LAUSR

Abstract Income inequality has been rising throughout the industrialized world, particularly in the United States. This long been thought to depress turnout, but extant research has yielded mixed findings. Here, I argue that the inequality-turnout relationship is conditional, depending crucially on election salience. I test this by using three decades (1984–2014) of panel data from the U.S. states and by leveraging the fixed and exogenous occurrence of presidential (higher-salience) and midterm (lower-salience) elections. Overall, I find a negative and statistically significant relationship between income inequality and voter turnout in midterm election years, but a substantively small and non-significant relationship in presidential election years. I attribute this to the ability of presidential contests, relative to midterms, to counteract the demobilizing influence of high inequality, by piquing voters’ interest and activating citizens who would otherwise abstain. Overall, these findings help us to better understand of the politics of electoral participation in an era of high, and rising economic inequality.