Is Marijuana a Gateway Drug

Abstract

Although many cocaine users initiated marijuana prior to cocaine, no formal evidence exists that marijuana consumption causes, or is a gateway to, cocaine consumption. This paper employs a two-stage instrumental variable procedure to estimate a structural effect of past marijuana demand on current cocaine demand using data from the National Longitudinal Survey of Youth. Extensive specification testing verifies that the instruments for marijuana demand, consisting of two state-level marijuana penalty variables, the state beer tax and an indicator of parental alcoholism, have sufficient explanatory power for marijuana demand and have no separate impact on cocaine demand. Results provide strong support for the gateway hypothesis, indicating that marijuana use in 1984 increases the probability of cocaine use in 1988 by 29 percentage points for respondents who have never used cocaine by 1984. The implication is that cocaine use can be more effectively deterred by redirecting some enforcement resources from cocaine to marijuana.