Spatial Modeling for the Populist Era

• In Event: Explaining Voters' Support for Populism

Abstract

How can spatial models of vote choice contend with rise of populism? Modern political trends such as the rise of populism pose an intractable problem to existing spatial models of vote choice. Factors such as rhetorical style and candidate attributes are increasingly decisive in contemporary elections. Existing spatial models fail to successfully incorporate these non-policy considerations into the vote choice calculation, resulting in suboptimal predictive capacity. Building on the existing spatial modeling literature, I develop a revised spatial model of politics that incorporates non-policy considerations, the credibility of party offerings, and a novel treatment of issue importance.

I test this model on recent election results from Argentina and Australia through Bayesian multinomial logistic approaches. I demonstrate the comparative utility of this approach in predicting vote choice over partisan identification and traditional spatial models, and the comparable performance to powerful indicators such as feeling thermometers and approval ratings. These results suggest an improvement in our understanding of the determinants of vote choice and may yield strategies to bolster democracy against populist threats in an increasingly unstable era.

Author

• Russell Luke, Georgia State University