Health expenditures, remittances, and climate vulnerability: Evidence from Bangladesh

Remittances’ effect on a household’s health outcome (e.g. Infant mortality) is ambiguous, but the impact on health expenditure is positive and less equivocal in literature. This paper puts the relationship between health expenditure and remittances into a stress test to see whether it survives the adverse impact of climate change. Using a natural experiment of rainfall-driven remittances, I provide an experimental measure for remittances’ effect on the health expenditure among rural households in southern Bangladesh. Health expenditure and remittances are jointly related; therefore, I use the instrumental variable 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 while remittances cause household health expenditures to increase, the marginal effect of remittances is heterogeneous and negative conditional on the household’s exposure to the level of vulnerability proxied by the household’s distance to cyclone shelter. In other words, the health expenditure-remittances nexus gets weaker with the adverse effect of climate change. Specifically, I find that an increase in remittances by a Taka increases health expenditure by 0.24 Taka (24 Paisa) in the absence of any climate hazard but reduces health expenditure by 0.10 Taka (or 10 Paisa) if the measure of climate vulnerability increases by one standard deviation from its mean value. For countries like Bangladesh, which is exceptionally vulnerable to natural hazards, climate vulnerabilities can render the financing of health care costs through remittances unsustainable even if households receive regular and sizable flows.