Many governments in weak democracies “manage” the electoral process to make their defeat very unlikely. We aim to understand the decision whether to vote or abstain in such managed elections. We focus on the 2018 general election in Bangladesh and randomize the salience of either (i) the citizens' view that their vote signals their policy preferences to the government or (ii) their view that their voting participation and high turnout increase the winning party's legitimacy. These treatments both increase voting participation in government strongholds, but decrease voting participation in opposition strongholds, with the legitimacy treatment having much stronger effects. We further document gender differences in treatment effects, which are partly explained by women who are constrained in their voting decisions. Our results have important implications for get-out-the-vote and information campaigns in weak democracies.