Electoral accountability refers to the mechanism that people are able to use their votes to award or punish incumbent governments based on their policy outcomes and political performance. Conventional studies focusing on electoral accountability in Latin America mainly address the relations between governments’ economic performance and incumbents’ vote share. However, their conclusions are diverse. We argue that this problem might be due to mistaking levels of analyses, incomplete datasets, mistakes in model selection, and ignoring the influence from political contexts. To deal with these problems, in this paper, we collect electoral results and economic indicators from 28 Latin American democracies. Then, by applying the mixed effects model, we introduce political contexts into my empirical analysis and control them as fixed effects. The statistical results prove that electoral accountability does exist in Latin American democracies. The results also show that political contexts influence electoral accountability. Whenever an incumbent government owns a supermajority in a legislature, it will be highly accountable for its electorate. Finally, this paper also finds that in contrast to their counterparts in parliamentary democracies, governments in presidential democracies are more accountable to their voters for their economic performance.