A Dynamic Structural Model of Rational Addiction: Evaluating Tobacco Product Flavor Bans and Nicotine Caps

This paper develops a dynamic model of rational addiction with quasi-hyperbolic discounting to evaluate flavor bans and nicotine caps. Using household-level panel data, I first present reduced-form evidence of habit-persistence, forward-looking behavior, and the short-run effects of statewide flavor bans to motivate the structural model. Then, I estimate the structural model via pseudo-maximum likelihood, and simulate counterfactual policies such as a national flavor ban and the FDA’s 2025 proposed nicotine cap, revealing long-run behavioral responses to such policies. The model accounts for nicotine intake along with heterogeneity in flavor preferences and demographics, offering insights into behavioral spillovers and regulatory design.