The Effect of Smokeless Tobacco (SLT) Excise Taxes on SLT Use Among Youth

Abstract

Background:

Smokeless Tobacco (SLT) use prevalence among youth in the United States (US) is comparable to youth prevalence of cigarette smoking. However, it is in general understudied compared to cigarettes and draws less attention nowadays compared to e-cigarettes (ECs).

Aim:

We estimate the own- and cross-tax elasticities of SLT use among US youth and explicitly test how SLT use changes in response to taxes on SLT, cigarettes, ECs, and beer.

Methods:

We standardized SLT taxes for chewing tobacco, moist snuff, dry snuff, and snus, and computed average SLT taxes. We implemented a logit regression model within the state- and year-fixed effects framework.

Results:

A 10% increase in SLT excise taxes reduced youth SLT use by 4% (p<0.01). This result is primarily driven by males, Whites, multiple non-Hispanic races, other races, and individuals living in non-Appalachian states. In addition, a 10% increase in cigarette taxes increases youth SLT use by 8% (p<0.05), suggesting substitutional effects. A 10% increase in EC and beer taxes reduce SLT use by 0.5% and 2.4% (p<0.01), respectively, suggesting complementary effects.

Conclusion:

Raising excise taxes on SLT products can effectively curtail their usage among the youth population. Furthermore, increasing EC and beer taxes reduces youth SLT use. However, an increase in cigarette taxes leads to an unintended consequence of promoting SLT use among youth. In addition, increasing SLT taxes does not appear to significantly impact the disparities in youth SLT use by whether living in Appalachian states. Future research is needed to assess whether SLT taxes reduce disparities in use by rural/urban divisions.

1. Introduction

Smokeless tobacco (SLT) is a class of tobacco products designed to be placed in the oral or nasal cavity, such as chewing (spit) tobacco, moist/dry snuff, snus, and dissolvable tobacco¹. (Centers for Disease Control and Prevention, 2021) Besides prolonging nicotine addiction, SLT use is linked to cardiovascular diseases, oral cancers, and early delivery and stillbirth among pregnant women. (Centers for Disease Control and Prevention, 2020; Gupta et al., 2018)

Youth are particularly vulnerable to addiction and substance dependence, making them a critical target for prevention efforts. (Biggar Jr et al., 2017; Cohn et al., 2018; Crost & Guerrero, 2012; Kenkel et al., 2001; Lopez-Quintero et al., 2018; Zuckermann et al., 2019) In 2023, 1.5% of high school students in the United States (US) reported the past 30-day use of SLT. (Birdsey, 2023) While this prevalence is lower than that of electronic cigarettes (ECs) at 10%, it is comparable to cigarette smoking (1.9%), cigar smoking (1.8%), and oral nicotine pouches which do not contain tobacco leaves (ONPs, 1.7%). (Birdsey, 2023) In addition, SLT use is higher among males, Whites, and those who are socioeconomically disadvantaged and live in rural areas (e.g., Appalachian regions). (Burke et al., 1988; Pesko & Robarts, 2017; Sinha et al., 2018) An international study reveals that among 356 million SLT users, 66% were male and 34% were female. In high- and upper-middle-income countries, 2% of males and 0.5% of females use SLT. In contrast, in low- and lower-middle-income countries, SLT use is significantly higher, with 18% of males and 10% of females using these products. (Sinha et al., 2018) In the US, Appalachian regions rank higher in tobacco use (including SLT use) and poverty than the national average. (Beatty et al., 2019; Pollard & Jacobsen, 2019) A study found that 7% of adolescents in rural areas used SLT, compared to 2.9% of adolescents in urban areas. (Pesko & Robarts, 2017) Furthermore, the co-use of SLT and cigarettes is high, with nearly 10% of young adults aged 18–24 who smoked cigarettes also reporting SLT use. (U.S. Dept of Health and Human Services, 2020)

While extensive research exists on the effects of cigarette taxes on smoking, studies on SLT taxation are limited. Several studies restricted the sample to male students and found that higher SLT excise tax rates were associated with decreased SLT use. (Chaloupka et al., 1997; Dave & Saffer, 2013; Ohsfeldt et al., 1998; Ohsfeldt & Boyle, 1994; Ohsfeldt et al., 1997; Tauras et al., 2007) With the exception of Dave & Saffer (2013), they generally rely on cross-state variations for identification since they used cross-sectional samples, which may have biased the effects of SLT taxes on SLT use. Furthermore, these studies express SLT taxes as a percentage of the wholesale price of snuff (ad valorem taxes), which has several drawbacks. First, if the wholesale prices vary a lot from state to state, the percentage of wholesale prices may distort the measurement of SLT taxes. Second, snuff tax may not be a good proxy for SLT taxes as SLT consists of a class of smokeless tobacco products.

Economic and public health research has demonstrated that pricing and cost have direct impact on purchase and use behaviors. (Blecher & Van Walbeek, 2004; He et al., 2018; Nargis et al., 2021; Organization, 2003; U.S. National Cancer Institute & World Health Organization, 2016) In addition, with budget constraints and common liability of substance use, the pricing and accessibility of one substance are expected to influence the consumption of another, either in a complementary or a substitutable way. (Markowitz & Tauras, 2009; Vanyukov & Ridenour, 2012) Such economic relationship has been widely studied in the context of tobacco, alcohol, and cannabis. (Cameron & Williams, 1999; Subbaraman, 2016) However, the relationship between SLT and cigarettes is empirically debatable. Some studies suggest they are substitutes, (Ohsfeldt et al., 1998; Ohsfeldt & Boyle, 1994; Ohsfeldt et al., 1997) while others find them to be complements. (Cotti et al., 2016; Dave & Saffer, 2013; Tauras et al., 2007; Zheng et al., 2017) Moreover, as ECs overtake cigarettes as the most popular nicotine or tobacco products among young people, growing attention has been given to the cross-tax elasticities between cigarettes and ECs, which evaluates how taxing ECs may impact cigarette smoking. However, little attention has been given to how EC taxes might impact SLT use. The legalization of medical and recreational cannabis and other substances (e.g., alcohol) may further shape SLT use among youth. Research indicates that medical cannabis legalization (MCL) is associated with increased cigarette-cannabis co-use, and recreational cannabis legalization (RCL) slightly increases EC use. (Coley et al., 2021; Weinberger et al., 2022) However, the impact of these legalizations on SLT use remains unclear.

SLT taxation varies across states in the US, with some imposing taxes based on product weight and others using ad valorem taxes based on retail, wholesale, or manufacturing prices. Furthermore, some states impose differential tax rates on different types of SLT. For example, Illinois imposed $0.03 on per ounce of moist snuff while imposing 36% of the wholesale price on chewing tobacco, dry snuff, and snus. The diversity in tax forms from one state to another and differential tax rates on various SLT within a state create a challenge to compare tax rates across states. The lack of uniform tax measures on SLT taxes hinders the evaluation of the effect of SLT tax on use. All previous studies had to exclude states that imposed specific taxes on SLT and only include states that imposed ad valorem taxes on SLT and used snuff taxes to proxy for SLT taxes. (Chaloupka et al., 1997; Dave & Saffer, 2013; Tauras et al., 2007)

To address these knowledge gaps in the current literature, we build on our prior research that standardized SLT taxes into uniform taxes per ounce and calculate average SLT taxes for each state. (He, 2024) We then examine the own- and cross-tax elasticities, testing how youth SLT use responds to taxes on SLT, cigarettes, ECs, beer, and cannabis legalization, using a logit regression model and a Difference-in-Differences (DiD) approach within the state- and year-fixed effect framework. We also explore the heterogeneity in SLT use by different demographic groups, such as sex, race/ethnicity, and location of residence. Our results will offer valuable insights into the impact of SLT and other related taxes, guiding policy design.

2. Data and Methods

2.1. Data

Outcome variable: Any SLT use

We used the Centers for Disease Control and Prevention (CDC) Youth Risk Behavior Surveillance System (YRBSS) data from 2007 to 2019 to estimate the prevalence of any SLT use among youth. (Centers for Disease Control and Prevention, 2007–2019) YRBSS is a cross-sectional nationally representative biennial survey of the 9th through 12th graders. The surveys are administered every other year to monitor youth risky behaviors, including tobacco and alcohol use. Students complete the self-administered questionnaire during one class period and record their responses directly in a computer-scannable booklet or on a computer-scannable answer sheet. We generated a dichotomous variable to measure any SLT use in the past 30 days based on the following questions: “During the past 30 days, on how many days did you use chewing tobacco, snuff, dip, snus, or dissolvable tobacco products, such as Copenhagen, Grizzly, Skoal, or Camel Snus? (Do not count any electronic vapor products).” The variable took the value of 1 if respondents reported any day of use during the past 30 days and 0 if not.

Key explanatory variables

State SLT excise taxes

State-level excise taxes on SLT (chewing tobacco, moist/dry snuff, snus, and dissolvable) were sourced from the CDC State Tobacco Activities Tracking and Evaluation (STATE) System. (Centers for Disease Control and Prevention, 2023) Since some states impose a specific excise tax on SLT based on the product’s weight, while others impose an ad valorem tax (i.e., a percentage based on the retail, wholesale, or manufacturer price), standardized SLT taxes are needed in the regression analysis.

Specifically, we converted ad valorem taxes to specific taxes ($/ounce) using SLT prices from the Nielsen Retail Scanner Data (NRSD), which contains the price and sales information of various products sold in 30,000–50,000 participating stores in the US. We calculated the average price per ounce of each type of SLT product except for dissolvable tobacco² for each state in 2007, and this price was used to estimate taxes per ounce for all the following years in states that impose ad valorem taxes. This procedure is taken to reduce the influence of time-varying factors, such as state-level tax changes, tax pass-through rate changes, etc., on SLT prices³. Then we calculated standardized taxes ($/ounce) for each type of SLT based on the following formula:

$${standardizedtax}_{s,t}=\frac{{averagepriceperounce}_{s,2007}}{1+{advaloremtaxrate}_{s,t}}\mathrm{*}{advaloremtaxrate}_{s,t}$$ (1)

where $\mathrm{s}$ denotes state and $\mathrm{t}$ denotes year. Since the Nielsen prices include excise taxes, but exclude sales taxes, we first converted prices inclusive of excise taxes to tax-free prices by dividing the average price in 2007 by (1+ad valorem tax rate). Then we calculated standardized taxes by multiplying the tax-free prices by the ad valorem tax rate. For states that imposed ad valorem taxes based on retail prices, the average price refers to the retail price. For states that imposed ad valorem taxes based on wholesale prices, the average price refers to the wholesale price. For states that imposed ad valorem taxes based on manufactural prices, the average price refers to the manufactural price. Following previous literature, we used a 20% markup rate at each transaction stage (manufacture to wholesale, and wholesale to retail) in the supply chain to calculate wholesale and manufacturer prices based on the Nielsen retail price. (Chaloupka & Tauras, 2020) Finally, we adjusted standardized taxes for inflation using the Consumer Price Index (CPI) published by the US Bureau of Labor Statistics. (U.S. Bureau of Labor Statistics, 2024) All SLT taxes are constant in 2010 dollars.

State cigarette, EC, and beer excise taxes

State excise taxes on cigarettes were obtained from Orzechowski and Walker, The Tax Burden on Tobacco. (Orzechowski and Walker, 2020) Cigarette excise taxes were expressed in dollars per pack of 20 cigarettes. State excise taxes on ECs were documented from multiple sources, and then standardized into specific taxes ($ per e-liquid volume) using the 35% retail-wholesale markup rate, following Cotti et al. (2021). (Cotti et al., 2021; Tobacconomics, 2023) State-level specific taxes on per gallon of beer with a 5% alcohol concentration and sold off-premises for each state between 2007 and 2019 were sourced from the Alcohol Policy Information System (APIS). (National Institute on Alcohol Abuse and Alcoholism, 2024) All taxes were adjusted to constant 2010 dollars by using the Consumer Price Index (CPI) published by the US Bureau of Labor Statistics. (U.S. Bureau of Labor Statistics, 2024)

State-level control variables

To account for the potential influence of other state-level tobacco control policies and contextual factors on SLT use, we also controlled for the following variables in our analysis.

Indoor smoking and vaping restriction coverages

State-level indoor smoking and vaping coverages were estimated using data from the American Nonsmokers’ Rights Foundation (ANRF) US Tobacco Control Laws Database, which includes the date of implementation, venues covered by the law (workplace, restaurant, and bar venues), and strength of the laws (100% smoke/vape-free, partial restrictions, and no restriction at all) at the state and local (county, city) levels. (Cheng et al., 2011)

The coverage measures are inclusive of state and local level policies, with adjustment of state preemption (If a state preempted local legislation, city- and county-level policies would no longer apply). Specifically, to calculate the coverages, we linked the policy data with the Census-Estimated Population (CEP) Cities and Towns (Vintage 2019 all states, all geographies data file) to estimate the proportions of state and county populations covered by comprehensive indoor smoking and vaping restrictions, respectively. Comprehensive indoor restrictions are defined as policies that ban smoking or vaping in all indoor areas, including workplaces, restaurants, and bars without exemptions.

The coverage measures range from 0 (no comprehensive restrictions) to 1 (the entire population of the state is covered by a comprehensive ban). We have used the measures in a series of previous studies that estimate the impact of indoor restrictions on tobacco use behaviors. (Cheng et al., 2011; Cheng et al., 2013; Gonzalez et al., 2013)

T21 laws for cigarettes and ECs

The Minimum Legal Sales Age (MLSA) Laws for cigarettes and ECs were sourced from the CDC STATE System. (Centers for Disease Control and Prevention, 2023) We coded the two variables as 1 if a state raised the MLSA for cigarettes and ECs to 21 years old and 0 if otherwise.

Recreational and medical cannabis legalization (RCL and MCL)

We sourced RCL and MCL data from the Insurance Institute for Highway Safety and ProCon.org. (Insurance Institute for Highway Safety, 2024; ProCon.org, 2023) The two variables were coded as 1 if a state has legalized recreational and medical cannabis use and 0 if otherwise.

Seasonally adjusted unemployment rate

State-level seasonally adjusted unemployment rates were obtained from the Bureau of Labor Statistics and controlled for local economic conditions. (U.S. Bureau of Labor Statistics)

Demographic variables

We utilized the following demographic variables from the YRBSS data: sex (female vs. male), grade (9^th, 10^th, 11^th, and 12^th), race/ethnicity (White, Black, Hispanic, Asian, multiple non-Hispanic races, and others), which are either dichotomized or categorical variables. The variable for sex was coded as 1 for male respondents and 0 for female respondents. For categorical grade and race/ethnicity variables, we used dummy coding for each grade level and each race/ethnicity group. These variables were used as explanatory variables in the regression analysis. (He, Fong, et al., 2024; He, Liber, et al., 2024)

2.2. Methods

Since the survey question asked in the YRBSS data did not specify which type of SLT products the respondent used in the past 30 days, we first created average SLT taxes by taking the average of standardized taxes across all types of SLT products⁴. Then we implemented a logit regression analysis within a two-way fixed effects framework. The analytical model is illustrated as follows:

$${SLTUse}_{ist}=\alpha +\beta SLT{Tax}_{st}+Z_{st}\lambda +X_{ist}\gamma +{\delta }_s+{\theta }_t+u_{ist}$$ (2)

where $\mathrm{i}$ denotes individual, $\mathrm{s}$ denotes state, and $\mathrm{t}$ denotes year. ${SLTUse}_{\text{ist}}$ denotes whether individual $\mathrm{i}$ who resided in state $\mathrm{s}$ used any type of SLT product in year $\mathrm{t}$. $SLT{Tax}_{st}$ denotes the average standardized SLT tax in state $\mathrm{s}$ and year $\mathrm{t}$. $Z_{st}$ denotes state-level taxes of relevant products and tobacco control policies in state $\mathrm{s}$ and year $\mathrm{t}$, such as cigarette/EC/beer tax, RCL, MCL, indoor smoking/vaping restriction coverage, T21 laws for cigarettes and ECs, and unemployment rate. $X_{ist}$ denote individual $\mathrm{i}$’s demographic variables. ${\delta }_s$ is the state-fixed effect which can account for time-invariant state-specific factors and ${\theta }_t$ is the year-fixed effect which can account for the common time trend across states. All regressions were weighted using YRBSS sample weights and conducted using Stata 17 (StataCorp LLC, College Station, TX). All standard errors are clustered at the state level to account for inter-temporal correlations within the same state.

The pooled surveys from 2007 to 2019 initially included a total of 101,860 observations. However, some respondents did not provide answers to certain questions (such as SLT use, sex, race, and grades), and some states had missing standardized SLT tax values due to unavailable NRSD sales records. As a result, the sample size for our regressions was reduced to 93,360. To address the missing data, we conducted additional regressions by creating a missing category for sex, race, and grades, which increased the sample size to 95,921. However, the results remained largely the same. Therefore, we chose to use only variables with non-missing values for our final analysis.

We further stratified the analysis by sex, race/ethnicity, and whether living in the Appalachian region (Alabama, Kentucky, Mississippi, North Carolina, New York, Ohio, Pennsylvania, South Carolina, Tennessee, Virginia, and West Virginia). The stratified analysis will provide information about whether taxes are an effective tool to reduce SLT use among specific groups.

3. Results

Table 1 presents summary statistics of the analysis data between 2007 and 2019 using YRBS sample weights. Between 2007 and 2019, about 6.9% of US high school students appeared to use any type of SLT products. During the study period, the average SLT tax was about $0.64 per ounce. Figure 1 shows that SLT use prevalence has been trending downward between 2007 and 2019. In contrast, SLT taxes have been increasing slowly from $0.48 per ounce in 2007 to 0.85 per ounce in 2019 (Figure 2). On average, cigarette tax was $2.231 per pack, EC tax was $0.088 per e-liquid ml, and beer tax was $0.253 per gallon. Statistics for other state-level control variables are provided. In terms of demographics, about half of the sample were female (49.3%). Grades of the sample are distributed quite equally but with slightly more 9^th graders. More than half of the sample were Whites, 13.8% of them were Black, and 21.3% of them were Hispanic. Figure 3 presents any SLT use prevalence by sex, race/ethnicity, and location of residence. It indicated that the prevalence of any SLT use was high among males (11.58%), non-Hispanic White (9.11%), multiple non-Hispanic races (6.34%), other races (9.74%), and youth who lived in Appalachian regions (9.48%).

Table 1:

Summary statistics (2007–2019, n=93,360)

Variables	Mean (SD) or %
Outcome variable
Any SLT use	6.9%
Explanatory variables
SLT taxes
Average SLT tax ($/ounce)	.640 (.446)
Taxes of relevant products
Cigarette tax ($/pack)	2.231 (.933)
Std. EC tax ($/ml)	.088 (.386)
Beer specific tax ($/gallon)	.253 (.244)
State-level policies
Recreational cannabis legalization (RCL)	7.8%
Medical cannabis legalization (MCL)	42.9%
Smoke-free air law strength score for cigarettes	72.2%
Smoke-free air law strength score for ECs	9.5%
T21 law for cigarettes	4.8%
T21 law for ECs	4.8%
SA unemployment rate (%)	6.448 (2.468)
Demographics
Sex-female	49.3%
Sex-male	50.7%
Grade-9th	27.5%
Grade-10th	25.8%
Grade-11th	23.9%
Grade-12th	22.8%
Race-White	56.0%
Race-Black	13.8%
Race-Hispanic	21.3%
Race-Asian	3.5%
Race-Multiple non-Hispanic races	4.1%
Race-others	1.4%

Note: Data were weighted using YRBS sample weights. All taxes were adjusted for inflation using 2010 dollars. SA: seasonally adjusted.

Figure 1: SLT use prevalence, 2007–2019.

Figure 2: Average standardized SLT taxes, 2007–2019.

Figure 3: SLT use prevalence by sex, race/ethnicity, and location of residence.

In Table 2, we provide the main results of the effect of SLT tax on any SLT use among youth. The coefficients reported are odds ratios. Across Models 1 through 3, we incrementally added state-fixed effects and state-level control variables. Our preferred specification is Model 3, which shows that a $1 increase in the average SLT tax is negatively associated with any SLT use among youth (p<0.01) with a corresponding tax elasticity of −0.393. This indicates that a 10% increase in average SLT tax will result in a 3.93% reduction in any SLT use among youth. Figure 4 depicts the own- and cross-tax elasticities from model 1 through model 3. The effects of state-level tobacco control policies and contextual factors on SLT use are not statistically significant.

Table 2:

The effect of average SLT tax on any SLT use among youth.

Variables	Model 1	Model 2	Model 3
Average SLT tax	.571*** (<0.001) [.442, .738] <−.337>	.791 (0.359) [.479, 1.306] <−.141>	.521** (0.005) [.331, .821] <−.393>
Cigarette tax			1.495* (0.028) [1.045, 2.139] <.841>
Std. EC tax			.543*** (<0.001) [.430, .687] <−.052>
Beer specific tax			.360*** (<0.001) [.233, .556] <−.240>
Recreational cannabis legalization (RCL)			1.116 (0.737) [.589, 2.114] < .1012>
Medical cannabis legalization (MCL)			.898 (0.488) [.661, 1.219] < −.101>
Demographics	Y	Y	Y
Year-fixed effects	Y	Y	Y
State-fixed effects	N	Y	Y
State-level tobacco control policies	N	N	Y
Observations	93,360	93,360	93,360

Note: The estimates reported are odds ratios. P-values, confidence intervals, and elasticities are reported in parentheses, square brackets, and angel brackets, respectively. Robust standard errors are clustered at the state level. All regressions are weighted using YRBS sample weights.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001

Figure 4: The own- and cross-tax elasticities of any SLT use.

In terms of cross-tax elasticities, we find that an increase in cigarette taxes appears to increase any SLT use among youth (p<0.05). The cross-tax elasticity of 0.841 implies that cigarettes and SLT products among youth are economic substitutes and that a 10% increase in cigarette taxes leads to an 8% increase in any SLT use among youth. For ECs and beer, the negative cross-tax elasticities suggest that ECs and beers are economic complements for SLT use among youth (p<0.001). Specifically, a 10% increase in EC and beer taxes leads to a 0.5% and 2.4%, respectively, decrease in SLT use among youth. RCL and MCL are not significantly associated with SLT use.

Table 3 presents the effect of SLT tax on any SLT use among youth by delineating the sample by sex. The results indicate that higher SLT taxes are associated with reduced SLT use among male youth. However, for female youth, SLT taxes did not significantly impact SLT use. Interestingly, recreational cannabis legalization (RCL) significantly increased SLT use among female youth, suggesting that recreational cannabis serves as an economic substitute for SLT in this demographic.

Table 3:

The effect of average SLT tax on any SLT use by sex.

Variables	Female	Male
Average SLT tax	.689 (0.188) [.395, 1.201] < −.235 >	.488** (0.009) [.285, .836] < −.412 >
Cigarette tax	1.697*** (<0.001) [1.344, 2.143] < 1.162>	1.462 (0.080) [.955, 2.237] < .751 >
Std. EC tax	.529** (0.003) [.347, .806] < −.059>	.540*** (<0.001) [.403, .723] < −.047>
Beer specific tax	.207*** (<0.001) [.114, .376] < −.386>	.388*** (<0.001) [.240, .626] < −.214>
Recreational cannabis legalization (RCL)	3.136** (0.007) [1.368, 7.189] <1.119>	.847 (0.613) [.444, 1.615] <−.147>
Medical cannabis legalization (MCL)	1.012 (0.948) [.696, 1.473] <.012>	.872 (0.452) [.610, 1.246] <−.121>
Any SLT use prevalence	2.11%	11.58%
Observations	46,970	46,150

Note: The estimates reported are odds ratios. P-values, confidence intervals, and elasticities are reported in parentheses, square brackets, and angel brackets, respectively. Robust standard errors are clustered at the state level. All regressions are weighted using YRBS sample weights.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001

In Table 4, we now delineate the analysis sample by race/ethnicity group. Overall, the results are comparable to those in Table 2, especially among groups that have a relatively high prevalence, including Whites (9.11% prevalence), multiple Non-Hispanic (6.34% prevalence), and others (9.74% prevalence). For groups with a relatively lower SLT use prevalence, such as Asians (3% prevalence) and Hispanics (4.43% prevalence), the SLT taxes were not significantly associated with SLT use. In addition, for black youth, higher SLT taxes increase SLT use, which does not follow the law of demand, suggesting that SLT may be inferior goods (i.e., demand decreases as prices drop) for this group. Given that the SLT use prevalence among Black youth is lowest among all race/ethnicity groups, they may prefer other products to SLT.

Table 4:

The effect of average SLT tax on any SLT use by race/ethnicity.

Variables	White	Black	Hispanic	Asian	Multiple NH	Others
Average SLT tax	.371** (0.004) [.188, .731] < −.574>	2.673** (0.008) [1.291, 5.536] <.471>	.909 (0.701) [.556, 1.483] <−.068>	.862 (0.840) [.202, 3.673] <−.110>	.326* (0.044) [.110, .969] <−.708>	.081*** (<0.001) [.024, .281] <−1.559>
Cigarette tax	1.655* (0.026) [1.063, 2.576] < 1.036>	.753 (0.283) [.448, 1.264] <−.565>	1.501 (0.108) [.914, 2.464] <.886>	2.206** (0.006) [1.260, 3.863] <1.889>	1.212 (0.688) [.473, 3.106] <.408>	2.281 (0.092) [.874, 5.952] <1.708>
Std. EC tax	.542*** (<0.001) [.428, .687] < −.053>	.215 (0.179) [.023, 2.026] <−.062>	.354** (0.002) [.183, .686] <−.097>	.341 (0.721) [.001, 124.781] < −.198>	.325 (0.203) [.058, 1.832] <−.105>	1.411 (0.669) [.292, 6.827] <.028>
Beer specific tax	.402*** (<0.001) [.278, .584] < −.195>	.283*** (<0.001) [.155, .517] <−.458>	.073*** (<0.001) [.022, .247] <−.543>	.099 (0.134) [.005, 2.039] <−.514>	.029*** (0.001) [.004, .226] <−.861>	.001* (0.012) [.000, .229] <−1.483>
RCL	.992 (0.978) [.572, 1.721] <−.007>	1.526 (0.619) [.288, 8.077] <.411>	1.788 (0.156) [.800, 3.995] <.555>	5.303 (0.096) [.746, 37.704] <1.615>	4.060 (0.143) [.623, 26.469] <1.311>	.623 (0.715) [.049, 7.887] <−.424>
MCL	.877 (0.415) [.641, 1.201] <−.119>	.642 (0.291) [.282, 1.462] <−.431 >	1.244 (0.355) [.783, 1.978] <.209>	1.179 (0.848) [.220, 6.307] <.159>	1.021 (0.969) [.346, 3.012] <.020>	.699 (0.492) [.252, 1.940] <−.321>
Any SLT use prevalence	9.11%	2.7%	4.43%	3.00%	6.34%	9.74%
Observations	40,771	16,412	26,432	3,361	4,054	1,832

Note: The estimates reported are odds ratios. P-values, confidence intervals, and elasticities are reported in parentheses, square brackets, and angel brackets, respectively. Robust standard errors are clustered at the state level. All regressions are weighted using YRBS sample weights.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001

Table 5 shows the impact of SLT taxes on any SLT use delineated by residency in the Appalachian regions. Among youth living in states with Appalachian regions, increasing SLT taxes did not result in changes in youth SLT use, despite a relatively high SLT use prevalence of 9.48% compared to the rest (6.1%). In contrast, higher SLT taxes significantly decreased youth SLT use among those who live in states without Appalachian regions (p<0.01). These findings suggest that while SLT taxes do reduce youth SLT use overall, they may not be effective in reducing SLT use disparities among youth in Appalachian vs. non-Appalachian regions.

Table 5:

The effect of average SLT tax on any SLT use by region of residence.

Variables	Non-Appalachian States	Appalachian States
Average SLT tax	.467*** (0.002) [.287, .761] < −.525>	.288 (0.158) [.051, 1.621] < −.427>
Cigarette tax	1.167 (0.390) [.821, 1.658] < .329>	2.889*** (0.001) [1.571, 5.314] < 2.074>
Std. EC tax	.548*** (<0.001) [.405, .742] < −.067>	12.619 (0.764) [.000, 1.98e+08] < .013>
Beer specific tax	.091*** (<0.001) [.027, .301] < −.501>	.366*** (<0.001) [.251, .532] < −.313>
Recreational cannabis legalization (RCL)	1.482 (0.245) [.764, 2.876] <.370>	Omitted
Medical cannabis legalization (MCL)	1.101 (0.586) [.779, 1.556] <.090>	.579* (0.034) [.349, .960] <−.495>
Any SLT use prevalence	6.1%	9.48%
Observations	73,478	21,467

Note: Appalachian states include AL, KY, MS, NC, NY, OH, PA, SC, TN, VA, and WV. The estimates reported are odds ratios. P-values, confidence intervals, and elasticities are reported in parentheses, square brackets, and angel brackets, respectively. Robust standard errors are clustered at the state level. All regressions are weighted using YRBS sample weights.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001

4. Discussion

While the extant literature has focused on understanding the impact of EC use among youth, the use of SLT has received relatively less attention despite its negative health impacts and non-marginal youth prevalence in the US, which is comparable to that of cigarette smoking. (American Cancer Society, 2020; Birdsey, 2023; Centers for Disease Control and Prevention, 2020) In this study, we estimate the own- and cross-tax elasticities of SLT use among the US youth using the recent data from YRBS between 2007 and 2019. Empirical results showed that increased SLT taxes lead to a reduction of any SLT use among youth with an inelastic tax elasticity of −0.393, which is comparable to the estimates of previous studies that ranged from −0.04 to −0.20. (Chaloupka et al., 1997; Levy et al., 2017; Tauras et al., 2007)

This result confirms that increasing taxes remains a significant and effective policy instrument to date to reduce tobacco use among youth. Existing literature has found that youth are very responsive to cigarette prices and taxes. (U.S. National Cancer Institute & World Health Organization, 2016) Our findings further add to the literature by showing the effectiveness of SLT taxes in reducing SLT use among youth, especially among male and White subgroups that have a relatively high prevalence of SLT use.

Given the effectiveness of SLT taxes in reducing SLT use among youth, policymakers may consider raising SLT taxes. As of 2020, the tax incidence on SLT is only around 20%, significantly lower than that on cigarettes and closed-system ECs. (He, Yang, et al., 2024) In addition, the tax pass-through to SLT prices appears to be lower for lower-priced SLT products, suggesting tax avoidance or price minimization opportunities. (He, Yang, et al., 2024) Future policies may consider reforming the bases and rates of SLT taxes, such as changing ad valorem taxes to specific taxes, to increase tax pass-through rates to prices at all levels. Price promotion restrictions and minimum pricing laws may further increase the effectiveness of taxes in reducing SLT use among youth. (Huang et al., 2016; McLaughlin et al., 2014)

While we found that SLT taxes to be effective in reducing SLT prevalence in various demographic groups, especially Whites, multiple non-Hispanic races, other races, and males, who have a relatively high SLT use prevalence, the potential of SLT taxes in reducing disparities in SLT use by Appalachian vs. non-Appalachian regions may be limited. This conclusion needs to be taken with caution. Due to the lack of information on county identifiers and rural vs. urban residence in YRBSS data, we conducted stratified analysis by whether a state has places or counties that fall into the Appalachian regions. While this stratification or definition did identify states falling into Appalachian regions on average have a higher SLT use prevalence, it may not fully reflect the disparities in SLT use by rural and urban areas, which have been documented in the literature. Nonetheless, this finding does shed light on the need for pricing policies beyond taxation, such as price promotion restrictions and minimum pricing laws that could enhance the effects of prices on behaviors.

Youth are particularly susceptible to developing substance dependence. (Crews et al., 2007; Gray & Squeglia, 2018; Whitesell et al., 2013) Existing evidence has shown that youth are at risk of experimenting with several substances and poly substance use prevalence is typically higher among youth compared to adult samples. (Conway et al., 2013; Kilpatrick et al., 2000; Moss et al., 2014; Zuckermann et al., 2020) Our findings on cross-substance impacts are consistent with previous findings, which demonstrate the significant economic relationship between SLT and cigarettes. (Da Pra & Arnade, 2009; Ohsfeldt et al., 1998; Ohsfeldt & Boyle, 1994) Specifically, higher cigarette taxes increase SLT use among youth, suggesting that SLT is an economic substitute for cigarettes and that increasing cigarette taxes may lead to unintended consequences of increasing SLT use. This finding is consistent with previous studies. (Ohsfeldt et al., 1998; Ohsfeldt & Boyle, 1994; Ohsfeldt et al., 1997) However, given that cigarettes are arguably the most harmful tobacco products, raising cigarette taxes is effective in reducing smoking, and that the SLT use prevalence is relatively low (1.5% among high school students in 2023), the unintended consequence of increasing SLT use may be limited and can be mitigated by increasing SLT taxes.

We further found that higher beer taxes reduce SLT use among youth, suggesting that SLT is an economic complement to beer. This is consistent with previous studies. (Ohsfeldt et al., 1997) An increase in EC taxes also leads to reduced SLT use among youth, suggesting that SLT also serves as an economic complement to ECs. These findings imply that increasing EC and beer taxes not only reduces the use of these products among youth but also leads to less SLT use. The cost and benefit analysis for EC and beer taxes therefore should take account of such benefits. Moreover, beer and other alcohol taxes have been significantly eroded by inflation and income increases over time. Our findings add to the rationale of increasing beer and other alcohol taxes for public health benefits.

We did not find a significant relationship between youth SLT use and MCL or RCL for the overall sample. However, we found a significant effect of RCL on SLT use among female students. This result should be viewed with caution as only 2% of female students use SLT compared to 11.58% of male students.⁵ It is worth noting that these results do not illustrate the economic relationship between cannabis and SLT for youth due to the lack of cannabis tax or price variables. As a growing number of states in the US legalize recreational cannabis and impose taxes on cannabis, youth are increasingly exposed to cannabis taxes. Future research may utilize cannabis taxes to ascertain whether SLT is a complement or substitute for cannabis.

In summary, our study contributes to the literature in several aspects. It provides recent evidence on the impact of SLT taxes on youth SLT use using a quasi-experimental DiD approach, updating older studies from 1994–2013. (Cotti et al., 2016; Dave & Saffer, 2013; Ohsfeldt et al., 1998; Ohsfeldt & Boyle, 1994; Tauras et al., 2007) By including all US jurisdictions and leveraging standardized SLT taxes, we offer a more comprehensive analysis. Additionally, we provide significant evidence of the relationships between SLT and cigarettes, ECs, and alcohol, respectively, which have been understudied in the literature. As ONPs, a distinct category of oral nicotine products similar to snus, have been gaining popularity (use prevalence among high school students increased from 1.1% in 2021 to 1.4% in 2022 to 1.7% in 2023), policymakers have been considering how to appropriately regulate them. (Birdsey, 2023; Gentzke, 2022; Park-Lee, 2022) Understanding the impacts of SLT taxes on youth use will inform possible nicotine pouch regulations as both products are similar in delivery mode, pack styles, and sizes.

Yet, our study has a few limitations. First, as we rely on large-scale self-reported survey data for our empirical analysis, the data itself may be subject to self-reported biases such as recall bias by respondents. YRBSS data also do not contain detailed information on which type of SLT that youth reported using. Therefore, we cannot carry out the analysis by SLT types. Second, our SLT tax measures are the average of taxes imposed on four types of SLT products, calculated without market share weighting due to the lack of market share data for each type of SLT. However, a concurrent study has shown that tax incidences for different SLT product types are very similar⁶, (He, Yang, et al., 2024) Therefore, we do not expect market shares to significantly impact the levels of standardized SLT taxes. Third, due to the lack of detailed geographical identifiers, we can only stratify analysis by whether a state has places falling into Appalachian regions, which does not reflect accurately Appalachian vs. non-Appalachian regions. This could also reduce the statistical power to detect tax impacts in the Appalachian regions or states. Fourth, we did not assess the economic relationship between ONP use and SLT use due to the lack of ONP tax data and use prevalence data in national surveys. Future research is needed when such data becomes available. Finally, although our studies demonstrate that taxes on SLT, beer, cigarettes, and ECs impact youth SLT use, whether these taxes can be designed and implemented in coordination to reduce tobacco use may face significant challenges. As income and inflation grow, specific taxes may also face erosions and thereby become less effective. Future research on implementation and dissemination is needed to facilitate taxation policies.

5. Conclusion

Raising excise taxes on SLT products can effectively curtail their usage among the US youth population, especially among Whites, multiple non-Hispanic races, other races, and male students. Furthermore, increasing EC and beer taxes reduces youth SLT use. However, an increase in cigarette taxes leads to unintended consequences of promoting SLT use among youth. In addition, increasing SLT taxes does not appear to significantly impact the disparities in SLT use by whether a state has Appalachian regions, which may be due to a lack of statistical power. Future research is needed to assess whether SLT taxes reduce disparities in use by rural/urban divisions.

Highlights.

• Our study provides the most recent evidence of the impact of SLT taxes on youth SLT use using a quasi-experimental design, a DiD approach, while previous studies date back to 1992–2013.

• We included all US jurisdictions to examine the SLT tax impact, by leveraging standardized SLT taxes in empirical analysis, which was not used in the existing literature.

• We provided significant evidence of the relationships between SLT and cigarettes, ECs, and alcohol (proxied by beer use), respectively, which has been understudied in the literature.

Data Availability Statement

The data sets generated and/or analyzed during this study are available from the corresponding author on reasonable request