This paper examines the impacts of alcohol expenditure on poverty. A main challenge of doing so is to remove the endogeneity arising from the bidirectional relationship between the two variables that seriously distorts the estimated correlation coefficients. This paper proves that an instrumental variable (IV) probit model outperforms the treatment effect and the panel probit models in terms of purging this endogeneity. When this IV probit model is applied to the Vietnamese data, the results show that alcohol expenditure increases the likelihood of consumers becoming poor; therefore, it is crucial to prioritize the regulation of alcohol spending as a fundamental element in the execution of poverty alleviation initiatives.