“We should do everything we can to prevent young people from smoking and save lives. Increasing the tobacco age to 21 will help achieve these goals… Increasing the tobacco age to 21 [will also] reduce the likelihood that a high school student will be able to legally purchase tobacco products for other students and underage friends.” Campaign for Tobacco-Free Kids (2020)

Tobacco use is the leading cause of preventable death in the United States, responsible for over 480,000 deaths each year (Centers for Disease Control and Prevention 2019). Combustible tobacco consumption (i.e., cigarettes, cigars, cigarillos, hookah) has been linked to increased risk of heart disease, cancers of the lung, liver, head, and colon, diminished respiratory function, and stroke (U.S. Department of Health and Human Services 2014). The social costs of tobacco consumption are substantial, with estimates of the annual health care costs of treating tobacco-related illnesses totaling nearly $200 billion (2019$) (Hall and Doran 2016). While the prevalence of exposure to secondhand tobacco smoke has fallen in the last decade (Brody et al. 2021), estimates of the external costs of smoking reach approximately $7 billion (USDHHS 2014).1

With the goal of preventing tobacco initiation and nicotine addiction, tobacco control advocates often target teenagers and young adults for reduced access to tobacco (Lantz et al. 2000). While the average age of smoking initiation has risen sharply over the last two decades (Barrington-Trimis et al. 2020), a substantial share of young adult smokers initiated cigarette use when they were teenagers. Moreover, electronic nicotine delivery systems (ENDS) has rapidly emerged as an alternative nicotine source (Creamer et al. 2019; Cullen et al. 2019; Centers for Disease Control and Prevention; 2019). In 2019, 50 percent of U.S. high school students reported they had used an ENDS product and nearly a third were current ENDS users, a rate nearly five times higher than the current smoking participation rate (CDC 2021).

While one's late teenage years have been identified as a potentially important window on the path from experimental tobacco use to everyday use (CDC 2020; Bonnie 2015; Department of Health and Human Services 2014; Khudar et al. 1999), such evidence does not necessarily imply that policy interventions to curb tobacco use will enhance social welfare. To the extent that teenagers’ initiation of tobacco product use is (1) experimental, utility enhancing, and non-habit forming, (2) the result of decisions that rationally consider the possibility of addiction (Becker and Murphy 1988), or (3) an explicit harm reduction strategy (i.e., using ENDS rather than combustible tobacco products), policy interventions to prevent self-harm to teens could be welfare diminishing. On the other hand, if youths give insufficient weight to the future costs of addiction when making current tobacco consumption decisions (Gruber and Köszegi 2001; Crettez and Deloche 2020) — perhaps due to time-inconsistent preferences caused by an underdeveloped prefrontal cortex (Romer 2010) — policies that restrict access to tobacco for such persons could, in theory, enhance social welfare.2

This study examines the effects of a far-reaching policy designed to reduce teenagers’ access to tobacco products: Tobacco-21 (T21) laws. T21 laws raise the minimum legal purchasing (and sales) age (MLPA) for all tobacco products, including combustible tobacco products (i.e., cigarettes, cigars, cigarillos, hookah), ENDS (i.e., e-cigarettes), and smokeless tobacco (i.e., snus) from 18 (in most states) to 21.3 Hawaii was first state to implement a statewide T-21 law on January 1, 2016, followed by California on July 9, 2016.4 Fifteen (15) additional states did so through 2019.5

Among the objectives of T21 laws were: (1) to reduce consumption of all tobacco products (i.e. combustible tobacco products and ENDS) among 18-to-20-year-olds, (2) to reduce tobacco consumption among teens under age 18 who may rely on other 18-year-olds in their high schools/peer groups to obtain their tobacco products (or may look at 18-year-olds’ behavior to role model), and (3) to slow the path to addiction (i.e., everyday tobacco use) among those under age 21. This study is the first to assess the impact of statewide T21 laws on these outcomes.

Importantly, our findings will help to inform how the Federal T21 law (adopted on December 20, 2019) might have impacted youth and young adult tobacco use. Providing direct evidence on the effects of the Federal T21 law is likely to be quite challenging given its coincident timing with the onset of the COVID-19 pandemic, including (1) local school closures and COVID-19 mitigation policies (i.e., mask mandates, stay-at-home orders), as well as (2) local respiratory-related pandemic health shocks (Nowakowski 2020), each of which also likely impacted tobacco consumption. Thus, the policy estimates we obtain are likely to be informative in assessing post-2019 tobacco related consumption changes that could reasonably be attributed to the Federal T21 law.

We document three key findings from our investigation. First, using data from the 2009-2019 Behavioral Risk Factor Surveillance Survey (BRFSS) and a difference-in-differences approach, we find that the adoption of a statewide T-21 law led to a 2-to-4-percentage-point decline in prior 30-day smoking participation among 18-to-20-year-olds. We also detect some evidence that ENDS use among 18-to-20-year-olds fell following T-21 adoption. An examination of (1) pre-treatment trends from event-study analyses (including those generated using stacked DD estimates and never-adopters of T-21 as counterfactuals), (2) falsification tests on young adults ages 21-and-older, and (3) estimates obtained from difference-in-difference-and-differences regressions that control for full state-by-year interactions, suggest that our estimated treatment effects can be causally interpreted.

Second, analyses of the 2009-2019 State Youth Risky Behavior Surveys (YRBS) shows that statewide T-21 adoption is associated with a 3-to-7-percentage-point reduction in cigarette use and a 6-to-12 percentage-point decline in ENDS use among 18-year-old high school students. We also find descriptive evidence consistent with the informal social market (i.e., sources from whom to borrow or bum e-cigarettes) mutes the net tobacco use effects of T-21 laws. In contrast, we find somewhat larger reductions in tobacco use when assessing everyday use, consistent with the hypothesis that reduced access to peers (i.e., 18-year-olds in high school) with legal access to tobacco products reduces youths access to a steady supply needed to support a daily addiction.

Third, our analyses uncover important spillover effects of T-21 laws. We find that statewide T-21 laws reduce cigarette use among 16-to-17-year-olds, a group that relies heavily on informal social sources — including 18-year-old peers — to access tobacco. In addition, we detect modest evidence that T-21 adoption is associated with a small reduction in marijuana use among some teens, consistent with the hypothesis the marijuana and tobacco are complements.

Together, our findings suggest that T-21 laws had important effects on youth and young adult tobacco use, and that informal social sources for tobacco products played an important role, both for (1) insulating some youths (18-year-olds) from the full tobacco consumption effects of T-21 laws, and (2) reducing younger teens’ (16-to-17-year-olds) informal supply of cigarettes. Our lower-bound back-of-the envelope calculation suggests that, under the assumption that our research design reasonably captures how 18-to-20-year-olds in non-T-21 states would have responded to the Federal T-21 law, that approximately 40 percent of the cigarette smoking decline among 18-to-20-year-olds observed in the period from 2019 to 2021 could be explained by the Federal law.