Abstract
Various studies predict large migration flows due to climatic and other environmental changes, yet the ex post empirical evidence for such migration is inconclusive. To examine the causal link between environmental changes and migration for a population residing along the Jamuna River in Bangladesh, an area heavily affected by floods and riverbank erosion, I relate the respondents' self-reported affectedness by environmental changes, their migration aspirations, and their capability to move to their migration likelihood. The analysis relies on a unique quasi-experimental research design based on original survey panel data of 1604 household heads. I find that erosion significantly and substantively increases the likelihood to migrate, leading to more than a doubling of the migration likelihood compared to the unaffected control group. Flooding has a significant effect only if it causes severe and irreversible impacts. Moreover, erosion affectedness increases the likelihood of moving permanently, with the whole household, and to a rural location. Individual, temporary moves to urban locations, by contrast, are primarily driven by low socio-economic status. Those who move with the whole household migrate mostly less than five kilometers. These findings call for a more nuanced understanding of the complex environment-migration nexus. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s11111-025-00478-7.