E-cigarette Price Impacts legal and Black-Market Cigarette Purchasing Under a Hypothetical Reduced-Nicotine Cigarette Standard

Abstract

Introduction

The Tobacco Control Act gives the U.S. Food and Drug Administration authority to establish a reduced-nicotine content standard in combusted cigarettes. This future potential regulation may pose a significant public health benefit; however, black markets may arise to meet demand for normal-nicotine content cigarettes among smokers unwilling to transition to or use an alternative product.

Aims and Methods

We determined the behavioral-economic substitutability of illicit normal-nicotine content cigarettes and e-cigarettes for reduced-nicotine content cigarettes in a hypothetical reduced-nicotine regulatory market. Adult cigarette smokers were recruited online to complete hypothetical cigarette purchasing tasks for usual-brand cigarettes, reduced-nicotine content cigarettes, and illicit normal-nicotine content cigarettes, as well as a cross-commodity task in which reduced-nicotine content cigarettes were available across multiple prices and illicit cigarettes were concurrently available for $12/pack. Participants completed two three-item cross-commodity purchasing tasks in which e-cigarettes were available for $4/pod or $12/pod alongside reduced-nicotine content cigarettes and illicit cigarettes.

Results

Usual-brand cigarette purchasing was greater than illicit normal-nicotine content cigarettes and less than reduced-nicotine content cigarettes. In the cross-commodity purchasing tasks, illicit cigarettes and e-cigarettes both served as economic substitutes for reduced-nicotine content cigarettes; however, when e-cigarettes were available for $4/pod, they were purchased at greater levels than illicit cigarettes and resulted in greater reductions in reduced-nicotine content cigarettes purchasing than when available for $12/pod.

Conclusions

These data suggest that some smokers are willing to engage in illicit cigarette purchasing in a reduced-nicotine regulatory environment, but e-cigarette availability at lower prices may reduce black-market engagement and shift behavior away from combusted cigarette use.

Implications

E-cigarettes available at low, but not high, prices were stronger substitutes for legal, reduced-nicotine content cigarettes than illegal, normal-nicotine content cigarettes in a hypothetical reduced-nicotine tobacco market. Our findings suggest the availability of relatively inexpensive e-cigarettes may reduce illicit cigarette purchasing and combusted cigarette use under a reduced-nicotine cigarette standard.

Introduction

The 2009 Family Smoking Prevention and Tobacco Control Act provided the U.S. Food and Drug Administration (FDA) authority to set a reduced-nicotine content standard in cigarette tobacco. In June 2022, the Biden-Harris Administration published plans for the FDA to develop this potential future regulatory action.¹ A reduced-nicotine content standard has been proposed as a means for preventing tobacco use initiation and reducing nicotine dependence among current smokers in the United States.² Indeed, clinical studies evaluating cigarettes with reduced-nicotine content have been shown to reduce cigarette consumption, reduce cigarette toxicant exposure, and increase both the likelihood and success rate of attempts to quit in current smokers.^3–5

Despite the potential public health benefits of a reduced-nicotine standard in the United States, it would represent an unprecedented change to cigarette composition, and black markets may emerge to meet normal-nicotine content cigarette demand among smokers unwilling to transition.^6–8 Cigarette black markets have developed in North America in response to taxation and usually involve interstate smuggling of cigarettes from low-tax states.^8,9 Cigarettes within these black markets can be sold by the pack or as individual “loosie” cigarettes,^8–11 and illicit tobacco sales have been estimated to comprise 20%–30% of the Canadian tobacco market¹² and between 8.5%–21% of the U.S. tobacco market (estimates from 2008 to 2011, respectively).⁸ Because a reduced-nicotine standard would apply federally, illicit normal-nicotine cigarette sales would require a different approach, likely involving international smuggling or illicit domestic manufacture. The prevalence of regular tobacco use in the United States significantly outnumbers regular use of illicit drugs,¹³ and a small fraction of current smokers willing to engage in illicit cigarette purchasing could provide a substantial customer base to motivate black-market sales. Although myriad factors would influence the nature and scale of a black market,^8,9 it could potentially provide revenue for organized crime (either large, central distributors or small, and localized organizations) and increase the prevalence of more-harmful unregulated cigarettes (eg, cigarettes containing adulterants from illicit manufacture),¹⁴ which could limit the maximum effectiveness of a nicotine standard. In response to a temporary, pandemic-associated cigarette prohibition in South Africa, local black markets emerged to meet local cigarette demand and survey data suggests sustainability for these channels despite significant price increases.¹⁵ Although a nicotine standard is not as severe as prohibition, the findings from South Africa highlight the importance of considering illicit trade amid a large-scale, non-price change in cigarette regulation.

Cigarette smokers reticent to break the law may adopt use of other, legally available nicotine products to replace or supplement reduced-nicotine content cigarettes, such as e-cigarettes. Although not harmless, e-cigarettes provide considerably less toxicant exposure than combusted cigarettes (eg,^16,17), and transitioning from combusted cigarettes to e-cigarettes poses a health benefit to current smokers.^18,19 E-cigarettes have been suggested as a reduced-harm alternative to combusted cigarettes and a means to reduce or remain abstinent from combusted cigarette use,^17,20,21 and preliminary clinical trials have indicated e-cigarette use may promote significant reductions in combusted cigarette consumption (eg,^22,23).

The potential outcomes of a reduced-nicotine marketplace can be simulated with hypothetical purchasing tasks. Hypothetical purchasing tasks, in which hypothetical tobacco consumption is reported for a specified time period across different prices and subsequently modeled with demand curve analysis, have been used extensively to evaluate the reinforcing properties of reduced-nicotine content cigarettes in both clinical trials and laboratory studies,^24–26 as well as online convenience samples in which the conditions of a reduced-nicotine tobacco standard were hypothetically modeled.^27,28 Dose-dependent reductions in hypothetical cigarette consumption have been observed in blinded laboratory studies.^24–26 Conversely, hypothetical reduced-nicotine content cigarette purchasing was greater than for normal-nicotine cigarettes²⁷ in online convenience samples, suggesting anticipated compensatory smoking when nicotine dose is made apparent as well as a non-nicotinic reinforcing effect of cigarette use.²⁸ Hypothetical purchasing tasks have also been used to evaluate e-cigarette demand and have demonstrated that e-cigarettes can substitute for combusted cigarettes (eg,^29–31). Demand curve analysis of anticipated tobacco consumption patterns in response to hypothetical reduced-nicotine and illicit cigarette marketplaces may inform policy and avoid possible unintended consequences of regulation. Thus, we aimed to determine the potential impact of a reduced-nicotine tobacco standard on consumption of illicit, normal-nicotine content cigarettes in a hypothetical black market and the extent to which legally available e-cigarettes may substitute for combusted cigarettes using hypothetical purchasing tasks.

Methods

Survey Release

Data were collected across two survey releases (September 2019; January–April 2020). Participant recruitment methods and compensation were identical across survey releases and participants who completed the study were prevented from completing a second time (ie, within or across releases).

Participants

Participants were recruited through the online crowdsourcing platform Amazon Mechanical Turk (mTurk) and data were collected using Qualtrics survey software (Provo, UT). The mTurk platform contains numerous tasks, such as surveys, and allows registered workers to self-select tasks to work on and complete, if eligible. Individuals were eligible for the study if they reported: ≥18 years of age, living in the United States, fluent in English, smoking ≥5 cigarettes/day, a mTurk approval rating of ≥95%, and previous completion of ≥100 mTurk tasks. The mTurk “approval rating” refers to the proportion of tasks completed by a worker that were approved (ie, deemed as being of adequate quality) by the task owners, and a higher approval rating with a larger total number of completed tasks indicates historically higher-quality response patterns, which may improve the likelihood of quality, thoughtful responses.^32–34 To mask the inclusion criteria and target sample, the study was advertised on mTurk as a survey on health and decision-making, the screening survey on Qualtrics asked participants to identify all the products they used from a list, which included cigarettes among various loosely related products, such as alcohol, cannabis, and over-the-counter medications, and participants were asked how frequently they visited a physician. Participants who met the inclusion criteria on the screening survey were provided a link to the full survey which began with an overview of the study and a consent page. Participants were paid $1.00 USD for completing the survey and an additional $2.00 for passing the embedded attention check (described below).

Procedure

Participants completed a basic demographic questionnaire, provided tobacco product use patterns (ie, frequency of use, magnitude of consumption), indicated whether they would purchase normal-nicotine content cigarettes illegally if a reduced-nicotine standard were implemented (“yes,” “no,” or “maybe”), and completed the Fagerström Test of Cigarette Dependence.^35,36 Two open-ended questions were included at the end of the survey which asked, “What do you like best about your usual brand of cigarettes?” and “Do you have any comments about the survey?.”

Hypothetical Purchasing Tasks

Participants completed up to six separate hypothetical purchasing tasks for legal, reduced-nicotine content cigarettes, illegal, normal nicotine “black market” cigarettes, and nicotine vapor pods. The full instructional sets for each task are presented in Supplementary Materials. Participants in both releases first completed three separate, single-commodity purchasing tasks for their current usual brand of cigarettes, legal reduced-nicotine content cigarettes, and illegal “black market” cigarettes. In each single-commodity task, only one type of cigarette was available for purchasing. In the “usual brand” single-commodity task, participants reported how many of their current, usual-brand cigarettes they would purchase across prices, and there was no mention of reduced-nicotine or illicit trade in the associated instructions. In the reduced-nicotine content cigarette single-commodity purchasing task, participants were told that a reduced-nicotine standard had been implemented nationwide and now the only available cigarettes contain 2% of their normal nicotine, but all other aspects of the cigarette (eg, taste, smell) were the same. In the “black market” single-commodity task, participants were told that a complete tobacco product ban had been implemented and the only way to purchase cigarettes, which are the same as their current usual brand, would be to purchase them illegally through a black-market dealer and that being caught could result in cigarette confiscation and a fine from law enforcement. We utilized a complete tobacco product ban in this task to address participants’ willingness to engage in black-market cigarette purchasing while avoiding possible implicit substitution (ie, participants reporting 0 purchasing with the intention of getting nicotine from alternative sources). For all cigarette purchasing tasks, participants purchased packs of cigarettes for use over the course of 1 week at the following prices per pack, which were presented on the same page: $0.01, 0.10, 1.00, 1.70, 3.20, 5.60, 10.00, 17.00, 32.00, 56.00, and 100.00. In the cigarette purchasing task, participants were asked four multiple-choice and true/false questions related to task instructions to verify understanding.

In the second release, participants also completed a cross-commodity hypothetical purchasing task in which legal, reduced-nicotine content cigarettes (2% nicotine) and illegal, normal-nicotine content cigarettes (current usual brand) were concurrently available for purchase. In this task legal, reduced-nicotine content cigarette pack prices increased across the same range as above while the price of illegal, normal-nicotine content cigarette packs remained fixed at $12/pack across reduced-nicotine content cigarette prices. This fixed price was an estimate by investigators regarding the possible price of illicit cigarettes, assuming a nearly doubling of the average price per pack in the United States. In addition to completing a cross-commodity purchasing task with adjusting-price reduced-nicotine content cigarettes and fixed-price black-market cigarettes, participants completed two additional cross-commodity tasks wherein they could purchase adjusting-price reduced-nicotine content cigarettes, fixed-price ($12) black-market cigarettes, or fixed-price e-cigarettes, which were described as “nicotine vapor pods” with a per-pod nicotine content equivalent to a pack of cigarettes, that were available for $4/pod in one task and $12/pod in the other. Each price was presented on a separate page.

Data Analysis

Multiple methods were used to ensure data integrity and prevent repeated sampling. Security settings within Qualtrics for both the screening survey and the main survey were enabled to prevent multiple submissions of either. We also embedded a script in each survey batch on mTurk to further prevent multiple attempts. An attention check was embedded within the Fagerström Test of Cigarette Dependence that asked “which of the following is not round” with the options: a brick, a tire, a frisbee, a basketball. Participants who did not select “a brick” failed the attention check and were excluded from analysis. Participants who reported purchasing 0 packs of their usual-brand cigarettes at $0.01/pack in the Usual-Brand single-commodity purchasing task were excluded from analysis. Finally, the open-ended responses were independently evaluated by two research staff to identify nonsensical or inappropriate response patterns indicative of inattention, rushing, non-fluent English, or potentially automated responding (ie, “bots”). If there were discrepant evaluations of a response between the two reviewers, a senior researcher made the final determination of response validity. Participants who provided responses deemed inappropriate were excluded from analysis.

Seventy packs/week (10 packs/day) was considered the biological limit and consumption values over 70 were recoded as 70 across all tasks (adjusting-price: n = 851 recoded of 53 570 total responses (1.59%); fixed-price: n = 138 recoded of 30 602 total responses (0.45%)). Responses were subsequently transformed such that outliers (Z > 3.29) were recoded as one higher than the highest non-outlier value at each price for each task (adjusting-price: n = 1048 of 53 504 total responses (1.96%); fixed-price: n = 416 recoded of 30 602 (1.36%)).³⁷ Consumption of fixed-price commodities (illicit cigarettes, e-cigarettes) was subsequently transformed by adding 1 and log-transforming to avoid zero consumption and price was log-transformed for subsequent linear analyses in log-log space.

Adjusting-price cigarette consumption was modeled using the exponentiated demand equation: $\mathrm{Q}=Q_0\ast {10}^{k(e^{-Q_{0}C}-1)}$ </mathgraphic>,³⁸ in which Q is consumption, Q₀ is demand intensity, or consumption at near-zero price, α is demand elasticity, or sensitivity to price, and k is the log range of consumption (fixed at k = 1.775). Q₀ and α were estimated using nonlinear mixed-effects methods, which allows simultaneous estimation of demand parameters (Q₀,α) and product/task and subject-specific parameters using maximum likelihood estimation (eg,³⁹). The parameters Q₀ and α were fit in log₁₀ space. In each model, a random intercept was specified for participant. A blocked covariance matrix was specified for each model with the first block containing a symmetric covariance matrix for participant and the second block containing a diagonal covariance matrix for product or condition, depending on the model (eg,⁴⁰).

Separate models testing for main effects of product (Usual-Brand, Reduced Nicotine, Illicit) were made for single-item demand analysis with one model testing the interaction between survey release and product and another without survey release as a predictor. Because both survey releases completed the single-item tasks, participants were nested within release in both models. Using the second release data, three models were constructed to evaluate change in reduced-nicotine cigarette demand across conditions: (1) alone versus with illicit cigarettes only, (2) alone versus with e-cigarettes at both prices and illicit cigarettes, and (3) alone versus illicit only versus with e-cigarettes at both prices.

Purchasing of fixed-price commodities was modeled using linear mixed-effects methods. Illicit cigarette purchasing was modeled across reduced-nicotine cigarette prices with random intercepts for participants and price. Illicit cigarette and nicotine vapor pod purchasing was simultaneously modeled with product (e-cigarettes, illicit cigarettes) and e-cigarette price ($4, $12) as interacting predictor terms across reduced-nicotine cigarette prices with random intercepts for participant and price.

ANOVAs were conducted for each model to determine statistical significance of model terms, and post hoc comparisons of estimated marginal means (EMM) for demand metrics were conducted to determine significance across products or conditions. Data analysis were conducted in R (version 3.6.2; R Core Team) using the following packages for primary analyses: nlme,⁴¹lme4,⁴²beezdemand,⁴³ and emmeans.⁴⁴

Results

Participants

A total of 3863 individuals completed the screening survey in the first release and 547 qualified, completed the survey, and were retained for analysis. A total of 5417 individuals completed the screening survey for the second release and 447 qualified, completed the survey, and were retained for analysis. Participant demographics of completers from both survey releases are presented in Table 1.

Table 1.

Participant Demographics

Demographic variable	First release (n = 547)	Second release (n = 447)	Combined (n = 994)
Race/ethnicity (n, %*)
White	422 (77.1)	344 (77.0)	766 (77.1)
Black/African American	56 (10.2)	45 (10.1)	101 (10.2)
Hispanic or Latino	31 (5.7)	22 (4.9)	53 (5.3)
Asian	25 (4.6)	21 (4.7)	46 (4.6)
Other race or ethnicity**	13 (2.4)	15 (3.4)	28 (2.8)
Gender (n, %)
Female	251 (45.9)	182 (40.7)	433 (43.6)
Male	296 (54.1)	265 (59.3)	561 (56.4)
Age (mean, SD)	37.2 (10.9)	36.6 (10.7)	36.9 (10.8)
Daily cigarette consumption (mean, SD)	13.5 (7.9)	12.6 (9.5)	13.1 (8.7)
FTCD score† (mean, SD)	5.2 (2.7)	4.3 (2.2)	4.8 (2.5)
Unwilling to purchase illicit cigarettes (n, %)	130 (23.7)	95 (21.3)	225 (22.6)

^*Percentages are rounded up and may not add to 100.

^**Other races or ethnicities reported include: American Indian or Alaska Native, Native Hawaiian or Pacific Islander, More than one race, “Other.”

^†Fagerström Test for Cigarette Dependence.³⁵

Single-Item Purchasing

Actual and estimated single-item cigarette purchasing are illustrated in Figure 1. An ANOVA of log-Q₀ determined a significant main effect of product (F(2,31766) = 66.56, p < .001) but no main effect of survey release (p = .632) or release-by-product interaction (p = .444), and an ANOVA of log-α determined no significant effects of product or survey release or a release-by-product interaction (all ps > .05). Thus, a separate model without survey release was used for estimating demand parameters, which also indicated a main effect of product for log-Q₀ (F(2,31772) = 66.53, p < .001) but not log-α (p = .336). EMM for log-Q₀ and log-α from this model are presented in Table 2. Planned comparisons of log-Q₀ EMM demonstrated significantly higher log-Q₀ for reduced-nicotine content cigarettes, and significantly lower log-Q₀ for illicit cigarettes under a tobacco ban, respectively, compared to usual-brand cigarettes.

Figure 1.

Single-commodity demand curves. Demand curves illustrating the median number of packs purchased and estimated consumption from the three single-commodity purchasing tasks for Usual-Brand cigarettes (filled circles, solid line), Reduced-Nicotine Cigarettes (empty circles, dotted line), and black-market, normal-nicotine cigarettes (triangles, dashed line). Note the logarithmic x-axis and linear y-axis.

Table 2.

Estimated Marginal Means for Log-Transformed Demand Intensity (Q₀) and Elasticity (a)

Condition	logQ0	loga
Single-Item*
Usual Brand	0.93 (.013)a	−2.73 (.016)
Reduced-Nicotine	0.96 (.016)	−2.72 (.016)
Illicit Cigarettes	0.77 (.018)a	−2.74 (.018)
Cross-Commodity**
Reduced-Nicotine + Illicit Only	0.87 (.028)a,b	−2.62 (.032)a
Reduced-Nicotine + Illicit + $4 E-cigarettes	0.78 (.028)a,b	−2.63 (.030)a
Reduced-Nicotine + Illicit + $12 E-cigarettes	0.83 (.027)a,b	−2.65 (.031)a

*Means determined from model without survey release as predictor (n = 994).

**Means determined from model with all reduced-nicotine cigarette conditions.

^asignificantly different from Reduced-Nicotine Content Cigarettes alone, p < .05.

^bsignificantly different from each cross-commodity condition, p < .05.

Illicit Cigarette Substitution

Actual and estimated purchasing of concurrently available reduced-nicotine content and illicit cigarettes are illustrated in Figure 2. Log-Q₀ for reduced-nicotine content cigarettes was significantly lower when illicit cigarettes were concurrently available (EMM = 0.891, SE = 0.026) compared to reduced-nicotine content cigarettes alone (EMM = 0.980, SE = 0.024; F(91,9363) = 27.065, p < .0001). Log-α was significantly higher for reduced-nicotine content cigarettes with concurrently available illicit cigarettes (EMM = −2.62, SE = 0.301) than for reduced-nicotine content cigarettes alone (EMM = −2.71, SE = 0.0280; F(1,9363) = 9.600, p = .002). In the linear model, reduced-nicotine cigarette price was not a significant predictor of illicit cigarette purchasing (F(1,446.07) = 2.9823, p = .085), suggesting economic independence.

Figure 2.

Illicit cigarette substitution. Demand curves illustrating the median number of packs purchased and predicted consumption for reduced-nicotine cigarettes available alone from the single-commodity purchasing task (empty circles, solid line) compared to the adjusting-price reduced-nicotine cigarettes (filled circles, dashed line) and fixed-price black-market, normal-nicotine cigarettes (triangles, solid line) from the cross-commodity purchasing task.

Three-Item Purchasing

Reported and estimated purchasing of concurrently available reduced-nicotine, illicit, and electronic ($4 or $12/pod) cigarettes are illustrated in Figure 3. In the first nonlinear mixed model, with reduced-nicotine content cigarettes alone and with $4 or $12 e-cigarettes and illicit cigarettes, reduced-nicotine content cigarette log-Q₀ significantly differed between each condition (F(2,14267) = 35.392, p < .0001), with highest intensity in the alone (EMM = 0.958, SE = 0.032), followed by $12 e-cigarette (EMM = 0.825, SE = 0.028), then $4 e-cigarette conditions (EMM = 0.779, SE = 0.028). Similarly, reduced-nicotine content cigarette log-α significantly differed across conditions (F(2,14267) = 5.757, p = .0032), with highest elasticity/log-α associated with concurrently available illicit cigarettes and $4 e-cigarettes (EMM = −2.64, SE = 0.03) or $12 e-cigarettes (EMM = −2.67, SE = 0.03) than alone (EMM = −2.73, SE = 0.03), but log-α did not differ between e-cigarette price conditions (ps > .05). EMM for log-Q₀ and log-α from the second nonlinear mixed model, which also included reduced-nicotine cigarette purchasing with only illicit cigarettes concurrently available, are presented in Table 2. Reduced-nicotine content cigarette log-Q₀ significantly differed between each condition (F(3,19171) = 29.615, p < .0001), with the highest intensity in the alone, then illicit substitution, $12 e-cigarettes, and $4 e-cigarettes conditions. Similarly, as in the first model, reduced-nicotine content cigarette log-α was significantly different across conditions (F(3,19171) = 5.321, p = .012), with concurrently available illicit cigarettes, and $4 e-cigarettes with illicit cigarettes or $12 e-cigarettes with illicit cigarettes being significantly more elastic than reduced-nicotine content cigarettes alone, but log-α did not differ between multi-commodity conditions (ps > .05).

Figure 3.

Impact of e-cigarette price on cigarette purchasing. Demand curves illustrating the median number of packs or pods purchased and estimated purchasing for reduced-nicotine cigarettes available alone from the single-commodity purchasing task (empty circles, solid line) compared to the adjusting-price reduced-nicotine cigarettes (filled circles, dotted line), fixed-price black-market, normal-nicotine cigarettes (triangles, dashed-dotted line), and fixed-price nicotine vapor pods (squares, dashed line) from the three-commodity purchasing task when nicotine vapor pods were available for $4/pod (left) or $12/pod (right).

The linear mixed model of illicit cigarette and e-cigarette purchasing at both e-cigarette prices ($4/$12) determined significant main effects of price (F(1,455.9) = 16.934, p < .0001) (ie, substitution), product (F(1,18785.5) = 27.718, p < .0001) and e-cigarette price (F(1,18785.5) = 622.585, p < .0001) and a product-by-price interaction (F(1,18785.5) = 1258.247, p < .0001), such that e-cigarette purchasing was greater than illicit cigarette purchasing at $4, but lower at $12.

Discussion

This study evaluated potential responses to a reduced-nicotine content standard by assessing hypothetical purchasing of usual-brand cigarettes, reduced-nicotine content cigarettes, illicit, normal-nicotine content cigarettes, and e-cigarettes at different prices. When examining hypothetical purchasing patterns of cigarettes without alternative products available, we found that demand intensity (ie, purchasing at prices approaching 0) for usual-brand cigarettes was significantly lower than legally available reduced-nicotine content cigarettes and significantly higher than illicit, normal-nicotine content cigarettes under a tobacco ban. When illicit, normal-nicotine content cigarettes were concurrently available for $12/pack, illicit cigarettes functioned as an independent product for reduced-nicotine content cigarettes and reduced demand intensity and increased demand elasticity (ie, price sensitivity) of reduced-nicotine content cigarettes compared to when available alone. When $4/pod e-cigarettes were concurrently available with illicit cigarettes and reduced-nicotine content cigarettes, they were stronger substitutes (ie, elicited greater consumption) for reduced-nicotine content cigarettes than illicit cigarettes and promoted the greatest reductions in demand intensity for reduced-nicotine content cigarettes; however, when e-cigarette price was increased to $12/pod, illicit cigarettes were stronger substitutes and the changes in demand intensity and elasticity were less pronounced.

The primary findings from these analyses indicate that many current U.S. smokers are willing to illicitly purchase cigarettes with normal nicotine content in a reduced-nicotine cigarette regulatory environment. More than three-quarters of participants across both survey releases reported a willingness (“yes” or “maybe”) to purchase black-market, normal-nicotine cigarettes in a marketplace where the only legally available cigarettes had reduced-nicotine content, and this was further evidenced by the high engagement in illicit cigarette purchasing in the hypothetical purchasing tasks. Furthermore, illicit cigarette purchasing was independent of reduced-nicotine content cigarette price, and reduced-nicotine content cigarette purchasing was significantly lower when available with illicit cigarettes relative to when available alone, which together suggest that some smokers may forego reduced-nicotine content cigarette purchasing in favor of illicit cigarettes. In the single-commodity tasks, illicit cigarette purchasing was lower than both reduced-nicotine content and usual-brand cigarettes, potentially indicating an aversion to illicit marketplace engagement by many participants who are likely to purchase more reduced-nicotine cigarettes legally or perhaps quit smoking altogether (ie, not purchasing either cigarette type). Nevertheless, the engagement by such a substantial proportion of the sample in black-market cigarette purchasing is concerning. Overall, these data buttress the findings of Hall and colleagues, who demonstrated substantial interest in illicit cigarette purchasing among smokers learning of a reduced-nicotine standard⁴⁵ and suggest an emergence of a black market for normal-nicotine cigarettes in a reduced-nicotine marketplace, consistent with the legal prohibitions against other psychoactive substances.¹⁴ Sufficient demand for illicit normal-nicotine content cigarettes amid a federal reduced-nicotine standard may require additional efforts by law enforcement to intercept cigarettes smuggled internationally and to monitor counterfeit cigarette production domestically and abroad as illicit trade shifts from interstate to international trafficking.

Although e-cigarettes have been demonstrated to substitute for combusted cigarettes in studies utilizing hypothetical purchasing tasks (eg,^29–31), these findings are novel in that a second alternative reinforcer (illicit cigarettes) was available and e-cigarettes were still readily purchased. Furthermore, when available for a lower price than illicit cigarettes, e-cigarettes were purchased to a greater degree, suggesting that when they are inexpensive relative to black-market prices, e-cigarettes are better substitutes than illicit cigarettes. Conversely, a significant price-by-product interaction indicated that when e-cigarettes and illicit cigarettes were available for the same price, illicit cigarette purchasing exceeded e-cigarette purchasing and reduced-nicotine cigarette consumption was greater than when e-cigarettes were available for $4. These data suggest that restricting access to e-cigarettes (ie, through excessive taxation) may push consumers toward combusted cigarette consumption, either through legal reduced-nicotine or illicit normal-nicotine markets. Indeed, analysis of consumer tobacco purchasing and market surveillance data has demonstrated significant increases in combusted cigarette consumption upon levying e-cigarette taxes.⁴⁶ Related to the current data, a substantial proportion of e-cigarette users were willing to purchase e-cigarettes illicitly in response to more comprehensive product bans in a study assessing hypothetical tobacco purchasing under different e-cigarette regulatory market conditions (no ban, non-tobacco flavor ban, and total vaping ban).⁴⁷ Although e-cigarette use is not harmless, it appears to be considerably less toxic than combusted cigarette use, and risks will be reduced further once only e-cigarettes with FDA marketing authorization are available on the market.^16–19 Therefore, it is within the interest of public health to provide access a less-toxic alternative nicotine source amid major policy change. The current experiments directly evaluated the influence of price (ie, taxation) on relative consumption of e-cigarettes versus combusted cigarettes, but the impacts on combusted cigarette consumption and illicit product use determined in this study may be applied to policy measures that restrict e-cigarette access in adult smokers (eg, state-wide product bans, severely limiting flavor availability). E-cigarette regulation has been a point of contention among U.S. clinicians, scientists, and policymakers, with e-cigarette proponents highlighting their potential to reduce cigarette-related harms and opponents viewing them as solely perpetuating nicotine addiction.⁴⁸ Our findings suggest e-cigarettes may reduce combusted cigarette consumption and limit harms posed by the illicit tobacco market by potentially reducing engagement in it.

Across both survey releases, demand intensity for reduced-nicotine cigarettes was greater than for usual-brand cigarettes, indicating anticipated compensatory smoking in this sample. These findings run counter to previous behavioral-economic analyses of reduced-nicotine content cigarette consumption that demonstrated comparable or dose-dependently reduced hypothetical consumption of reduced-nicotine cigarettes relative to normal-nicotine cigarettes^24–26; however, these previous analyses were conducted under double-blind conditions in which the participants were unaware of the nicotine content in their combusted cigarette or that nicotine content had been manipulated at all. The implementation of a reduced-nicotine content cigarette standard would essentially remove the blind, and smokers would be aware that the nicotine content in their usual-brand cigarettes had decreased and may resort to smoking more to satisfy their cravings. An unblinded clinical study comparing use of exclusively available normal-nicotine content or reduced-nicotine content cigarettes demonstrated compensatory use of reduced-nicotine content cigarettes (ie, increased cigarettes/day, expired carbon monoxide) only in the initial 24–48 hours of the 4-day study,⁴⁹ and in subsequent interviews, participants indicated that they initially anticipated smoking more reduced-nicotine cigarettes prior to beginning the study.⁵⁰ These data from unblinded clinical studies may explain the discrepancy in hypothetical cigarette purchasing between the current data and blinded clinical studies,^24–26 and together with the current data, these findings suggest that smokers may anticipate smoking more when switching to reduced-nicotine content cigarettes. However, it is possible that any increases in cigarette consumption would be transient. These findings highlight the necessity of conducting additional unblinded clinical studies to model a potential reduced-nicotine regulatory environment more realistically, and to determine the influence of expectancy effects on reduced-nicotine content cigarette consumption.

This study has limitations. First, the study was hypothetical in nature among a convenience sample of smokers without reduced-nicotine cigarette experience. Hypothetical purchasing tasks have been shown to approximate cigarette consumption,^51,52 and to model acceptability and uptake of reduced-nicotine content cigarettes in reduced-nicotine-naïve individuals,^27,28 but the hypothetical nature of the study and lack of experience with reduced-nicotine content cigarettes are considerable limitations nonetheless. The study also only assessed two tobacco products (combusted and electronic cigarettes), but smokers may turn to alternative products (eg, cigars, smokeless tobacco, and nicotine replacement therapy) if nicotine restrictions are not similarly placed on them. Furthermore, in the instructional sets for tasks featuring black-market purchasing, participants were told they were purchasing their current usual-brand cigarettes from an illicit dealer and that being caught by law enforcement could result in confiscation and a fine. However, products sold in illicit markets may have limited diversity (ie, only certain brands or illicitly manufactured, generic cigarettes available) and lack of usual brand may limit the appeal of illicit cigarettes. Similarly, the severity of legal consequences associated with illicit tobacco trade under a potential nicotine-restriction policy are currently unknown and might vary geographically, which may influence illicit purchasing behavior. Results from a study addressing the influence of monetary-fine magnitude on hypothetical illicit e-cigarette purchasing provided some insight into this issue by demonstrating reductions in the likelihood of illicit marketplace engagement as monetary fines increased.⁴⁷ Our data provide a basis for future studies to investigate the impact of product availability (ie, cigarette brand and type), demographics, and punishment severity on hypothetical illicit cigarette purchasing and alternative product substitution. Additionally, in both survey releases, participants were predominately white and between 25 and 45 years of age (the latter is likely representative of the younger-skewing demographics of mTurk’s worker population^32–34), which may limit the generalizability of these findings to the broader population of U.S. cigarette smokers.

Taken together, the findings from this study suggest that current smokers anticipate compensatory smoking and report a willingness to engage in an illicit cigarette marketplace to maintain consistent nicotine consumption under a reduced-nicotine standard. E-cigarettes may provide a less-toxic alternative to combusted cigarettes in this marketplace; however, restricting access to e-cigarettes through punitive taxation may turn tobacco users to combusted cigarettes in one marketplace or another. Before implementing a reduced-nicotine standard, coordination among governing bodies to standardize the regulatory approach for e-cigarettes at federal, state, and local levels based on empirical evidence will be essential to mitigate harms and maximize public health benefits of this tobacco standard and deter the emergence of a combusted cigarette black market.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

Support for this study, including participant compensation and author effort, was provided by National Institute on Drug Abuse grants R01 DA042527 (MWJ), R01 DA035277 (MWJ), and T32 DA007209 (SBD). These funding sources had no involvement in the study beyond financial support.

Declaration of Interests

None declared.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.