How Polarized are Citizens? Measuring Ideology from the Ground-Up

Abstract

Strong evidence has been emerging that major democracies have become more politically polarised, at least according to measures based on the ideological positions of political elites. We investigate whether the general public (`citizens') followed the same pattern. To this end, we propose a novel methodology to identify the underlying ideologies of citizens by applying Latent Dirichlet Allocation (an unsupervised machine learning algorithm) to political survey data. This approach indicates that in addition to a left-right scale, confidence in institutions defines another major ideological dimension. Using this framework, we are able to decompose the shift in ideological positions across the population over time and create measures of `citizen slant' and polarisation. Specifically, we find evidence of a `disappearing centre' in a sub-group of countries with citizens shifting away from centrist ideologies into anti-establishment `anarchist' ideologies over time. This trend is especially pronounced for the US.

Keywords: Polarisation, Ideology, Populism, Unsupervised Learning

JEL Classification: D72, C81