More information about text formats
NOT PEER REVIEWED
We thank Pesko for his comments and the opportunity for us to respond and clarify.
First, we appreciate Pesko’s clarification that Cotti et al. (2020) clustered standard errors to account for clustering. In the present study, we used multilevel analysis not only to account for clustering of respondents (i.e., design effects) but also to incorporate different error terms for different levels of the data hierarchy which yields more accurate Type I error rates than nonhierarchical methods where all unmodeled contextual information ends up pooled into a single error term of the model.
Second, we understand that Cotti et al. (2020) evaluated the magnitude of e-cigarette tax values, which does not contradict to our statement because our study focused on the effects of e-cigarette excise tax policies on individual e-cigarette use and prevalence rather than aggregated sales at state or county levels. We also clearly described the reason why we examined the e-cigarette excise tax policy implementation indicator rather than its magnitude in our paper’s discussion section.
Third, our study used a nationally representative sample of young adults (rather than a nationally representative sample of general adult population). While we understand Pesko’s concern that a sample’s representativeness might be lost when subgroups are explored, we believe our use of sampling weights in analysis has reduced such a concern.
Fourth, in Table 3,...
Fourth, in Table 3, please note that vaping product excise tax policy indicator is a time-variant variable in Model 1. However, to present results of a standard difference-in-differences model with a binary indicator, the policy implementation status was operationalized as a time-invariant variable in Model 2, which is not unusual.
Disclosure: We did not receive any funding from the tobacco industry.
1. Cotti, C. D., Courtemanche, C. J., Maclean, J. C., Nesson, E. T., Pesko, M. F., & Tefft, N. (2020). The effects of e-cigarette taxes on e-cigarette prices and tobacco product sales: evidence from retail panel data. National Bureau of Economic Research. NBER Working Paper No. w26724.
NOT PEER REVIEWED
We appreciate the comments from Bates and the opportunity for us to respond and clarify.
First, Bates' argument heavily relies on the assumption that e-cigarettes and combustible cigarettes are substitutes, which is theoretically possible as some consider vaping as a harm reduction alternative to combustible cigarettes. Empirically, however, there have been mixed findings about whether e-cigarettes and combustible cigarettes are substitutes (or complements). Bates cited Pesko et al. (2020) that concludes e-cigarettes and combustible cigarettes are substitutes, whereas other studies have shown that they are complements. For example, Cotti et al. (2018) found that higher cigarette excise taxes, in fact, decrease sales of both e-cigarettes and combustible cigarettes, suggesting that they are complements. Such mixed results abate Bates' argument that taxing ENDS could lead to more use of combustible cigarettes.
Second, Bates might have ignored that our study focused on young adults aged 18-24 years rather than general adults when examining the effect of vaping product tax on e-cigarette use. Although Pesko et al. (2020) suggests that e-cigarettes and combustible cigarettes are substitutes, the findings are based on the general adult population (average age: 55 years) which may not be generalizable to the young adult population. In fact, one study conducted by Abouk and Adams (2017) indicates that e-cigarettes and combustible ci...
Second, Bates might have ignored that our study focused on young adults aged 18-24 years rather than general adults when examining the effect of vaping product tax on e-cigarette use. Although Pesko et al. (2020) suggests that e-cigarettes and combustible cigarettes are substitutes, the findings are based on the general adult population (average age: 55 years) which may not be generalizable to the young adult population. In fact, one study conducted by Abouk and Adams (2017) indicates that e-cigarettes and combustible cigarettes are not substitutes for young people. Established cigarette smokers may use e-cigarettes as a cessation tool but it is less common in young adults. In addition, even if e-cigarettes and combustible cigarettes are substitutes to some degree, the direction of substitution as well as co-use versus subsequent use should not be overlooked. Studies have shown that e-cigarettes may serve as a gateway to future combustible cigarette smoking among young people. For example, a study conducted by Hair et al. (2021) shows that youth and young adults who reported ever e-cigarette use had significantly higher odds of ever cigarette use one year later. Therefore, e-cigarette use versus combustible cigarette smoking is not simply an issue of substitution in particular among young people.
1. Abouk, R., & Adams, S. (2017). Bans on electronic cigarette sales to minors and smoking among high school students. Journal of Health Economics, 54, 17-24.
2. Cotti, C., Nesson, E., & Tefft, N. (2018). The relationship between cigarettes and electronic cigarettes: Evidence from household panel data. Journal of Health Economics, 61, 205-219.
3. Hair, E. C., Barton, A. A., Perks, S. N., Kreslake, J., Xiao, H., Pitzer, L., ... & Vallone, D. M. (2021). Association between e-cigarette use and future combustible cigarette use: Evidence from a prospective cohort of youth and young adults, 2017–2019. Addictive Behaviors, 112, 106593.
4. Pesko, M. F., Courtemanche, C. J., & Maclean, J. C. (2020). The effects of traditional cigarette and e-cigarette tax rates on adult tobacco product use. Journal of Risk and Uncertainty, 60(3), 229-258.
NOT PEER REVIEWED
I have a number of concerns with the paper as currently written.
1) The authors write: “Besides, none of the previous studies except Pesko et al (15) that examined the associations between vaping product excise tax adoption and ENDS use has accounted for the clustering of respondents within the same localities…” This is not accurate, as citation 19 also clusters standard errors at the locality level in all specifications.
2) The authors write: "A working paper reported reduced ENDS sales, but not ENDS use prevalence or behaviours, after implementation of a vaping product excise tax policy. (19)” This is not accurate, as the cited study uses the magnitude of e-cigarette tax values, rather than an indicator variable for tax implementation. States have adopted e-cigarette taxes of different magnitudes and a number of them (such as California) have changed the magnitudes of these taxes after adoption. All of this variation is used in citation 19, contrary to the current study’s description. It's also unclear from the sentence whether citation 19 studied use and found imprecise estimates, or did not study use. It's the latter and this should be clarified. It's also unclear why the authors did not use magnitude of e-cigarette taxes themselves in the current paper, as has been commonly done in the referenced literature.
3) Authors write they use a “nationally representative sample of US young adults.” I do not beli...
3) Authors write they use a “nationally representative sample of US young adults.” I do not believe this is not accurate. The TUS-CPS sample itself may be nationally representative, but this representativeness may be lost when subgroups are explored.
4) The “vaping product excise tax policy” variable in Table 3 appears to be re-defined mid-table. Based on the discussion of the results, in column 1 it appears that this variable is an indicator equal to 1 only at the time when a state has an e-cigarette tax in place. In column 2 though, this indicator equals 1 when a state ever has an e-cigarette tax in place (even prior to it being in place). The use of the same row for a variable that changes across columns is unusual and can easily lead to the wrong interpretation.
NOT PEER REVIEWED
I would like to make three comments by way of a brief post-publication review.
1. The impacts of vaping tax on smoking have been completely overlooked
For a study of e-cigarette taxation to have any public health relevance, it must consider the impact of e-cigarette prices on *cigarette* demand. Cigarettes and e-cigarettes are economic substitutes. The demand for one responds to changes in the price of the other, an idea well understood in economics and quantified through the concept of cross-elasticity. The paper appears to pay no regard to the impact of vaping taxes on cigarette demand, Yet such effects might easily overwhelm any benefits from reduced e-cigarette use - in fact, impact on demand for other tobacco products and the development of informal markets are by far the most important impacts of a vaping tax. By way of example, a 2020 paper by Pesko et al.  concluded:
"Our results suggest that a proposed national e-cigarette tax of $1.65 per milliliter of vaping liquid would raise the proportion of adults who smoke cigarettes daily by approximately 1 percentage point, translating to 2.5 million extra adult daily smokers compared to the counterfactual of not having the tax."
2. The case for reducing adult vaping by taxation has not been made
The authors have based their paper on an unexamined assumption that it is a justifiable goal of policy to lower rates of adult e-cigarette use. Why should...
The authors have based their paper on an unexamined assumption that it is a justifiable goal of policy to lower rates of adult e-cigarette use. Why should this be a policy goal any more than reducing caffeine use or moderate alcohol use? The goal of public health policy is to address significant harms or self-destructive patterns of use, not to modify behaviours that the authors find distasteful. What are the harms that justify state intervention to reduce adult vaping with a tax? Further, they appear indifferent to welfare costs and the distributional impact of imposing a regressive tax burden on people who use vaping products. Tobacco control advocates should become more familiar with the idea that punitive policies impose harm on users, even though these users are supposed to be the intended beneficiaries. For example, a vaping tax harms families by drawing on the household budget of those who continue to vape.
3. The analysis to support the policy recommendations is wholly inadequate
The authors make over-confident policy recommendations without considering the full range of impacts of the measures they are proposing.
"Our findings suggest that adopting a vaping product excise tax policy may help reduce ENDS use and suppress the increase of ENDS use prevalence among young adults. Considering that there are still a number of US states that have not implemented vaping product excise tax policy, wider adoption of such policy across the nation would likely help mitigate ENDS use prevalence."
Without considering all the possible responses to the tax they support, they may easily be proposing tobacco control policies that do more harm than good. In fact, the most important public health impact of this policy is entirely excluded from the analysis. That is the effect of a vaping tax on smoking or other tobacco use. Given the two orders of magnitude difference in risk between smoking and vaping, only a tiny uptick in smoking would be needed to completely offset the benefits, if any, arising from reduced vaping
 Pesko MF, Courtemanche CJ, Maclean JC. The effects of traditional cigarette and e-cigarette tax rates on adult tobacco product use. J Risk Uncertain 2020;60(3):229–258. https://link.springer.com/article/10.1007/s11166-020-09330-9