Working Paper: NBER ID: w24212
Authors: Henry Saffer; Daniel Dench; Dhaval Dave; Michael Grossman
Abstract: Over the past few years, adult use of e-cigs has been increasing while adult smoking has been declining. Although there is a negative time correlation, it is important to determine if there is a causal effect of e-cig use on smoking. This is important because of the known health hazards associated with smoking. A key concern with most prior studies of e-cigs and smoking is that causality between e-cig use and cigarette use is ignored. One contribution of this paper is to estimate structural and reduced form equations that replace e-cig use with e-cig price in order to avoid this endogeneity problem. The data employed to estimate the empirical models come from the Tobacco Use Supplements (TUS). These data are from the combined July 2014, January 2015 and May 2015 waves of the TUS. The results show that e-cig use increases the probability of a quit attempt, the probability of a quit failure, the number of quit failures and the probability of a quit success. It is also estimated that a 10% federal excise tax on e-cigs would reduce the number of quitters in the US by more than 250,000 per year.
Keywords: e-cigarettes; smoking cessation; public health; taxation
JEL Codes: I18
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| e-cig use (L96) | probability of quit attempt (C25) |
| e-cig use (L96) | probability of quit failure (C41) |
| e-cig use (L96) | probability of successful quitting (C25) |
| probability of quit attempt (C25) | probability of quit failure (C41) |