We aim to identify robust determinants of current account imbalances in 87 emerging and developing countries in the context of model uncertainty and selection bias, from 1980 to 2017. The uncertainty arises from the diversity of competing theories and the contradictory findings of empirical works, while the systemic unavailability of current account data causes selection bias. To overcome these issues, we extend Bayesian Model Averaging (BMA) approach to HeckitBMA in order to address model uncertainty in the presence of selection bias. We find that out of 21 potential determinants, only 10 are robust determinants. However, these determinants differ according to country groupings, and since the 2008 financial crisis, some determinants have gained importance, such as inflation targeting, macroprudential policies, and income inequality. Finally, we find that the determinants of current account surpluses are not strictly the same as those of deficits. Our results shed more light on how policy makers can manage current account imbalances and which levers to use.