Hooked on Flavor: Addiction, Present Bias, and the Consequences of E-Cigarette Flavor Policy

Flavored e-cigarettes have captured millions of users who are disproportionately teens and young adults, while bans on them have triggered one of the most highly controversial regulatory debates in recent nicotine public health policy. To evaluate e-cigarette flavor policy, I develop and estimate a dynamic discrete choice model of cigarette and e-cigarette demand with forward-looking, present-biased consumers, featuring nicotine-driven addiction stocks, a flavored habit stock, and heterogeneous flavor preferences that vary with household age composition. Using NielsenIQ scanner data for 2021–2023, I first present reduced-form evidence documenting strong habit persistence, moderate cross-category substitution patterns, and heightened flavored e-cigarette demand among households with younger members. I then estimate the dynamic model via a two-stage pseudo-maximum-likelihood procedure and simulate counterfactual flavor policies under rational and time-inconsistent discounting. Three findings emerge. First, the feared cigarette substitution backfire does not materialize: the dominant response to a ban is market exit rather than switching to cigarettes, driven primarily by strong flavor preferences. Second, enforcement scope is salient: restricting the ban to only FDA-authorized brands captures only 84% of the addiction reduction, with the gap driven by substitution to unauthorized products. Third, a 0.50/mL excise tax achieves 92% of the ban’s addiction reduction at less than half the welfare cost. Present bias does not alter this policy ranking, but reduces the true welfare cost of both instruments by converting part of the measured loss into an internality correction.