PROJECT SUMMARY
The past 20 years have seen significant changes in the US’s legal cannabis landscape alongside marked
changes in tobacco use, the leading cause of preventable mortality. Yet recreational cannabis legalization’s
(RCL’s) effects on tobacco product use remain unclear: few peer-reviewed studies directly test these effects
and quasi-experimental analyses—methods to generate causal estimates in the absence of randomization—
yield mixed findings on whether cannabis is an economic complement or substitute for cigarettes. These
dynamics are crucial: laws increasing cannabis access will decrease use of its substitutes and increase use of
its complements. Moreover, quasi-experimental studies largely ignore RCL’s effects on use of non-cigarette
tobacco products like cigars and electronic nicotine delivery systems (ENDS), as well as potential effects of
state cannabis policy details (e.g., tax rates, formulation restrictions) and local cannabis laws (e.g., local taxes,
home delivery bans). If such effects are not anticipated, policymakers’ expectations of RCL’s costs and
benefits will be incorrect, and both state and local policymakers may miss opportunities to structure cannabis
laws in a manner that better protects their community’s health. To address this, we will compile a cannabis
policy database covering state RCL policy details, local cannabis policies, and retailer density, and match it to
nationally representative, restricted-use data from the National Survey on Drug Use and Health. Aim 1A will
use these data to characterize variation in local access to cannabis retailers, local cannabis policies, and state
RCL policy details likely to impact cannabis access and use within RCL states. Aim 1B will assess these policies’
relationships to cannabis retailer density, perceived ease of cannabis access, and cannabis use, elucidating their
strength as unconfounded proxies for local cannabis retailer density and ease of access, and clarifying whether
the policies’ effects on cannabis use are strong enough to allow instrumental variable analyses of cannabis use’s
effect on tobacco product use. Considering cigarettes, cigars, ENDS, and blunts, Aim 2 will estimate RCL’s direct
effects on adult tobacco product use, test for effect modification from state RCL policy details and local
cannabis laws, and simulate implications for tobacco product use under alternative RCL scenarios (e.g.,
federal RCL, different cannabis tax rates, preempting local bans on retail sales). Aim 3 will conduct parallel
analyses and simulations for 12-20 year-olds. Results will extend the literature on RCL’s effects on tobacco
product use by accounting for a range of tobacco products, and increase its rigor by considering effect
modification due to variation in state policy details and local cannabis laws, separately for underage versus
21+ age-groups. Moreover, simulations of alternative policies’ effects on tobacco and nicotine use will clarify
potential unanticipated costs or benefits of different state and local cannabis policies, helping policymakers
identify policy options to best ensure continued progress towards eliminating tobacco’s toll in their community.