The Limits of Duverger’s Law
New book sets standard for research on electoral systems
Posted: November 14, 2012
Here’s a law if ever there were one: every scientist wants a law named after them. Political science cannot point to many such laws, but the discipline has typically assumed it has at least one — Duverger’s Law. Robert Moser, in his new book coauthored with Ethan Scheiner (University of California, Davis), rains on this parade, although it is clear that their modifications to existing thought push the discipline’s scientific rigor forward.
Maurice Duverger wrote his famous work on political parties in the mid-20th Century. Duverger’s Law asserts that a First-Past-the-Post electoral system — an election in which there is one seat to be won and the winner is the candidate who gains the most votes — tends to produce two-party competition at the district level. “Electoral Systems and Political Context: How the Effects of Rules Vary Across New and Established Democracies” demonstrates that the applicability of Duverger’s Law is actually quite limited. Rather than being a covering law, Duverger’s Law applies more narrowly to relatively homogenous electoral districts within established democracies with well-developed political parties, a context which fails to describe new democracies and socially diverse districts in any democracy.
A major reason why Duverger’s Law is limited is because the law rests on voters and elites having significant information about likely outcomes and/or political parties that clearly structure the vote. Given a context where it is unclear which candidates are truly competitive, Duverger’s Law will not work. Because of this, Duverger’s Law cannot reliably be applied to new democracies, especially new democracies with poorly established political parties, and in these cases a First-Past-the-Post system might actually encourage candidate proliferation, because information is lacking and it may only take relatively few votes to win an election.
Moser and Scheiner use mixed-member electoral systems, which give voters two votes — one for a candidate in a single-member district and one for a party under proportional representation (PR) — as a social laboratory to study electoral system effects. Based on comparisons of outcomes under the two different rules used in mixed-member systems, the book highlights how electoral systems’ effects — especially strategic voting, the number of parties, and women’s representation — tend to be different in new democracies compared to what one usually sees in established democracies. By showing that the effects of electoral institutions vary depending on a country’s democratic experience, the relative institutionalization of the party system, and a district’s level of social diversity, Moser and Scheiner demonstrate that specific differences in political context systematically condition the effects of electoral rules. This argument is contrary to the dominant paradigm in political science, which is that rules condition context. However, the argument also runs counter to opponents of the dominant paradigm who argue that generalization is impossible because context is determinant.
Although Moser and Scheiner demonstrate that context mitigates the way electoral institutions affect outcomes, they likewise demonstrate that these differential effects are systematic and predictable, or put another way, that context is a variable and its effects can be generalized. The authors apply this framework to a range of electoral system effects. For example, scholars have long argued that PR systems tend to promote the election of women. However, Moser and Scheiner show that the effect of electoral rules on women’s representation depends on the level of party fragmentation and the degree to which the society supports women’s representation. Where there are few parties and relatively high support for female leaders, women win office more frequently under PR systems. However, where there are many parties and lower support for female leaders, PR systems are not likely to significantly increase women’s representation over that found under First-Past-the-Post systems.