and an increasing ideological distance be- A more fundamental question, however, is how to
tween parties. Parties may adopt the strategy of becom- interpret the convergence of parties and the increasing
ing more extreme to distinguish themselves from mod- similarity of their programmes. It is difficult to separate
erate parties. This leads to the radicalization of parties the impact of competition from other factors such as (1)
wishing to maintain a distinctive character. the development of a large and homogeneous middle
In conclusion, the crucial determinant is the distribution class and the disappearance of other classes, namely the
of the electorate. If we know the shape of the curve we can working class; (2) the reduction of social inequalities and
predict the behaviour of parties. However, to know what the secularization of society with the disappearance of
the voters’ distribution looks like is a matter of empirical religious conflicts; (3) the integration of societies and
research namely through surveys asking respondents to the disappearance of ethnic and linguistic particularities
place themselves on a left–right scale (Laver 2001). through nationalization and globalization. Is the conver-
In spite of critiques (see Box 13.5) these models remain gence of parties a result of this evolution rather than a
useful. In all electorates a number of less ideological vot- product of competition?
ers are ready to change their vote. This is an available
electorate around which competition turns and on which
these models focus. This electorate is composed of opin- Conclusion
ion voters or ‘pocket-book voting’, that is, based on pri-
Understanding party systems requires the combination
vate interests, values, and opinions, rather than identity
of the various perspectives presented in this chapter.
voters or ‘socio-tropic voting’ based on socio-economic
and identification factors. The models described in this 1. The macro-sociological approach must be combined
section apply less to segments that are encapsulated in with institutional and actor-oriented models. They
strong identifications. For this reason these models ap- complement each other and are not mutually exclu-
ply to the left–right dimension along which voters are sive. We cannot understand party systems without
available rather than other dimensions (ethnic, linguistic, reference to the social cleavages from which parties
religious) along which identities are stronger and voters have emerged. However, we must also take into ac-
count parties’ capacity to act independently from
KEY POINTS social conditions—in fact, to shape them—through
ideology and policy. The motivations of parties are
● In the electoral market parties (the supply side) present
not entirely determined by their origins. Parties’
programmes and platforms to appeal to a large num-
strategies, in turn, must take into account the rules
ber of voters whose vote is determined by the proxim-
of the game—electoral laws being the most important
ity of their preferences (the demand side) with the par-
ones—influencing the number and size of players.
ties’ offer. Voters are assumed to be rational, that is,
2. Both descriptive and explanatory research are needed.
informed about alternative party proposals and able to
The ultimate goal of research is to account for the
chose the alternative closest to their top preferences.
shape and dynamic of party systems. However, be-
● The dynamics of party systems is determined by par-
fore searching for causes, party systems should be de-
ties’ search for the optimal location on the left–right
scribed carefully. As seen with counting parties, this
axis. Depending on the distribution of the electorate
is often more complicated than appears at first sight.
along the scale, parties move to a position where the
3. Finally, we cannot understand party systems in isola-
support in votes is largest.
tion. We need comparison to assess whether or not
● The prediction of competition models is that parties they are fragmented or unstable, as well as a long-term
converge toward the centre of the left–right axis as perspective rather than a myopic focus on just the
the optimal location, as the point where most votes most recent elections. This is the only way of assess-
concentrate, and as the point where voters are less
ing how exceptional a given party system or a given
rigidly ideologized.
change really is.

uestions

1. What are the National and the Industrial Revolutions? 3. How should the number of parties in a system be counted?
2. What are Stein Rokkan’s four main social cleavages and 4. What are the characteristic features of a two-party
which party families emerged from them? system?
 OUP UNCORRECTED PROOF - FINAL-PROOF, 01/24/11 LASERWORDS

258 DANIELE CARAMANI

5. What does ‘effective number of parties’ mean? 8. Describe centripetal and centrifugal party competition in
6. What is the effect of electoral systems on the shape of Downs’s model.
party systems? 9. Are voters really rational?
7. What does it mean that parties are ‘vote-maximizers’? 10. Can the space of competition be reduced to one left–
right dimension?

Further reading

Classical texts on party systems For a more extended bibliography see the works cited
throughout in this chapter.
Downs, Anthony (1957) An Economic Theory of Democracy
In addition, up-to-date reports on party systems can be
(New York: Harper Collins).
found in journals. Detailed country-by-country developments
Duverger, Maurice (1954) Political Parties (New York: Wiley).
from 1991 onwards are reported in the Political Data
Lipset, Seymour M., and Rokkan, Stein (1967) ‘Cleavage Yearbook, an annual supplement of the European Journal of
Structures, Party Systems, and Voter Alignments: An Political Research. Students may also find useful material in
Introduction’, in Lipset and Rokkan (eds), Party Systems and journals such as the American Political Science Review, British
Voter Alignments (New York: Free Press), 1–64. Journal of Political Science, Comparative Politics, Comparative
Sartori, Giovanni (1976) Parties and Party Systems: A Framework Political Studies Electoral Studies, Party Politics, and West
for Analysis (Cambridge: Cambridge University Press). European Politics.

Web links

www.electiondataarchive.org collection which includes also information on parties and

The largest ever produced databank on elections at the level party systems.
of single constituencies since the beginning of competitive
www.georgetown.edu/pdba
elections covering the entire world.
Website of the Political Database of the Americas including
www.electionresources.org information on parties and party systems.
Manuel Álvarez-Rivera’s Election Resources in the Internet.
http://dodgson.ucsd.edu/lij
www.psr.keele.ac.uk Website of the Lijphart Election Archive with information on
Richard Kimber’s website on Political Science Resources party systems, electoral systems, and recent election results
(University of Keele). around the world.

www.parties-and-elections.de www.cia.gov/cia/publications/factbook
Database of Parties and Elections about parliamentary Website of CIA’s The World Factbook with information on
elections, parties, and political leaders in Europe. institutions, social structures, economic data, and party
systems for most countries of the world.
www.electionworld.org
Website includes information on political parties around the www.idea.int
world with up-to-date election results and other information Website of the International Institute for Democracy and
on the party system and the main institutions. Electoral Assistance (IDEA).

www.gksoft.com/govt/en/parties.html www2.essex.ac.uk/elect/database/aboutProject.asp
Webpage of Government on the WWW devoted to Website of the project on Political Transformation and the
political parties and party systems around the world. The Electoral Process in Post-Communist Europe (University of
main page includes additional information on heads of Essex).
state, parliaments, executives, courts, and other
www.eiu.com
institutions.
Country Reports and Country Profiles published by the
www.epicproject.org Economist Intelligence Unit are very useful for an overview
Website of the EPIC Project on election process information and recent data.

For additional material and resources, please visit the Online Resource Centre at:
www.oxfordtextbooks.co.uk/orc/caramani2e