Various studies predict large environmental migration flows due to climate change, yet the ex post empirical evidence for such migration is inconclusive. In particular, we lack robust evidence on the impact of repeated and concurrent climate events on migration. For instance, a climate event such as a flood might not lead to migration if it occurs once, but might make people move if it occurs several times in a row, or if it occurs in combination with another climate event, such as a storm. We investigate these questions relying on original survey panel data of 1537 household heads residing along the Jamuna River in Bangladesh, an area heavily affected by environmental changes. We relate respondents’ self-reported affectedness by riverbank erosion and floods in three consecutive monsoon seasons to their likelihood to migrate. We find that affectedness by erosion significantly increases the likelihood of migration. However, this effect gets weaker if households are affected repeatedly. Flooding has a negative effect, which becomes significantly stronger for repeated exposure. In terms of concurrent events, respondents who are affected severely by both flooding and erosion in the same monsoon season are more likely to migrate than those affected by only one of the two. A decline in socio-economic status after suffering repeated or concurrent impacts might lead to “trapped populations”. These findings have significant policy implications by underlining that repeated and concurrent affectedness by environmental shocks might reduce out-migration from vulnerable regions.