Association for Public Policy Analysis & Management

Background:  Since 1996, 23 states plus the District of Columbia have passed medical marijuana laws (MMLs).  In general, MMLs provide limited legal protections from criminal penalties for the sale, possession, and cultivation of marijuana intended for medical purposes. Recent research has found that the implementation of MMLs led to a decrease in traffic fatalities involving alcohol, suggesting potentially positive public health impacts that may be realized from increasing access to marijuana.  However, very little has been done to evaluate the impact of MMLs on cigarette use, which is the leading cause of preventable death in the U.S.

Objective: This study assesses whether MMLs affect cigarette smoking among American adults aged 18 and older.  Specifically, the analysis considers whether particular characteristics of MMLs (i.e. dispensaries, in-home cultivation, and required registration of medical marijuana patients) have differential impacts on the probability of cigarette use, and whether the effects of MMLs differ by age and sex.  Current smoking status is measured as the use of cigarettes on at least one or more days in the past 30 days.

Approach: This study uses the 1994-2010 surveys of the Behavioral Risk Factor Surveillance System (BRFSS), resulting in a pooled sample of over 3.7 million observations.  During this period, 15 states implemented MMLs, creating natural variation that is used to identify the effects of MMLs.  Using a fixed effects ordinary least squares identification strategy (i.e. differences-in-differences), I estimate the impact of MMLs on the probability that an individual is a smoker, while controlling for a number of other demographic, economic, and policy characteristics at both the individual and state levels.  To illustrate the differential impact the specific characteristics of MMLs might have on cigarette use, additional models that control for the presence of dispensaries, the allowance of in-home cultivation, and the implementation of a required medical marijuana registry are also estimated. I then estimate the effect of these MML characteristics on smoking using samples stratified by sex and age. 

Results: Preliminary estimates indicate that states that allow medical marijuana dispensaries significantly increase the probability of cigarette smoking, while states that require registration of medical marijuana patients significantly decrease the probability of smoking.  Among females, the effects of MML characteristics are insignificant after controlling for individual and state-level characteristics.  In contrast, initial estimates indicate that the probability of smoking significantly increased for adult men in states that allowed dispensaries and in-home cultivation by approximately 10%. 

Conclusions: The results of this study suggest that increased access to marijuana via storefront dispensaries and in-home cultivation may result in increased cigarette use, especially among adult males.  These results highlight the need for policymakers to consider the possible negative externalities on public health that might result from increased access to marijuana.