This article re-examines Lijphart's (1997) idea that compulsory voting is an effective instrument to reduce income inequality. Using a quasi-experimental research design based on a matching technique, the results show that compulsory voting does not have a significant impact on redistributive policies. Examining the experience of Latin America, where mandatory voting prevails along with high income inequality, the article argues that the empirical problem with Lijphart's argument lies in the implicit assumption that there is a programmatic linkage between politicians and voters. An alternative hypothesis is proposed, that the combination of high electoral participation and high inequality may be due to the prevalence of clientelistic linkage, frequent in democracies with weak institutions.