Policy divergence, sorting and electoral polarization

University Of Illinois At Urbana-Champaign, Urbana IL

Investigators

Abstract

There exists a wide-spread notion that political polarization today is considerably larger than it was a generation ago. Many observers are concerned that this leads to increased gridlock in a given political system, which could threaten the functioning of government and thus, ultimately, the economic welfare of a nation and the stability of its democracy. Understanding which political and economic causes drive polarization, as well as what the effect of polarization on voters and parties is, therefore represent central questions for our society. However, precisly identifying the notion of increased polarization in scientific terms is conceptually challenging because measuring candidates' positions and comparing them over time is difficult. Since the pertinent policy issues in two different campaigns are usually very different, directly comparing the difference between the candidates' platforms is not a feasible approach to the measurement problem. Instead, the PIs provide a fundamentally new method for estimating changes in platform divergence over time, using data from opinion polls on the ideal positions of voters regarding different issues, as well as on their preferred candidate. Based on a spatial model of policy preferences, this project shows how changes in candidate positions affect voter behavior, and how changes in voter behavior provide information about candidate positions. This study will also analyze the question of "voter polarization". The existing literature is somewhat vague about what exactly constitutes polarization and how to measure it. Based on their model, the PIs provide a new measure of this concept and are able to disentangle the two effects that might be responsible for increased polarization: voter "radicalization" (i.e., a change in the distribution of voter ideal points) and "sorting" (an effect arising from more policy divergence between politicians). This method can also be used to identify the different ideological types of swing voters and their demographic composition. Moreover, one can identify which ideological and demographic voter types are underrepresented in the electorate because of abstention, and thus, the effect of changes in participation rates. A unique feature of the U.S. political system are the primaries in which voters select the candidates of their respective parties. The second part of this project will analyze a dynamic model of policy divergence and polarization, based on the interaction between candidates and policy-motivated primary voters who select them.