This thesis investigates electoral accountability for incumbent leaders and their parties in Latin America. It addresses two central questions. First, it seeks to explain sources of contextual and institutional variation in the degree to which voters punish and reward incumbents. Second, it asks how voters hold incumbents accountable for other areas of government control besides the macro-economy. e first paper develops a framework of executive accountability as dependent on the degree to which the presidential candidate of the incumbent party is identified with the performance of the outgoing president. It differentiates between presidents running for re-election, successors, and non-successors. Estimating random-intercept random-slopes models on an original dataset, it shows different levels of accountability for the three types of candidates. e second paper examines whether endorsements from incumbent politicians to co-partisans lead to more electoral sanctioning. It uses a randomised experiment embedded in a national survey conducted in the run-up to the Mexican general election to demonstrate that Senate candidates endorsed by the outgoing president are held more electorally accountable. Using a difference-in-difference design, the third paper finds causal evidence of tactical allocation of federal funds to municipalities governed by co-partisans of the president in the run-up to the Mexican presidential election. It shows that voters rewarded the party in the federal government for these additional transfers to their municipalities by voting for the ruling party. Taken together, the three papers have important theoretical implications for the study of electoral accountability, comparative electoral behaviour, and the quality of democracy in Latin America.