The marijuana use landscape is changing rapidly across North America. Canada became the first major nation to legalize cannabis in 2018.¹ As of year-end 2019, eleven states in the U.S. approved marijuana for recreational use – while only three states completely prohibit the use of medicinal marijuana. Increasing social acceptance with legalization over the past few years has resulted in an upward trend in marijuana use in the U.S. general population and, by extension, in the life insurance buying population. This increase in marijuana use may also be explained by a decrease in the overall stigma of use, prompting more honest answers. In the U.S., 34.8% of people between age 18 to 25 used marijuana in 2018 as well as 13.3% of people older than 25.² Given the increasing trend and growing concerns about excess mortality of heavy users, it is critical for life insurers to effectively assess the risk associated with marijuana use on their blocks of business.

Munich Re’s North American Biometrics team performed a sensitivity analysis to estimate the mortality impact of extreme what-if marijuana use scenarios on its in force block and new business in the U.S.

Data Sources

Clinical Research Laboratories (CRL) published one of the few publicized studies regarding marijuana use  in the U.S. life insurance buying population in 2016. ³  An update to the study included 500,060 applicant urine samples tested for THC^* metabolite at CRL between 1995 and 2015.⁴ This sample included 3,569 deaths. The same update estimated that greater than 4% of tested life applicants between 2017 to April 2019 were positive for THC vis-à-vis the original study with greater than 2% THC positive – confirming the upward trend in marijuana use.

Population studies, public data sources, and the CRL study were all used to develop assumptions for mortality relative risk (RR) and prevalence, the two primary building blocks for any mortality impact estimation exercise.

Relative Mortality Risk of Standalone Marijuana Use

Relative Risk (RR) assumptions based on internal research and Munich Re Medical Director feedback were set as follows: 

The “heavy”^** users listed in Table 1 are most likely to be those who test THC positive in insurance lab testing. The RR estimates in Table 1 are chosen keeping in mind how debits solely for marijuana use would most likely be applied in a production underwriting environment. Additionally, these RR estimates assume recreational marijuana users. Medical marijuana, which has an additional level of complexity depending on the underlying impairment(s), is not in the scope of this exercise. Future updates can be based on a more granular RR structure – such as age or gender breakdowns – as more experience becomes available.

For the marijuana what-if scenarios described in this paper, we assume the RR as described in Table 1 will remain unchanged in the future. As a result, the overall mortality impact will be primarily driven by an increase in the prevalence of marijuana use.

^{* The usual cut-off for a THC positive test is 50 ng/mL}

^{** There is no standard definition of “heavy” users with regard to frequency of use}