Estimating the Effects of Electoral Thresholds Using Regression Discontinuity

• In Event: Effects of Democratic Representation on Public Opinion

Abstract

Electoral systems are characterized by a multiplicity of strict, arbitrary cutoffs. Regression discontinuity design (RDD) is a method that exploits such cutoffs, treating them as if they were a mechanism for random assignment. Therefore, under certain assumptions, if a discontinuity in the estimated parameter is observed at the cutoff value, then this discontinuity can be used to estimate a local average treatment effect. Though one of these cutoffs--the cutoff that determines the plurality winner in first-past-the-post elections--has been thoroughly explored using RDD, this study highlights the general utility of RDD for research in electoral systems. As an illustrative example, using data from the Comparative Study of Electoral Systems project, we exploit the threshold for representation (common in many proportional representation regimes) as an assignment mechanism, and attempt to identify a causal effect of representation on respondents’ attitudes toward the state of the economic and political system in their country.

Authors

• Reuben Kline, SUNY, Stony Brook University

• David Stack, Cleveland State University