The Two Effects of District Magnitude: Venezuela as a Crucial Experiment*
I wish to thank James Danziger, Bernard Grofman, William Kimberling, Arend Lijphart, and Caesar Sereseres for their comments and encouragement. Special thanks to Rein Taagepera for patiently criticizing earlier drafts of this paper. Thanks also to Gonzalo Palacios of the Venezuelan embassy in Washington, DC, and David Myers for help in locating data.
ABSTRACT
Following Duverger, low district magnitude (low number of seats per electoral district) produces a mechanical effect, the underrepresentation of small parties, and a psychological effect, discouraging voters from voting for small parties. The two effects usually occur together and cannot be tested separately. However, in Venezuela an experimental ‘control’ is provided by the requirement that voters simultaneously cast a vote for the same party in each of three legislative tiers (Senate, Chamber of Deputies, and State Assembly), each with a different district magnitude. Hence the psychological effect can exert itself on only one tier. In the other tiers exceptionally high or low deviations from proportional representation are observed. The separate effects of district magnitude are analysed by their consequences on the number of electoral parties and the deviation from proportional representation.
