Mortality is considered one of the main costs of dispersal. A reliable evaluation of mortality, however, is often hindered by a lack of information about the fate of individuals that disappear under unexplained circumstances (i.e. missing individuals). Here, we addressed this uncertainty by applying a Bayesian mortality analysis that inferred the fate of missing individuals based on information from individuals with known fate. Specifically, we tested the hypothesis that mortality during dispersal is higher than mortality among non-disperser using 32 years of mark-resighting data from a free-ranging population of the endangered African wild dog (Lycaon pictus) in northern Botswana. Contrary to expectations, we found that mortality during dispersal was lower than mortality among non-dispersers, indicating that higher mortality is not a universal cost of dispersal. Our findings suggest group living can incur costs for certain age classes, such as limited access to resources as group density increases, that exceed the mortality costs associated with dispersal. By challenging the accepted expectation of higher mortality during dispersal, we urge for further investigations of this key life-history trait, and propose a robust statistical approach to reduce bias in mortality estimates.