Social cleavages, district magnitude, and the number of parties

(Both co-authored plan to be present.) (Tentatively for panel 2, Electoral Systems and Parliamentary Party Strategies, chair: Thomas Zittel) *** There is a close connection between the electoral system (especially of district magnitude) and the format of the party system (Rae, 1967; Taagepera & Shugart, 1989). The literature has also shown that the effect is moderated by social cleavages (Taagepera & Grofman, 1985; Amorim Neto & Cox, 1997; Ordeshook & Shvetsova, 1994; Clark & Golder, 2006; Li & Shugart, 2016; Mozaffar, Scarritt, & Galaich, 2003). However, none of these approaches has investigated the difference between the territorial dimension of cleavages. Both are supposed to work in affect party systems in opposed directions. Studying national parliamentary elections, we amend the models of electoral system and cleavage effects, distinguishing between functional and territorial cleavages. The social groups underlying functional cleavages live widespread in a country. Henceforth, functional cleavages are not only present at the national level of politics, but also at the subnational level, even in small electoral districts. Accordingly, the size of the electoral districts will limit the translation of multiple functional cleavages into party competition. Territorial cleavages, however, are related to groups, which live concentrated in a delimited area of the country. While such groups might be relatively small at a national scale, they will be quite large in those electoral districts, where they reside. Accordingly, small electoral districts will not block the representation of territorially based groups into national parties. In brief, while district magnitude will affect the party system format in the presence of functional cleavages, it will not do so if cleavages are primarily territorial. This paper provides for a formal model on the effect of cleavages and district magnitude on the format of party systems. And it provides for an empirical test, based on a global dataset of electoral systems and party systems in the first chamber of parliament of 70 democracies. To measure the effect of different types cleavages, we rely on data on ethnic diversity, and identify ethnic parties. Where our modeling differs from all previous work is that we treat the existence of ethnic parties as a crucial intermediating variable between ethnic diversity and party proliferation.