Using a natural experiment of a rainfall-driven remittances, I provide experimental measures of how remittances affect rural household’s choice of cylinder gas (LPG) as a cooking fuel over other alternative fuels in southern Bangladesh. Household choice of LPG and remittances are jointly related; therefore, I use the instrumental variable probit (IV-Probit) approach. The treatment of remittances is randomly assigned to households who suffered losses due to a natural shock from the cyclone-Roanu enabling the instrument – exogenous variation in rainfall interacted with cyclone-affected migrant household’s distance to the local weather stations – to identify the average treatment effect for the treatment group (cyclone-affected remittances recipient households). I find that an exogenous increase in remittances by 1,000 Taka causes the probability of using LPG to rise by 1%. In terms of percentage change, the implied elasticity shows that a 10% increase in remittances income can raise the probability of using LPG by 2%. I also find the impact of remittances is conditional on household’s health expenditures. In particular, controlling for the household’s health expenditures interacted with the provision for clean water and sanitary toilet in the dwelling, the marginal effects of remittances get stronger, i.e. households are more likely to use LPG as cooking fuel. These findings counter some existing case studies and views of many policy makers that economic factors are less significant in promoting cleaner energy for the household. The results of the paper are robust to potential violations of the exclusion restriction, to alternative specifications and instruments, and possible omitted variable bias.