This article seeks to draw possible lessons for adaptation programmes in Bangladesh by examining whether cyclone preparedness and relief interventions are subject to corrupt practices. Based on a random sample survey of 278 households, three focus-group discussions and seven key-informant interviews, the article investigates the nature and extent of corruption in pre- and post-disaster interventions in Khulna before and after Cyclone Aila in May 2009. Ninety nine percent of households reported losses from corrupt practices. Post-disaster interventions (such as food aid and public works schemes) suffered from greater levels, and worse types, of corruption than pre-disaster interventions (such as cyclone warning systems and disaster-preparedness training). Using an asset index created using principal component analysis, the article assesses how corruption affected wealth quartiles. Ultra-poor households were affected more by corruption in pre-disaster interventions, the wealthiest quartile more in certain post-disaster interventions, in particular public works and non-governmental interventions. These findings may hold lessons for attempts to increase resilience as current adaptation measures mirror some cyclone preparedness and relief efforts.