Distribution of votes across districts and the effective number of parties in 50 democracies (1945-2012)

• Grant number: 13/02942-3
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Start date: June 05, 2013
• End date: June 04, 2014
• Field of knowledge: Humanities - Political Science
• Principal Investigator: Fernando de Magalhães Papaterra Limongi
• Grantee: Fabricio Jorge Vasselai
• Supervisor: Kenneth W. Kollman
• Host Institution: Faculdade de Filosofia, Letras e Ciências Humanas (FFLCH). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: University of Michigan, United States

Abstract

This research investigates the effects of the district distribution of electoral support of parties on the national effective number of parties, as well as the determinants of such distribution. I depart from Cox's remarks that the propositions of Duverger operate in the electoral district level and, therefore, can only explain the effective number of parties within each of those districts. Hence, the correct understanding about the number of parties in the national level depends on the electoral coordination between districts. I intend to demonstrate empirically and formally that, to the extent the each partisan electoral support is heterogeneously distributed across the districts of a given country, different party subsystems arise and thus result in different national number of parties. I rely on the growing literature on electoral nationalization to find measures and proxies of the district distribution of votes and then, I propose models that include this variable as a possible determinant of the national number of parties, besides the classical variables (electoral institutions and social cleavages). As a natural consequence, I will investigate which variables impact on this homogeneity of district distribution of votes, to better understand the factors that may indirectly impact on the number of parties. This approach allows to test, for instance, whether social cleavages actually only impact on the number of parties when they are geographically biased. To overcome the main empirical difficulty needed for those kind of assessment, this research builds a dataset with electoral results for Lower Chambers of 50 countries (1945-2012) with results for all parties disaggregated at the electoral district level of each country. (AU)