A Test for the Rational Ignorance Hypothesis: Evidence from a Natural Experiment in Brazil

This paper tests the rational ignorance hypothesis by Downs (1957). This theory predicts that people do not acquire costly information to educate their votes. We provide new estimates for the effect of voting participation by exploring the Brazilian dual voting system- voluntary and compulsory- whose exposure is determined by citizens' date of birth. Using a fuzzy RD approach and data from a self-collected survey, we find no impact of voting on individuals' political knowledge or information consumption. Our results corroborate Downs' predictions and refute the conjecture by Lijphart (1997) that compulsory voting stimulates civic education.