This paper estimates the price elasticity of demand for alcohol using Health and Retirement Survey data. To account for unobserved heterogeneity in price responsiveness, we use finite mixture models.
We recover two latent groups, one is significantly responsive to price but the other is unresponsive. Differences between these two groups can be explained in part by the behavioral factors of risk aversion, financial planning horizon, forward looking and locus of control. These results have policy implications.
Only a subgroup responds significantly to price. Importantly, the unresponsive group drinks more heavily, suggesting that a higher price could fail to curb drinking by those most likely to cause negative externalities. In contrast, those least likely to impose costs on others are more responsive, thus suffering greater deadweight loss yet with less prevention of negative externalities.
Request Reprint E-Mail: pa7@duke.edu
________________________________________________________________