Revisiting the effects of cigarette taxation on smoking outcomes

Abstract This study reassesses the efficacy of cigarette taxation in curtailing smoking by leveraging recent advancements in the difference-in-differences (DiD) literature to account for heterogeneous treatment effects. Using data from the Behavioral Risk Factor Surveillance System Selected Metropolitan/Micropolitan Area Risk Trend (BRFSS SMART) for the sample periods 2004-2010 and 2015-2020, the study reveals three key findings. Firstly, the TWFE estimate for the 2004-2010 sample is only 48% of the average treatment effect on the treated (ATT) estimate obtained through the DiD framework. Secondly, event-study-type estimates demonstrate a gradual increase in magnitude following the treatment year, highlighting dynamic treatment effects overlooked by the TWFE estimate. Third, the ATT estimate for the 2015-2020 sample is approximately 66% of the ATT estimate for the 2004-2010 sample. Overall, the study underscores the potential bias toward zero in elasticity estimates when relying solely on TWFE models.