Marijuana Legalization and Interest in Marijuana: Evidence from Google Trends

On November 5, 1996, California passed Proposition 215, making it the first state in the nation where marijuana became an option for patients with debilitating conditions. Since then, twenty-eight states and D.C have introduced medical marijuana legalization, offering protection to patients who use marijuana under the recommendation of a physician. Public support for marijuana legalization continues to rise and, in 2012, recreational marijuana legalization was introduced in Washington where adults over 21 can use marijuana legally. By the end of August 2017, twenty-nine states have some form of medical marijuana legalization and eight have recreational marijuana legalization. Over time, state policy on marijuana has become more fluid and lax, while federal marijuana policy has been increasingly restrictive and punitive.

With legalization of marijuana at the forefront of state and federal policy discussions, considerable debate is focused on determining the likely impact legalization has when marijuana use is no longer subject to criminal prosecution. A large number of studies have looked into the behavioral effects of medical marijuana legalization using government survey data, but rigorous evidence on the effects of recreational marijuana legalization is still thin, owing largely to the lags in the dissemination of survey data.

The rising popularity and utilization of the Internet presents unprecedented opportunities for policy evaluation. In this study, I use data from Google Trends to examine the effects of medical and recreational marijuana legalization on search activity for terms related to marijuana news, marijuana use, and health concerns of marijuana use. An extensive literature in the business and marketing sectors has shown that information-seeking predicts choice and decision, which also provides a direct lens with which to understand the motives behind these decisions. My analyses therefore serve two purposes. First, they provide insights into the likely effects of marijuana legalization laws on marijuana use. This can be especially valuable when adequate post-treatment data on use (e.g., from government surveys) are not yet available, as is currently the case with the recreational laws that were implemented very recently. Second, by investigating multiple categories of marijuana-related search terms, my analyses can shed light on possible mechanisms through which any effect on use occurs. Using difference-in-differences-style econometric methods, I find that recreational marijuana legalization exerts a strong impact on search interest in all dimensions whereas the influence of medical marijuana legalization is limited to the areas of news and laws. Event study regressions suggest that these effects are partly transitory and partly persistent.

As a secondary analysis, I use data from the combined national and state Youth Risky Behavior Survey to investigate the association between state-level marijuana Internet search activity and adolescents’ use of marijuana and other substances. Search activity is significantly associated with marijuana use but not smoking or drinking. These results provide evidence that the increased search activity following recreational marijuana legalization may in fact be indicative of increased use, even though sufficient post-treatment survey data are not yet available to test this hypothesis directly.