The estimates from Table 3.7 are not statistically different from those of Table 3.2, but they are more precisely estimated in the restricted sample. The magnitude of the effects is largest for weekend and nighttime accidents, again suggesting that increased medical marijuana availability for older adults leads to positive externalities as they treat marijuana and alcohol as substitutes. Overall, access to medical marijuana significantly decreases alcohol and opioid abuse for older adults. The results suggest that, for adults aged 45-64, any costs of increased marijuana use may well be outweighed by a reduction in the substantial health costs associated with heavy consumption of these other substances. This is consistent with surveys of medical marijuana patients, who often report that they use medical marijuana to quit or decrease their use of alcohol and prescription pain medications . However, there is a policy trade off, as growth in the legal market significantly increases traffic fatalities caused by young drivers aged 15-24. These findings do not contradict those of past research, but they highlight the importance of considering the mechanisms by which medical marijuana liberalization generates spillovers to young recreational users. Individuals whose cannabis consumption decision is affected by law passage may have a different likelihood of joint use of marijuana and alcohol or opioids compared to an individual whose consumption changes with increased marijuana availability. By estimating the effects of growth in the legal market and not initial MML enactment alone, I show that spillovers to youths generate additional negative externalities as younger users treat marijuana and alcohol as complements. However, youths and adults differ significantly in their demand relationships for alcohol and cannabis. This is consistent with some past work showing cross-price elasticity estimates between alcohol and marijuana differ depending on the population studied. For instance, Farrelly et al. find higher alcohol prices significantly lower marijuana use among youths aged 12-20, but have no effect on cannabis use for adults aged 21-30.
Additionally, Williams and Mahmoudi find that the complementary relationship between alcohol and cannabis is strongest for those who use both substances simultaneously ,vertical grow rack system and poly substance use is much more common about adolescents and young adults. From the 2013 National Survey of Drug Use and Health, the prevalence of past-month use of both alcohol and marijuana peaks at age 21, when about 20% of individuals report past-month use of both substances. Participation in the use of both alcohol and cannabis declines slowly until sharply dropping to about 5% for individuals aged 35-49 and 3% for individuals aged 50-64 . Simultaneous use of alcohol and marijuana shows a similar pattern. About 11% of 18-20 years-old report using marijuana and alcohol together in the past month, while only 2% of 35-64 years-old report joint use . While this study’s use of aggregate data limits identification of the mechanisms driving these differences, below I outline a few insights from behavioral economics that might explain these findings and will be explored in future study. Evidence from the neuroscience literature suggests that differences between adolescents and adults in risk-taking behavior can be attributed to age differences in the stage of brain development. Development of the brain’s prefrontal cortex between ages 12-25 is associated with lower impulse control, increased sensation-seeking, and limited resistance to peer pressure , which may make adolescents and young adults more likely to jointly consume marijuana with other addictive substances or to drive under the influence of drugs and alcohol. Differences in brain development between adolescents and adults do not necessarily indicate that youths behave irrationally. In fact, experimental research has largely found that youths and adults are similar in their awareness of potential consequences and in their perceptions of the likelihood of facing those consequences. By age 15, individuals have logical reasoning comparable to that of adults in perceiving risk and estimating their vulnerability to it . A number of economic studies have similarly shown that youth substance use responds to price and perceived risk , which is consistent with some degree of rational decision-making.
This suggests that adolescents and adults are similarly able to assess the expected costs and benefits of addictive substance use within a rational framework. However, there is evidence that youths differ from adults in how they value the outcomes of their decisions. When presented with risky situations that have both potential rewards and costs, adolescents are more sensitive than adults to variation in rewards but less sensitive to variation in costs . In experimental studies, the presence of peers significantly increases risk-taking among teenagers, moderately among college-age individuals, and not at all among older adults .Differences in how youths and adults value consequences reflect differences in preferences. Traditionally, economists have dismissed preference-based explanations of human behavior since differences in preferences “explain everything and therefore nothing” . However, this is true only if there is no empirical evidence available to place structure on a model of preference heterogeneity. In the case of risky behavior, adolescence is shown to be a period when less value is placed on self-assessed potential negative consequences compared to the potential gains from experimentation, novelty-seeking, and social acceptance. Models of identity-formation , experimentation , and peer acceptance offer methods for incorporating these preferences in an economic framework. The utility derived from the consumption of alcohol, marijuana, and opioids is likely driven by some latent demand for intoxication. Motivational models of substance use have categorized these demand drivers along two dimensions, positive vs. negative affect regulation and intrinsic vs. extrinsic reasons for use . Crossing these two dimensions results in four classes of motives : positive/intrinsic , positive/extrinsic , negative/intrinsic , negative/extrinsic . A large body of literature has shown clear age patterns in self-reported reasons for using alcohol and cannabis.
Most adolescent users report use related to social motives, some for enhancement reasons, and very few for coping motives . However, youths tend to “age out” of these positive and extrinsic motives for substance use. Individuals who continue alcohol and marijuana use through their mid-20s are significantly less likely to report reasons for use related to having fun with friends, fitting in, and increasing the effects of other drugs; instead, adults are significantly more likely to report use of both alcohol and marijuana to relax, of alcohol to sleep, and of marijuana to decrease the effects of other drugs . Differences in these latent demand properties have been shown to generate heterogeneous substitution behavior. A recent study of college-aged individuals shows that most students use alcohol and cannabis for social reasons, and they tend to treat these substances as complements. However, those students who report using alcohol or cannabis to relieve negative affect are significantly more likely to treat them as substitutes . Increased marijuana availability for youths who consume to enhance positive mood or social interactions may thus have the unintended consequence of increasing alcohol use; for older individuals who consume alcohol to alleviate negative mood, access to marijuana may offer a substitute manner of coping. Potential economic models to incorporate affect in the utility-maximization process include Hermalin and Isen and Loewenstein . Adults and youths also differ in where they access and use alcohol and cannabis. Compared to older individuals, youths under the age of 21 are less likely to consume alcohol at bars or restaurants, and are more likely to drink in others’ homes, parties, outdoors, and in vehicles. They are also significantly less likely to purchase either alcohol or cannabis. Rather, for underage youths , opportunities for alcohol and cannabis access frequently occur in shared social settings. Increased supply of marijuana may make co-use more likely due to shared availability in the social markets where youth consumption occurs. In contrast, adults over age twenty-one have legal access to alcohol consumption in public places such as bars or restaurants. For adults who drink socially at bars and restaurants,grow rack with lights increased marijuana availability may keep drinkers away from these establishments and reduce their social consumption of alcohol.9 Even if alcohol consumption is not reduced, a shift of alcohol use from public places to an individual’s own home could generate significant reductions in alcohol-related traffic accidents. The role of environment may help reconcile differences between this paper’s findings and studies of the minimum legal drinking age that show marijuana and alcohol are substitutes.10 Being of MLDA decreases the total cost of consuming alcohol, but increases the cost of jointly consuming alcohol and cannabis . Figure 3.5 provides evidence in support of this hypothesis, as the prevalence of past-month use of both marijuana and alcohol peaks at age 21, but there is a sharp drop in the share of concurrent users reporting simultaneous consumption. Bernheim and Rangel present an economic model of addiction that explicitly incorporates the role of environment in determining an individual’s consumption decision. Briefly, their model describes an individual who chooses an environment or “lifestyle” activity , and some allocation of resources between an addictive substance and a non-addictive good. The decision-maker enters each period in a “cold” mode and rationally chooses her lifestyle activity. Along with her history of use, this decision determines the probability of encountering cues that trigger a “hot” mode. If triggered, consumption of the addictive good occurs regard- less of whether it is optimal; if not triggered, the decision-maker rationally chooses whether or not to use.
With growth in the legal medical marijuana market, the likelihood of encountering triggers for marijuana consumption increases for both adults and adolescents. However, an important distinction between adolescents and adults is that adolescents face a more limited choice set for environment or lifestyle. Youths also likely face more uncertainty in the probability of entering the hot mode with respect to multiple addictive substances. For example, an adult wanting to engage socially may choose a restaurant environment where he knows the probability of encountering an alcohol trigger is high, but the probability of encountering a cannabis trigger is low. Alternatively, he could choose a cannabis club environment where he knows the probability of encountering a cannabis trigger is high, but the probability of encountering an alcohol trigger is minimal. An adolescent, on the other hand, will often face more uncertainty about the probability of encountering one or more cues, and the “exposure” environments available to youths will often have high probability of triggering hot modes for multiple substances. Overall, access to medical marijuana significantly decreases alcohol and opioid abuse for older adults. This is an especially important finding given new research showing that the mortality reversal among white non-Hispanics aged 45-54 in the last decade is largely attributable to increased deaths from drug and alcohol poisoning and chronic liver diseases . Availability of medical marijuana as a substitute for these more lethal substances thus has the potential to lower mortality and morbidity from these causes. However, there is a policy trade-off, as youths treat marijuana and alcohol as complements. For adolescents and young adults, increased cannabis consumption due to growth in legal marijuana markets has additional negative consequences by increasing alcohol-related poisonings and fatal traffic accidents. For optimal regulation considerations, there needs to be some understanding of how we should trade off these effects. In Table 3.9, I calculate back-of-the-envelope estimates under a number of assumptions to quantify these health consequences. The numbers of lives saved and lost due to medical marijuana market expansion after the Ogden Memo are calculated by comparing actual mortality counts post-Ogden to those predicted under a counter-factual regime in which no registration rate growth had occurred. The counter-factual predictions are determined using results from Chapter 2 and the coefficient estimates from the regressions used to produce Tables 3.2 and 3.4. Overall, Table 3.9 shows that there are substantial benefits to expanding legal marijuana access for older adults, and regulations should be focused on limiting access to adolescents and young adults. These results imply that legal marijuana markets should be segmented, and costs should be increased for young adults and decreased for older adults. One such policy would be to differentially price marijuana by age, subsidizing the use of medical marijuana for older individuals while taxing younger adults. Since differential unit pricing may generate arbitrage opportunities, an optimal policy could consist of two-part pricing, in which medical marijuana purchases are subject to higher taxes but higher registration fees are subsidized for older adults.