Explaining Variation in the Competitiveness of U.S. Senate Elections, 1922-2004
44 Pages Posted: 5 Feb 2013 Last revised: 24 Feb 2013
Date Written: January 5, 2013
Abstract
Understanding the nature of political competition is a central issue in political economy. This paper offers an explanation for observed variation in the competitiveness of U.S. Senate elections since direct elections to the Senate became fully effective in 1922. We deliberately abstract away from candidate-specific factors to look at more general features of the electoral landscape. In the framework we develop, national party-specific constraints on the ideological positioning of local Senate candidates interact with the heterogeneity of state electorates to determine the number of highly competitive electoral contests. The model predicts that the number of local contests that are highly competitive will be greater: (1) if a party's national caucus is ideologically more heterogeneous, signaling a decline in the ability of the national party to impose restrictions on local candidates and/or greater tolerance on the part of contributors of deviation from the national party image; (2) if voters in a state are more ideologically heterogeneous; and (3) if only registered party members may vote in a (closed) state primary. To test these hypotheses, we use a spline estimation procedure to predict the effect of key factors identified by the theoretical framework on the size of a highly competitive segment or set of elections, as opposed to segments where one or the other party dominates. The results provide some support for the hypothesis concerning primary elections, and strong support for the two predictions concerning the heterogeneity of the national party caucus and of local electorates.
Keywords: electoral competition, political competiveness, U.S. Senate electons, ideological heterogeneity, national party constaints, primary elections, spline regression, asymmetric breakpoint
JEL Classification: D72
Suggested Citation: Suggested Citation