ABSTRACT
A great number of studies since Duverger’s seminal work (1954) have dealt with the determinants of the number of political parties. Some emphasize the importance of social structural and other non-institutional variables, but the majority of them take an institutional approach to the issue. This is due, in part, to the fact that the institutional approach is theoretically more sophisticated, in addition to its sound arguments. In other words, the institutional approach puts forward simple, clear, and testable hypotheses that link electoral systems or district magnitude to the number of parties. Yet, its theoretical hegemony does not necessarily mean that it is empirically unchallengeable. In fact, we often find a number of notable exceptions to the conventional Duvergerian and other rules. Meanwhile, non-institutional arguments often call attention to social structural cleavages, which are supposed to account for a structure of partisan divisions. Yet, they are largely narrative and descriptive. They fail to challenge the theoretical dominance of institutionalism if only because they do not present a systematic alternative explanation linking those non-institutional factors to the number of parties.
