Political parties need to differ – within reasonable limits

Riccardo Pelizzo, Griffith University


Democracies are fragile. No matter how long a regime has been democratic, it is always confronted with the possibility of becoming dysfunctional, of becoming less democratic, and possibly breaking down. Some scholars believe that polarisation is the single most important cause of the failure of democracy, because if there is too much disagreement the political system cannot possibly work. Others argue that without polarisation, parties fail to provide the voters with viable alternatives and, by doing so, undermine the quality of democracy.

Democracy is ‘unthinkable save in terms of parties’ (Rosenblum 2008) when parties adequately perform their representative function. But democracy runs into troubles when parties provide either excessive or negligible policy alternatives. On one hand, when the main parties are under-representative, they can create the conditions for the emergence and/or the success of anti-system parties such as the National Front in France, The Freedom Party in Austria, the Northern League in Italy or One Nation in Australia. On the other hand, when parties over-emphasise political differences, they may endanger the democratic regime by turning disagreement on policy into disagreement over fundamentals. This type of disagreement, or polarisation, which threatens the concept of a loyal opposition, is the most problematic development in US politics. There voters of both parties are convinced that their political opponents are morally inadequate and politically unfit to govern.


Political scientists working on political parties, party systems, and voting behaviour often talk of polarisation. Regardless of whether the concept of polarisation is applied to the study of parties, party systems, or electorates, it is very clear: there is polarisation ‘when we have ideological distance’ (Sartori 1976). The concept of polarisation can be used to describe both the amount and intensity of disagreement.

If there is too much disagreement, the political system cannot possibly work.

When there is a widespread agreement on the importance of a specific issue, the electorate is not polarised. For example, for most of the time between 1970 and 1990, a sizeable majority of German and Italian voters felt that fighting unemployment was the single most important issue for the government, and that the government should do so by increasing the amount of spending on labour market programs and industry assistance (Pelizzo 2008). Similarly, more than 60 per cent of the Australian voters in surveys following the 1993, 1996 and 1998 elections said that fighting unemployment was either important or very important. By contrast, when there is considerable disagreement on a specific issue, the electorate is polarised. In the United States, for example, the electorate is polarised on issues such as abortions, gay rights, or the legalisation of drugs (Flanigan & Zingale 1998). A large majority of liberals is in favour of each of these issues, and an overwhelming majority of conservatives strongly opposes them.

The concept of polarisation is also applied to the study of party systems. A party system is polarised when there is a large ideological distance between the parties. As Rosenblum (2008) recently noted, voters at election time are surveyed in most advanced industrial societies. In the course of these electoral surveys, they are asked, among other things, to locate themselves and political parties on the political spectrum. In Australia and in most Western European nations, voters are asked to locate themselves and parties on a left-right continuum, while in the United States voters are asked to locate themselves and parties on a liberal-conservative scale.

The reason why election surveys ask these kinds of questions is straightforward. In 1957, Anthony Downs proposed a simple model to predict the result of the elections. According to Downs, voters and parties have preferences and these preferences can be represented by a point in the political space. Downs’ idea is very clear: I am much more likely to vote for a party which stands for what I believe in than for a party with which I have many and/or profound disagreements. Since the position of the median voter is, by definition, majority-preferred, a party in a two-party system has an incentive to converge toward the position of the median voter to appeal to increase its chances to win the elections.

Electoral surveys can also be used to measure the polarisation of the party system. For example, the 1996 American National Election survey asked voters to locate the presidential candidates of both the Democratic and the Republican parties on a seven point scale, ranging from very liberal (1) to very conservative (7). If we take the positions of the presidential candidates as proxies for the positions of their respective parties, we find that the Democratic Party received a score of 3.15 and the Republic party received a score of 5.15—which means that the distance for the 1996 elections was 2. In 2004, the Democratic candidate received a score of 2.99, while the Republican candidate received a score of 5.18—in this case, the distance was of 2.19. We can, then, say that the American party system was more polarised in 2004 than it had been in 1996.

Political scientists have identified several reasons why party systems polarise.

Western democracies usually have more than two relevant parties. These parties are relevant because they are instrumental in the formation of a government, because they can prevent the formation of a government, or because they can influence the direction of political competition (Sartori 1976). A party influences the direction of competition when its sheer existence forces all the other parties to modify their policy stances, their positions, and their strategy. Between 1980 and 2000, first, and unified Germany later, had four relevant parties: the Social-Democratic Party, the Christian Democratic Union, the Freedom Party that joined any CDU-led government, and the Greens that joined any SDP-led government. In 1980, 1983, and 1987 the right-most party was the CDU (for the sake of simplicity I treat the Christian Social Party as part of the CDU) and the left-most party was the Green party. The distance between these parties and, hence, the polarisation of the party system increased from 4.11 in 1980, to 5.07 in 1983 and to 5.80 in 1987. This distance was not at all negligible if one keeps in mind that it was expressed on an 11-point scale.


What factors are responsible for changes in the levels of polarisation in party systems? Political scientists have identified several reasons why party systems polarise. Duverger (1954) and Lipset and Rokkan (1967) argued that societies are divided by conflict lines or social cleavages, and that some of these cleavages are politically relevant as they can be used to build voter identities (Campbell, Converse, Miller & Stoke 1960), loyalties (Rose & MacAllister 1990), or alignments (Lipset & Rokkan 1967). For other scholars (Sartori 1976; Baldassarri & Gelman 2008), the polarisation of a party system is a consequence of the number and depth of these social cleavages. While this line of research posits a relationship between polarisation and structural conditions—a party system is polarised because society is deeply divided—other studies have shown instead that changes in the level of polarisation can also be due to contextual factors. In work with a colleague, I have argued that the worsening of economic conditions, especially the rise of hyper-unemployment, is related to increases in the level of polarisation in pluralist party systems—party systems that are characterised not only by major ideological differences between the parties but also by the presence of more than five relevant parties (Pelizzo & Babones 2007). Party systems become more polarised in periods of economic crisis because voters are more receptive of radical messages than they would be otherwise. Hence, party leaders and political entrepreneurs have an incentive to radicalise their policy stances and discourse to appeal to voters and/or to prevent the emergence of even more radical forces.


Is polarisation of the party system good or bad? Party politics experts have argued that in the course of the past century or so, parties as organisations have undergone a major transformation. With the extension of universal suffrage, the cadre or elite parties of the pre-democratic era were supplanted by mass parties of social integration. These parties represented the interests of specific social groups, for example Labour, Socialist, and Social Democratic parties represented the interests of working class. In the mass party age, which lasted from about the 1880s to the late 1950s, party competition was very ideological because parties’ electoral competitiveness depended on their ability to mobilise the members of the groups they represented. From the 1960s onward, parties and electoral strategies changed. Parties, in order to remain electorally competitive, had to appeal to voters belonging to different social groups and they could not do so by sticking to rigid ideological principles. There was less polarisation from the 1960s onward quite simply because there were few and often negligible ideological and policy differences between parties. In the last 30 years or so, some political scientists have argued, policy differences between parties have also disappeared (Katz & Mair 1995). On this view, parties adopt roughly the same platforms and advocate roughly the same solutions. By doing so, they fail to adequately represent voter preferences and behave like a cartel of oligopolistic firms. If the main parties’ are not able to represent voters’ preferences adequately, more radical parties can emerge to undermine government stability and performance. Hence, scholars proposing the ‘party cartelisation’ theory believe that clear differences, as indicated by policy distance or polarisation, are beneficial for the proper functioning of a democratic system.

Issues with political impact evoke emotions in the voters.

In a recent article, Russell Dalton (2008) continued the economic metaphor, suggesting that polarisation should also be viewed as a proxy for product differentiation in the political arena. His empirical analysis shows that countries in which the party system is more polarised have higher levels of voter participation in elections. If the quality of democracy is a function of electoral participation (Lijphart 1999), then polarisation is good for democracy.


Other scholars have instead pointed out that party system polarisation is detrimental to the functioning of democracy for two different reasons. First, the polarisation of the party system reflects the polarisation of society. The literature documented that ideological polarisation is responsible for less durable governments (Taylor & Herman 1971), and that ideological polarisation leads not only to government instability, but also to government ineffectiveness, loss of legitimacy and democratic breakdown (Sartori 1976). More recently George Tsebelis (2002) has suggested that polarisation remains a problem for the quality of democracy, if not its survival. He shows that ideological polarisation is responsible for a variety of suboptimal outcomes, such as lower levels of government stability, budgetary problems, and lower quality of legislation. So even if the polarisation of the party system reflects the spread of opinion in society, so that in some respects parties are adequately performing their representative function, polarisation remains a problem because it makes governments work badly.

Proximity models of electoral choice, like that proposed by Downs, have also come under some criticism from directionality theorists. The most important problem, these critics claim, is that ‘the vast majority of voters (do) not see issues in the sharp positional fashion that the traditional theory assumes’. On the contrary, directionality theorists argue, ‘issues are perceived rather diffusely’ (Rabinowitz & Macdonald 1989). On this view, for issues to have a political impact they must be able to evoke emotions in the voters. To understand the impact of a political issue, then, we need to know whether a voter feels favourable or not toward that issue (that is we need to know the voter’s direction) and we also need to know the magnitude of her feeling toward the issue (that is we need to know the intensity of the voter’s feelings about an issue). Thus, for directionality theorists, voters’ assessment of a given party does not reflect how close that party is to the voter’s position. Instead, voters’ assessment of a given party reflects (1) whether the voter and the party are on the same side of a given issue (direction) and (2) how important that issue is for both the voter and the party (intensity). The combination of direction and intensity generates what directional theorists call the ‘directional effect’.

Figure 1

For example, consider Figure 1. According to proximity theory the voter V would prefer party B over party A because the distance between V and B is only two units while the distance between A and V is of four units. The opposite is true for directional theory. According to directionality theory the voter V will prefer A over B because A and V are on the same side of the issue.

Figure 2

If there were a third party C, as shown in Figure 2, located between the voter V and party A, the voter would prefer party A over party C because party A feels more strongly and hence holds a more radical position about this issue. Why would a voter support a more radical position? Because voters know that when parties win the elections, they need to moderate their stances and compromise. Hence, our voter knows that if she wants taxes to be reduced by 10 per cent, she cannot vote for the party that advocates a 4 per cent tax reduction because after the government party has compromised with the opposition, the tax reduction will be below 4 per cent and quite far from the 10 per cent that she had hoped for. Hence, if a voter wants the tax cut to be close to the 5 per cent that she wants, she has to vote for parties that advocate more radical tax cuts.

Polarisation is detrimental under some conditions but not others.

But since parties know that voters will support a party advocating more radical solutions, parties in their turn have an incentive to radicalise their policy positions. The implication of this theory is that a party system may become polarised in spite of the relative homogeneity of voters’ preferences. When parties adopt this strategy of outbidding each other, they engage in irresponsible behaviour. For the sake of outbidding each other, they end up advocating measures that are not very representative of voter preferences and that are quite possibly detrimental to the national interest. Worse, this policy of outbidding can induce voters to believe that their political opponents are positively antithetical to the good of the country and/or that a new political system is needed. In other words, one of the problems associated with party induced polarisation is that it may turn disagreement on issues into a disagreement on fundamentals such as the existence of the political system.


Why can political scientists argue, simultaneously, that polarisation is a positive and a negative feature of democratic politics? The answer is ‘conditionality’. Polarisation is detrimental under some conditions but not others. The literature identifies two conditions under which polarisation is detrimental. The first condition is the number of parties in the party system (fragmentation). Sartori found that when a party system has more than five relevant parties and is polarised, it is difficult to form stable government majorities and make democratic governments work—which is why in all those cases in which the party system displayed high levels of fragmentation and ideological polarisation, the constitutional order broke down (Pelizzo & Babones 2007).

Polarisation is also detrimental when there is disagreement between opposing parties over the nature of the political system itself. In polities like the United States, there is disagreement and polarisation, but the focus is policy issues—abortion, gay rights, and so on. In other countries, however, the disagreement between parties does not concern specific issues or policies, but concerns instead the existence of the political system. In the Weimar Republic between the two World Wars, some parties wanted to preserve the status quo, the Communists wanted to topple the democratic regime and establish Communist rule, while the Nazis wanted to establish–and succeeded in establishing—a new political regime that was neither Communist nor democratic. In countries where ideological disagreement concerns the nature of the political system, high levels of polarisation are bad because they indicate that the democratic regime enjoys very little support, has little legitimacy, and is likely to be overthrown.


The key lesson here for parties is that they need to be representative of voters’ preferences and responsive to voters’ demands. Whenever parties do not adequately perform their representative function, they undermine the functioning of the political system and the quality of democracy. In fact, when parties represent only a portion of social interests, they create the conditions for the emergence of new, radical, possibly anti-system parties that may not be beneficial for the quality or the survival of democracy. Alternatively, when parties adopt a strategy of outbidding each other to gain electoral support, they polarise the party system and may turn legitimate policy differences into a disagreement as to what is good for the country, as to what is the best political system, and as to whether the political opponents are fit to govern—and when this occurs, the quality of democracy is seriously compromised.

Australian parties have historically been able to find the proper balance. Unlike their European counterparts, Australian parties have been able to reaffirm inter-party differences without resorting to a strategy of outbidding. The question is whether in the future they will be able to preserve differences and disagreement and to keep them within reasonable limits.


Baldassarri, D. & Gelman, A. 2008, ‘Partisans without constraints’, American Journal of Sociology, vol. 114, no. 2, pp. 408–446.

Campbell, A., Converse, P.E., Miller, W.E, & Stokes, D.E. 1960, The American Voter, University of Chicago Press, Chicago.

Dalton, R. 2008, ‘The quantity and the quality of party systems: Party system polarization, its measurement, and its consequences’, Comparative Political Studies, vol. 41, no. 7, pp. 899–920.

Downs, A. 1957,  An Economic Theory of Democracy, Harper and Row, New York.

Duverger, M. 1954, Political Parties, Methuen, London.

Flanigan, W.E. & Zingale, N.H. 1998, Political Behavior of the American Electorate, CQ Press, Washington DC.

Katz, R.S. & Mair, P. 1995, ‘Changing models of party organization and party democracy: The emergence of the Cartel Party’, Party Politics, vol. 1, no. 1, pp. 5–28.

Lijphart, A. 1999, Patterns of Democracy, Yale University Press, New Haven.

Lipset, M.S. 1959, ‘Some social requisites of democracy’, American Political Science Review, vol. 53, no. 1, pp. 69–105.

Lipset, M.S. & Rokkan, S. 1967, ‘Cleavage structure, party systems and voter alignment, in Party Systems and Voter Alignment, eds M.S. Lipset & S. Rokkan, Free Press, New York, pp. 1–64.

Pelizzo, R. 2008, Cartel Parties and Cartel Party Systems, VDM Verlag Dr Mueller, Saarbruecken.

Pelizzo, R. & Babones, S.J. 2007, ‘The political economy of polarized pluralism’, Party Politics, vol. 13, no. 1, pp. 53–67.

Rabinowitz, G. & MacDonald, S.E. 1989, ‘A Directional Theory of Issue Voting’, American Political Science Review, vol. 83, no. 1, pp. 93–121.

Rosenblum, N.L. 2008, On the Side of Angels, Princeton University Press, Princeton.

Rose, R. & McAllister, I. 1990, The Loyalties of Voters, Sage, London.

Sartori, G. 1976, Parties and Party Systems, Cambridge University Press, New York.

Taylor, M. & Herman, V. 1971, ‘Party systems and government stability’, American Political Science Review, vol. 65, no. 1, pp. 28–37.

Tsebelis, G. 2002, Veto Players, Princeton University Press, Princeton.

Riccardo Pelizzo is a research fellow at the Centre for Governance and Public Policy of Griffith University. His articles have appeared in several international journals such as Comparative European Politics, Party Politics, and West European Politics.