Objective/context: This article analyzes the efficiency level of candidate selection methods for presidential elections in Latin America by governing parties, according to the degree of openness and internal fractionalization. The guiding hypothesis is that methods with greater openness generate greater internal fractionalization and, consequently, offer worse electoral outcomes; whereas more exclusionary methods generate less fractionalization and better electoral outcomes. Methodology: Twenty-three candidate selection processes for the presidency by ruling parties, and an equal number of presidential elections in Argentina, Chile, Mexico, and Uruguay between 1990 and 2020, under conditions of electoral integrity, are comparatively reviewed, and the degrees of openness are analyzed based on the proposal of an internal cohesion index and its effects on electoral success. Conclusions: From the cases analyzed, it can be concluded that there is no relationship between the openness of the selection methods and the fractionalization of the parties and electoral success, in this case, of those in government; and that while there is the idea that more internal democracy increases the probability of winning elections, this actually depends on other variables. Originality: The article aims to contribute to the literature on presidential candidate selection processes, particularly of governing parties, since elections are also an evaluation of the government, and the continuity or change is at stake. The selection method has effects on the parties and can define their efficiency or lack thereof in maintaining power.