Political Rational Inattention: A New Measure With an Application to Political Polarization

Abstract

Attention to political information stands at the core of political theory, yet measuring it is a challenging endeavor. We propose a novel measure of political rational inattention, based on political forecast data. Capitalizing on the unique features of the American National Election Studies, which recurrently collect political forecast data at a daily frequency prior to elections, we construct a state-by-year dataset of political rational inattention, covering U.S. presidential election years from 1952 to 2020. We document substantial heterogeneity in our measure across both states and elections. We then apply this measure in the context of political polarization. First, we present a rational inattention-persuasion model that illustrates how heightened societal attention may exacerbate political polarization. Second, we empirically test the model’s predictions using our political rational inattention measure and survey data on political opinions. We find that a standard deviation increase in political attention raises average polarization by 6%.

Keywords: Political rational inattention, political expectations, polarization