Large shifts in countries’ external current account positions can be disruptive, often reflecting sudden stops in the flows of external finance and leading to exchange rate and banking crises. As a result, an empirical literature has emerged on the sustainability of, and the determinants of large swings in, current account positions. We shed further light on this issue by testing for the presence of unit roots in the current account balance-to-GDP ratios of a large set of mature and emerging-market economies using a methodology that allows for structural breaks in intercepts and trends. We then construct a chronology of current account reversals that is consistent with sustainability of external positions and use it to estimate the factors explaining the likelihood and magnitude of such reversals using a selection model with ordered probit in the first stage. We find that most of the factors that explain the probability of reversals, such as trends in capital flows, in the budget balance and in external positions, also influence their magnitude. But there are a few exceptions. For instance, the stance of monetary policy and the magnitude of external imbalances prior to a reversal seem to be more powerful predictors of the probability of reversals than of their magnitude.