Gradual switching regression estimates of alcohol demand elasticities

Given that alcohol remains a heavily taxed good, studies of the demand for alcohol are numerous in the literature. This paper uses annual data from 1964-92 to estimate key elasticities of the US demand for distilled spirits. Unlike previous studies, which typically assume elasticities are constant over time, we allow elasticities to vary over time by estimating a gradual switching regression model. The results indicate that the demand for distilled spirits today differs substantially from the 1960s, suggesting that the efficacy of proposed policies is dependent on knowledge of this change in demand.