How Medical Marijuana Smoothed the Transition to Marijuana Legalization in the United States

Abstract

Public support for legalizing marijuana use increased from 25% in 1995 to 60% in 2016, rising in lockstep with support for same-sex marriage. Between November 2012 and November 2016, voters in eight states passed ballot initiatives to legalize marijuana sales for nonmedical purposes—covering one-fifth of the US population. These changes are unprecedented but are not independent of the changes in medical marijuana laws that have occurred over the past 20 years. This article suggests five ways in which the passage and implementation of medical marijuana laws smoothed the transition to nonmedical legalization in the United States: (a) They demonstrated the efficacy of using voter initiatives to change marijuana supply laws, (b) enabled the psychological changes needed to destabilize the “war on drugs” policy stasis, (c) generated an evidence base that could be used to downplay concerns about nonmedical legalization, (d) created a visible and active marijuana industry, and (e) revealed that the federal government would allow state and local jurisdictions to generate tax revenue from marijuana.

INTRODUCTION

When the authors first began analyzing drug prohibition and its alternatives, the topic was barely more respectable among academics than, say, JFK assassination conspiracies or the search for Bigfoot. Reuter (1992, p. 15) aptly labeled the legalization debate “a parlor sport for intellectuals.” Now, several decades later, “cannabis” is currently fourteenth in the Google Trends list of trending business topics.¹ A massive new industry is rapidly emerging, a change facilitated by new state laws and hindered by old federal laws.

Together with various collaborators (especially Jonathan Caulkins, Mark Kleiman, Rosalie Pacula, and Peter Reuter), in recent years we have published a series of lengthy analyses of what can and cannot be forecast about the effects of legalizing marijuana (Caulkins et al. 2012, 2015, 2016; Kilmer et al. 2010, 2013; MacCoun 2013; MacCoun & Reuter 1997, 2001; MacCoun et al. 1996), much of which is reviewed in a previous volume of the Annual Review of Law and Social Science (MacCoun & Reuter 2011). Those analyses attempted to draw lessons from various case studies, especially the decriminalization (or more accurately, depenalization) of marijuana in more than a dozen states (Caulkins et al. 2016, MacCoun et al. 2009) and the de facto legalization of marijuana in the Netherlands (MacCoun 2011; MacCoun & Reuter 1997, 2001). They also included exercises in simulation based on data-informed models of the links between marijuana policies, markets, and consumption. We do not attempt to summarize any of that work here, other than to highlight some basic conclusions.

First, a change in drug laws involves inevitable value trade-offs. For a variety of reasons, legalization is likely to reduce the average harm per dose of a psychoactive drug, where we use harm as a catchall term to refer to myriad dimensions of threat to health, safety, justice, and liberty. But for a variety of reasons, it is likely to increase the total number of doses consumed. Whether total harm—roughly, average harm × total doses—rises or falls will depend on the relative strengths of the harm-reduction and use-promotion effects, and on one’s personal weighting of the various components.

Second, our simulation modeling (Caulkins et al. 2015; Kilmer et al. 2010, 2013), using extensive structural and stochastic sensitivity analyses, demonstrated the wide range of possible outcomes, even for some of the most basic outcome indices (prices, revenues, prevalence, and total consumption). Irrespective of one’s philosophical and political views, it is very difficult to defend any predictive stance other than agnosticism (MacCoun & Reuter 2011).

But given the recent rapidity of marijuana policy change, anything written even a few years ago is out of date. Since 2012, eight states have legalized retail sales of recreational marijuana, and the District of Columbia has legalized possession and home cultivation.

Table 1 shows that each of these states legalized medical marijuana first. (All but Maine depenalized marijuana possession before that.) With the exception of Washington, DC, each jurisdiction has adopted some variation on the for-profit alcohol model. Despite that similarity, there are numerous variations in the details of regulation, e.g., in packaging, industry structure and licensing, and local control. All but one (Washington State) also allow some home cultivation. They vary considerably in their tax schemes.

Table 1.

Vanguard states for marijuana law reform

	Medical marijuana enacted	Voters passed recreational marijuana	Allow home cultivation?	Allow for-profit companies to sell?	State taxes (excludes sales and local excise taxes)a
Colorado	2000	2012	Yes	Yes	15% of the average market rate at wholesale and 15% retailb
Washington	1998	2012	No	Yes	37% retail
Alaska	1998	2014	Yes	Yes	$50/oz wholesale
Oregon	1998	2014	Yes	Yes	17% retail
Washington, DC	2010	2014	Yes	No	N/A
California	1996	2016	Yes	Yes	15% retail, $2.75-$9.25/oz wholesale
Maine	1999	2016	Yes	Yes	10% retail
Massachusetts	2012	2016	Yes	Yes	3.75% retail
Nevada	2001	2016	Yes	Yes	15% retail

^aSome of these tax rates are being debated as we go to press and could be different when the retail stores open.

^bColorado’s marijuana tax is commonly characterized as 15%, but it is much more complex (see Caulkins et al. 2015, pp. 79–80; Oglesby 2017).

Thus, there will be plenty to keep social scientists and legal scholars busy in the coming years, and the analytical infrastructure is far more advanced than in earlier eras (e.g., the repeal of alcohol prohibition, the adoption of marijuana prohibition). Indeed, some analysts have already begun to offer empirical evaluations of the change in laws (e.g., Dills et al. 2016, Rocky Mt. High Intensity Drug Traffick. Area 2016, SAM 2015). But in our opinion (also Hall & Lynskey 2016, Subritzky et al. 2016), these attempts are grossly premature.

The legalization of marijuana is not a single event but rather a series of unfolding changes—the vote itself, the gradual establishment of a regulatory system, the emergence of retail sales outlets, and the development of an adequate legal supply of the product. Retail shops have to jump through a host of regulatory hurdles involving, e.g., electricity and water usage and environmental safety. An economic corollary is that prices are less likely to fall in the immediate term than they are in the long run. Indeed, the very term legalization is problematic, because these new state laws are in conflict with the federal Controlled Substances Act (Baude 2015, Chemerinsky et al. 2015, Mikos 2013).

Moreover, existing data-collection systems were not designed to evaluate state-level changes. Most of the major monitoring systems, e.g., for tracking drug treatment admissions and emergency room incidents, provide neither a complete census nor true probability samples. The foremost tool for assessing state-level marijuana use in the general population—the National Survey on Drug Use and Health—publishes only two-year rolling averages at the state level, because the state-specific sample sizes are unreliably small in any given year.

A further complication is that drug data collection has endogenous relationships with drug policy. For example, roughly half of all marijuana treatment admissions are criminal justice referrals, so any changes in treatment rates after legalization might simply reflect the decline in arrests. And legalization may change the stigma associated with reporting one’s drug use in a survey (Kilmer & Pacula 2016).

Finally, there are enormous challenges of causal identification. Marijuana use was already more prevalent than average in the states that legalized. Postaccident toxicology tests can detect marijuana many days after intoxication, and testing procedures may be endogenously related to legalization. And many cities in Colorado and Washington already had visible and accessible medical marijuana dispensaries prior to nonmedical legalization.

But medical marijuana is far more than an analytic nuisance to be controlled for in multivariate analyses; it seems clear that it played an essential role in enabling state legalization, by facilitating the emergence of both industrial and regulatory structures, but more importantly, by enabling the psychological changes needed to destabilize the “war on drugs” policy stasis. Thus, our primary focus in this article is on medical marijuana, not recreational marijuana. At this moment in history, that focus seems apt because one cannot understand the success of the legalization movement without understanding the enormous strides taken by the medical marijuana movement.

EVOLUTION OF THE MEDICAL MARIJUANA MOVEMENT IN THE UNITED STATES

Since 1996 there have been tremendous changes in state laws about the production, distribution, and possession of medical marijuana in the United States.

Medical Marijuana Laws in the United States²

Marijuana has been used for therapeutic purposes for thousands of years in various cultures for various ailments (Grinspoon & Bakalar 1993, O’Shaughnessy 1843). However, in medicine as in other uses, marijuana had been largely replaced with newer substitutes even before it became prohibited. Support for medical marijuana in the United States increased in the late 1980s and 1990s, partially in response to public empathy for patients battling cancer and AIDS. There was anecdotal evidence that marijuana improved appetite while reducing nausea and pain for patients with AIDS, for whom loss of appetite was a life-threatening condition (Treaster 1993). Synthetic tetrahydrocannabinol (THC) in pill form (Marinol) was approved for these purposes, but swallowing a pill is problematic if nausea and vomiting are the issue. Moreover, pure THC was poorly tolerated by many patients, who complained of anxiety and unpleasant overintoxication.

In November 1996, California became the first state to exempt medical marijuana patients from prohibition when Proposition 215 passed. Prop 215 provided that

seriously ill Californians have the right to obtain and use marijuana for medical purposes where that medical use is deemed appropriate and has been recommended by a physician who has determined that the person’s health would benefit from the use of marijuana in the treatment of cancer, anorexia, AIDS, chronic pain, spasticity, glaucoma, arthritis, migraine, or any other illness for which marijuana provides relief.

This last clause, combined with the lack of any regulatory control over physician behavior or providers, made it extremely easy for anyone 18 and older to obtain a recommendation. For example, one study examining 1,655 patients seeking a physician’s recommendation for medical marijuana in California found that the most common conditions reported were those that were hard to verify: chronic pain, mental health conditions (primarily anxiety and depression), and insomnia (Nunberg et al. 2011).³

California’s approach was copied by some other states, whereas others created far more restrictive medical marijuana programs. Caulkins et al. (2016, p. 204) grossly generalize that “medical marijuana regimes created by state legislatures east of the Mississippi generally appear to be good-faith efforts to provide compassionate access to people with well-defined medical conditions, while the regimes originally created by voter propositions farther west are extremely permissive and easy to manipulate.” Twenty-eight states have now enacted medical marijuana laws—roughly half through ballot propositions. More than a dozen other states have passed laws to make cannabidiol (CBD) oils legal to those suffering from a very short list of conditions (primarily youth suffering from intractable seizures), but these CBD-only laws are in their infancy, and some states passed laws without providing a supply mechanism, thus making them almost impossible to implement.

A Trojan Horse?

Critics have been suspicious of the motivations behind medical marijuana for decades. Clinton’s Director of the White House Office of National Drug Control Policy, General Barry McCaffrey, argued that “[m]edical marijuana is a stalking-horse for legalization. This is not about compassion. This is about legalizing dangerous drugs” (Levine 1998). A New Republic account of the California Prop 215 campaign quoted “pinstriped professional, Bill Zimmerman,” the campaign manager, bemoaning the behavior of activist Dennis Peron: “He was pictured on election night smoking a joint and saying, ‘Let’s all get stoned and watch election night returns,’” Zimmerman recalls. “That kind of behavior supports the opponents’ view that we are a stalking horse for legalization … He could ruin it for the truly sick” (Rosin 1997).

The stalking horse metaphor is less common than another equine analogy, the Trojan horse. In their comprehensive 1999 review of medical marijuana, the Institute of Medicine noted,

The argument against the medical use of marijuana presented most often to the IOM study team was that ‘the medical marijuana movement is a Trojan horse’; that is, it is a deceptive tactic used by advocates of marijuana decriminalization who would exploit the public’s sympathy for seriously ill patients.

(Joy et al. 1999, p. 18)

In a 2009 Washington Post column, Charles Lane (2009) argued that “‘Medical marijuana’ is obviously a Trojan horse for legalization of pot as a recreational drug. In a democracy, people should pursue their policy objectives openly, not under false pretenses.” Passkik & Tickoo (2010) argued that the gap between support for medical marijuana and support for full legalization “provides just enough wiggle room for politicians lacking in backbone to squeeze through and try to make everyone happy. The problem is that they are using physicians as a Trojan horse to do it and pain physicians should be up in arms.”

The Trojan horse, of course, is a memorable but fictional tale in The Odyssey, attributed to the memorable but also possibly fictional author Homer (d’Angour 2014). As the story goes, the Greeks constructed a large wooden horse outside the gates of Troy, hid soldiers inside it, and then allowed the Trojans to believe it was left as a sacrifice to Athena by the departing invaders. The hapless Trojans brought it into the gates, and the rest is history—or rather, historical fiction. As an idiom, it refers to any situation in which someone is duped into letting an enemy past their defenses. The dupe is at least foolish and possibly greedy as well. The idiom stalking horse is sometimes used interchangeably, but actually unrelated; it refers to a hunting tactic (hiding behind a slow-walking horse). Thus, both phrases connote an act of deception.

Is medical marijuana in fact a deception? Many commentators clearly think so. General McCaffrey “derided the propositions as ‘hoax referendums,’ and insisted that voters had been ‘duped’ by deceitful ad campaigns whose real intent was to legalize drugs” (Fine 1997). A 2012 New York Times article (Onishi 2012) was titled “Marijuana only for the sick? A farce, some Angelenos say.” The article described how, “on the boardwalk of Venice Beach, pitchmen dressed all in marijuana green approach passers-by with offers of a $35, 10-minute evaluation for a medical marijuana recommendation for everything from cancer to appetite loss.” Onishi describes one civic leader’s comment upon watching a young man briskly leaving a dispensary: “I’m going to go out on a limb, but that’s not a cancer patient.” Passkik & Tickoo (2010) decry the “sudden epidemic of social phobia among college students and such; coached patients faking a variety of maladies to gain access to marijuana using a physician as the sympathetic gate keeper.”

As to whether the claim of deception is valid, there are really three different questions. First, do some physicians and their clients use the system deceptively? There seems little doubt that the answer is yes (Drake 2014, Sullum 2011), although it is difficult to estimate empirically (Nunberg et al. 2011, Reiman 2007, Reinarman et al. 2011). Former Congressman Patrick Kennedy has asserted that 80% of medical marijuana patients do not meet qualifying conditions (McKinney 2013); he reached this figure largely by discounting referrals for severe pain, which is in fact a qualifying condition and is statistically far more common than cancer, AIDS, or glaucoma. What we do not know is what fraction of referrals for pain and other difficult-to-verify presenting conditions are actually sincere. O’Brien (2013) interviewed 40 undergraduate medical marijuana cardholders. Surprisingly, the paper offers no attempt to assess whether the students had actually sought marijuana for accepted medical uses, focusing instead almost exclusively on the advantages of buying in a licit rather than an illicit market. O’Brien (2013, p. 436) does note that “students were typically not using medical marijuana for traditionally accepted conditions such as glaucoma, cancer, AIDS, or Multiple Sclerosis, but they denied that using marijuana for other reasons was wrong, and explained how it was appropriate.” Stogner et al. (2014) surveyed 2,349 university students and found that approximately 4% reported having intentionally deceived a physician to obtain a prescription drug; they did not ask specifically about medical marijuana but report that deception was more common among those who had used marijuana (6%).

Second, did activists act deceptively? Activist Dennis Peron infamously claimed that “I believe all marijuana use is medical—except for kids” (Rosin 1997). Prop 215’s campaign manager acknowledged that “some people supporting medical marijuana initiatives are without question using it as an attempt to legalize marijuana. Other people are supporting marijuana policy changes out of a genuine concern for patients. It’s a free country” (Fine 1997). Ethan Nadelmann, Executive Director of the Drug Policy Alliance, has always been quite open about how the medical marijuana movement fit into a larger reform agenda. “We started polling and found a couple of issues where the public said the drug war’s gone too far” (Dickinson 2013). He recognized that California’s medical marijuana “could change the public face of the marijuana consumer” from a “17-year-old high school dropout with dreadlocks” to adult chemotherapy and AIDS patients (Dickinson 2013).

Third, are citizens deceived by medical marijuana? When the Pew Research Center (2010) asked “Does legalizing medical marijuana make it easier for people to get marijuana even if they don’t have a real medical need?” in a 2010 poll, 46% said yes and 48% said no. When Fox News (2015) asked, “Do you think most people who smoke medical marijuana truly need it for medical purposes or just want to smoke marijuana?” in a 2013 poll, 30% said “truly need it,” 47% said “just want to smoke marijuana,” and 12% felt that it “depends on the person.” Finally, in a 2014 poll (Dutton et al. 2014), CBS News asked, “Do you think most of the marijuana that is being purchased in this country through state authorized medical marijuana programs is being used to alleviate suffering from serious medical illnesses, or do you think most of it is being used for other reasons?” “Serious illness” was selected by 34%, and 56% selected “other reasons.” The pollsters noted, “Nevertheless, 77% of Americans who think most of the medical marijuana purchased is being used for other purposes still think doctors should be allowed to prescribe medical marijuana.” These figures hardly paint a picture of a naïve and duped populace.

Public Opinion on Marijuana

Figure 1 shows trends in public support for legalized medical marijuana, legalized recreational marijuana, and the legalization of same-sex marriage. The first pattern to note is that medical marijuana has enjoyed strong and fairly stable support in the 70% to 85% range for more than 15 years. Further, there was a dramatic rise in support for marijuana legalization, which roughly three-quarters of Americans opposed in 1995 (and for many years before that) but which nearly six in ten supported by 2015. Finally, support for marijuana legalization has risen in lockstep with support for same-sex marriage, a phenomenon we return to below.

Figure 1.

Trends in public support for legalizing medical marijuana, recreational marijuana, and same-sex marriage (Fox News 2015; Gallup Poll 2016a,b).

What appeared in an earlier time to be an age effect—a generation gap—in attitudes toward marijuana legalization now appears to be a mix of period and cohort effects. With respect to period, analyses of the General Social Survey by Nielsen (2010) suggest a curvilinear trend, with greater hostility toward legalization during the Reagan-Bush era than in either the Ford and Carter administrations or the Clinton, G.W. Bush, and Obama administrations. With respect to cohort, Nielson found that the Baby Boom marked a turning point, with subsequent cohorts increasingly supportive of legalizing the drug. Galston & Dionne (2013) point out that the only cohorts still opposed to legalization are those over age 65. This pattern leads many to conclude that full federal legalization is a historical inevitability, an inference we question at the end of this essay.

That views toward legalizing marijuana and same-sex marriage have risen in tandem seems unlikely to be a pure coincidence. Twenge et al. (2015) used hierarchical linear modeling techniques to analyze age, period, and cohort effects in various indicators of social tolerance in the General Social Survey in the years from 1972 to 2012. With respect to period, “tolerance for those espousing controversial views was markedly higher among American adults in the early 2000s and 2010s compared with the 1970s and 1980s, showing a fairly linear increase over time” (Twenge et al. 2015, p. 386). They show that this period effect is correlated with a similar trend for individualism (relative to more collectivist attitudes). With respect to cohort, Baby Boomers show greater tolerance than either earlier or later cohorts. After controlling for these period and cohort effects, Twenge and colleagues found that tolerance declines with age. Unfortunately, it is difficult to directly compare this analysis with the Nielsen study; although both use the same data source, they focus on different attitude measures and time spans.

It is perhaps more accurate to say we are seeing more of a shrinking of opposition to legalization than a growth in enthusiasm for legal marijuana. Galston & Dionne (2013, p. 3) note that “[a]ttitudes toward legalization are marked by ambivalence, especially on the conservative side. Many of those who favor legalization do so despite believing that marijuana is harmful or reporting that they feel uncomfortable with its use.” They suggest that conservative disenchantment with marijuana prohibition may reflect views on states’ rights and on the ineffectiveness of government regulation more than anything to do with the drug itself.

HOW HAVE MEDICAL MARIJUANA LAWS INFLUENCED USE?

Some of the consequences of using marijuana—both good and bad—have been documented in several recent reviews (e.g., Caulkins et al. 2015, Hall 2015, Hill 2015, Volkow et al. 2014, Whiting et al. 2015). With respect to the medical benefits, a detailed review is beyond the scope of this article, but the most authoritative recent source is a meta-analysis (Whiting et al. 2015) examining the effects for various medical conditions of “smoked THC” and other cannabinoid drugs. The review of 79 controlled trials concluded the following:

There was moderate-quality evidence to suggest that cannabinoids may be beneficial for the treatment of chronic neuropathic or cancer pain (smoked THC and nabiximols) and spasticity due to MS (nabiximols, nabilone, THC/CBD capsules, and dronabinol). There was low-quality evidence suggesting that cannabinoids were associated with improvements in nausea and vomiting due to chemotherapy (dronabinol and nabiximols), weight gain in HIV (dronabinol), sleep disorders (nabilone, nabiximols), and Tourette syndrome (THC capsules); and very low-quality evidence for an improvement in anxiety as assessed by a public speaking test (cannabidiol). There was low-quality evidence for no effect on psychosis (cannabidiol) and very low-level evidence for no effect on depression (nabiximols).

(Whiting et al. 2015, p. 2467)

But as noted by Caulkins et al. (2016, p. 83),

absence of evidence is not evidence of absence of effectiveness. There have been few clinical trials involving smoked or vaped marijuana, largely because of prohibition, medical research culture, and hurdles to conducting these studies. Also, even if a study shows that cannabinoids do produce medical benefits, it may not be the best treatment available.

There has been very little research about possible nonmedical benefits of consumption (Caulkins et al. 2016), although in addition to pleasure or fun, users often cite enhanced creativity and aesthetic appreciation, relaxation, and self-insight. In contrast, there has been considerable research on the physical, psychological, and behavioral harms associated with using marijuana, mostly using longitudinal or quasi-experimental designs that raise questions about whether the identified correlation is causal or spurious (see Caulkins et al. 2016). Again, these issues are reviewed in detail elsewhere and are not our focus. Of more immediate relevance is the question of whether and how medical marijuana laws might influence recreational marijuana use. We address this below from both theoretical and empirical perspectives.

Medical Marijuana Laws and Recreational Use: Theory

Medical marijuana laws could influence the use of marijuana for medical and nonmedical purposes in myriad ways. Obviously, creating legal protections for those receiving medical recommendations and those supplying them will make marijuana a more attractive therapeutic agent, especially for those who have had limited success with more traditional treatments. Here we highlight several mechanisms by which easing prohibition and allowing medical marijuana could influence consumption, building on a framework developed by MacCoun (1993).

Legal risk.

The risk of arrest increases the expected cost of obtaining and using a prohibited substance (Reuter & Kleiman 1986). If a medical marijuana law reduces this cost by preventing those with a recommendation from being arrested or gives them an affirmative defense at trial for possession, this could make consumption more attractive to some individuals (of course, one would also need to factor in the costs associated with obtaining a medical recommendation).

Stigma.

A change in the legal risk could also affect the stigma associated with the activity. For example, users’ family and friends may be more accepting of marijuana use and home production if they know the state no longer prohibits these activities for patients. There may also be less of a stigma if medicalization reduces the belief that consumption is harmful.

Forbidden fruit.

But the inverse may also be true. If some people find marijuana more attractive to use because it is illegal and perceived as dangerous, medicalization could be seen as mainstreaming the activity and, thus, making it less attractive for some individuals.

Availability.

As noted above, there is a lot of variation in how medical marijuana laws are implemented across the United States. Most states allow medical marijuana dispensaries (stores that sell only medical marijuana), but the ratio of dispensaries to patients varies dramatically. This can reduce the search costs for obtaining marijuana and can also increase access to a variety of marijuana products. One can also think about availability in terms of the number of medical users who may be more likely to share their supply (purchased and/or grown) or make a straw purchase for those who do not have medical recommendations.

Promotion.

Not only does allowing dispensaries increase the availability of marijuana, the physical stores can also serve as advertisements. And in places where there is increased competition, there are also incentives to market in alternative news weeklies and other outlets. One recent study of junior high students in greater Los Angeles—an area saturated with marijuana dispensaries—found that the level of exposure to medical marijuana advertising was associated with a greater propensity to use marijuana a year later (D’Amico et al. 2015). Of course, anti-marijuana messages can also be promoted, but we are not aware of any that are directly linked to the implementation of medical marijuana laws.

Price.

Price is related to both availability and legal risk, and research suggests that the probability of using marijuana increases when the price decreases (Davis et al. 2016, Gallet 2014, Pacula & Lundberg 2014).⁴ Because a large portion of what users spend on illegal drugs is used to compensate the dealer and everyone else along the supply chain for their risk of arrest, incarceration, and violence (Caulkins & Reuter 2010, Kilmer et al. 2010, Reuter & Kleiman 1986), reducing this risk should lead to lower production and distribution costs. Whether or not this gets passed on at the retail level depends on the supply architecture (e.g., government-run monopoly), regulatory decisions, and tax policies (Caulkins et al. 2015). Currently, most medical marijuana states levy only the state sales tax, but there are some noteworthy exceptions.

A change in marijuana use may also influence the use of other substances like alcohol, opioids, and nicotine. Although there is a lot of evidence suggesting that marijuana and tobacco are economic complements, the evidence about the relationship between marijuana and alcohol consumption is all over the map (Caulkins et al. 2015, appendix A; Guttmannova et al. 2016). The evidence about the cross-price effects of marijuana on opioids is nascent, but much of the emerging evidence comes from studies of medical marijuana laws, which are discussed in the next section.

Medical Marijuana Laws and Recreational Use: Evidence

Table 2 summarizes results from major published analyses of the effects of medical marijuana laws on various outcomes. We exclude strictly cross-sectional studies and report only outcomes for which there are at least two eligible studies. The studies in the table differ somewhat in their years under study and their data sources. They differ considerably in their analytic approaches, covariates, use of fixed effects, and causal identification strategies.

Table 2.

Peer-reviewed longitudinal analyses of the association between medical marijuana laws and various outcomes

Outcome variable	Source	Association with medical marijuana
Marijuana prevalence among youth	Khatapoush & Hallfors (2004)	Not statistically significant
	Harper et al. (2012)	Negative, but statistical significance depends on specification
	Wall et al. (2012)	Not statistically significant once Montana or Vermont is dropped
	Lynne-Landsman et al. (2013)	Not statistically significant
	Schuermeyer et al. (2014)	Not statistically significant for commercialization on prevalence; marginally significant positive effect on abuse/dependence
	Choo et al. (2014)	Not statistically significant
	Anderson et al. (2015)	Not statistically significant
	Wen et al. (2015)	Positive for initiation for those 12–20; positive for dispensaries and past-month use
	Pacula et al. (2015)	Not statistically significant for dispensaries, mixed for other aspects of MML
	Hasin et al. (2015); Keyes et al. (2016)	Negative for eighth-graders; no statistically significant effect for other grades
	Martins et al. (2016)	Not statistically significant
	Stolzenberg et al. (2016)	Positive
	Wall et al. (2016)	Not statistically significant
	Johnson et al. (2017)	Not statistically significant for MML, but positive with higher possession limits and a voluntary registration
Marijuana prevalence among adults	Gorman & Huber (2007)	Not statistically significant
	Harper et al. (2012)	Not statistically significant
	Schuermeyer et al. (2014)	Not statistically significant
	Pacula et al. (2015)	Positive association with dispensaries (all ages; no test for adults only)
	Wen et al. (2015)	Positive for adults 21+
	Martins et al. (2016)	Positive for adults 26+
Alcohol use	Anderson et al. (2013)	Negative for adults
	Sabia et al. (2015)	Negative for binge drinking for adults
	Wen et al. (2015)	Not statistically significant for those 12–20; positive for frequency of binge drinking among those ages 21+
Opioids	Bachhuber et al. (2014)	Negative for overdose deaths
	Wen et al. (2015)	Not statistically significant for prescription painkiller misuse
	Bradford & Bradford (2016)	Negative for pain medication prescriptions in Medicare D population
	Kim et al. (2016)	No statistically significant effect for drivers in fatal crashes testing positive for opioids; negative for those aged 21–40
	Bradford & Bradford (2017)	Negative for pain medication prescriptions among Medicaid enrollees
Suicide rates	Anderson et al. (2014)	Negative
	Rylander et al. (2014)	Negative
	Grucza et al. (2015)	Not statistically significant

Abbreviation: MML, medical marijuana law (dichotomously coded).

Much of the early research on medical marijuana laws involved state-year panel analyses that used a binary variable to define whether a state had enacted a medical marijuana law. This is problematic because there is tremendous variation in how these laws are implemented across states (Pacula et al. 2015). Other studies focus on the provisions of these laws (e.g., whether they allow dispensaries, whether pain is an eligible condition), which captures some of the heterogeneity in these laws and how they are being implemented. There is also growing interest in studies that use the number of individuals with medical recommendations in a state to proxy the supply of medical marijuana (e.g., Smart 2015).

Table 2 suggests a bewildering mix of outcomes. A forthcoming review of the medical marijuana literature by Pacula & Smart (2017) finds the bulk of the research suggests these laws (or aspects of these laws) are positively associated with adult prevalence and not associated with youth prevalence:

While findings tend to be mixed when the literature is looked at as a whole, some consistent themes seem to emerge when the literature is instead considered with an eye toward differences between policies and populations. For example, studies that examine medical marijuana markets in a manner that is attentive to the development of these markets, either through measures of the presence of active dispensaries or the size of the market, seem to consistently show a positive correlation of these policies with use among high-risk users (arrestees, those in need of treatment, and polysubstance users). Similarly, many studies have shown a positive association with adult use of marijuana while most have found no association with youth prevalence or frequency of use in general school populations.

We agree with this assessment, but we would add the following caveat. Few of the studies in Table 2 link the observed association between medical marijuana and an outcome of interest to any intervening causal mechanism of the sort we discussed in our theoretical analysis above. More will be learned about these relationships when individual-level survey data with geographic identifiers and improved questions about marijuana consumption (including concurrent use with other substances) are analyzed.

HOW HAVE MEDICAL MARIJUANA LAWS INFLUENCED PUBLIC POLICY ON NONMEDICAL MARIJUANA?

In this section we highlight five potential ways in which the passage and implementation of medical marijuana laws may have smoothed the transition to nonmedical legalization by

1. Demonstrating the efficacy of using voter initiatives to change marijuana supply laws

2. Enabling the psychological changes needed to destabilize the “war on drugs” policy stasis

3. Generating an evidence base that could be used to downplay concerns about nonmedical legalization

4. Creating a visible and active marijuana industry

5. Revealing that the federal government would allow state and local jurisdictions to generate tax revenue from marijuana

This list is not exhaustive, and our goal is not to determine which factor mattered more in the success of nonmedical legalization. Rather, we ultimately want to convince readers that one cannot fully explain the legalization phenomenon in the United States without understanding the significant changes in medical marijuana laws that preceded these efforts.

Demonstrating the Efficacy of Using Voter Initiatives to Change Marijuana Supply Laws

There have been hundreds of marijuana-related initiatives on state and local ballots since the 1970s (Ballotpedia 2016, Crombie 2016, Staggs 2016). Although discussions about marijuana legalization often begin with California in 2010, California first voted on legalization in 1972 (it lost 33.5% to 66.5%). Many of the early votes focused on reducing the penalties for possession of small amounts or making marijuana offenses the lowest priority for law enforcement officials. More recently, some local jurisdictions have voted on whether to levy additional taxes on marijuana or allow retail sales.

Of the eight states that legalized medical marijuana from 1996 to 2000, all but one did it via the ballot box.⁵ What was different about these initiatives and those that eventually passed elsewhere is that they usually did more than address possession of marijuana; they often made allowances for supply. Of course, some of these initiatives were a bit vague about supply and required additional guidance. For example, California’s Proposition 215, which led the medical marijuana movement in 1996, stated that existing laws “shall not apply to a patient, or to the patient’s primary caregiver, who possesses or cultivates marijuana for the personal medical purposes of the patient.” California legislation passed in 2003 created a framework for the medical collectives, and in 2015 a series of laws were passed to create a state-regulated commercial market.

These ballot initiatives also highlighted how easy it was for the proponents of medical marijuana laws to outspend their opponents. Many of the medical and nonmedical initiatives have been supported by billionaires, such as George Soros and the late Peter Lewis, and there has been very little money raised for the opposition. Much of the opposition funding typically raised for these campaigns comes from law enforcement groups. One notable exception was the campaign opposed to legalizing medical marijuana in Florida in 2014. Casino magnate Sheldon Adelson donated over $5 million to the opposition—nearly matching the amount spent by the proponents—and the initiative lost with 58% of the vote (it needed 60% to pass because it was going to be a constitutional amendment). We do not know the extent to which exit polls and focus groups about these early initiatives shaped future medical and nonmedical efforts, but it is not hard to imagine that the campaigns learned from each other—especially when they were supported by some of the same donors.

Enabling the Psychological Changes Needed to Destabilize the “War on Drugs” Policy Stasis

The political success of the medical marijuana movement may also be attributable to the discomfort of ambivalence. In a sense, we think that a fig leaf is a better metaphor for medical marijuana than a Trojan horse. Table 3 draws on two influential taxonomies of values (Rokeach 1973, Schwartz 1992) to suggest the tensions inherent in the marijuana legalization debate. One column shows core values that many people might associate with marijuana prohibition; the other shows values that might be associated with marijuana legalization.

Table 3.

Two influential taxonomies of values

Source	Values that might be consistent with prohibition	Values that might be consistent with legalization
Rokeach (1973)	Family	Freedom
	Security	Pleasure
	Self-control	Honesty
	Obedience	Forgiveness
	Responsibility
Schwartz (1992)	Conformity	Stimulation
	Tradition	Hedonism
	Security	Self-direction

Tetlock’s value pluralism model (e.g., Tetlock et al. 1996) argues that because conflicts between core values are aversive, people use a variety of methods to resolve them, on a continuum from denial and bolstering (the simplest), to lexicographic ranking, to conceptual reasoning strategies requiring considerable integrative complexity. Tetlock and colleagues argue that social content and context moderate the choice of resolution strategies: “When resource scarcity compels people to make decisions that violate the normative ban on taboo trade-offs, people will make massive impression management efforts to conceal, obfuscate or redefine what they are doing.” Observers, who do not have to make the hard decisions, “will respond indignantly” (p. 38).

The marijuana legalization debate can be interpreted in this light, but there are some complications. In this domain, legalizers have been the observers and prohibitionists controlled policy, yet it was the former who embraced trade-offs and the latter who expressed indignation. An analysis of several decades of op-ed essays in the New York Times (MacCoun & Reuter 2001) shows that prohibition advocates display relatively low integrative complexity (essentially, “drugs are harmful and so they should be illegal”), and legalization advocates display relatively high levels of integrative complexity (acknowledging the need to manage trade-offs between health and safety and personal freedom).

A related theoretical account, focusing more on specific goals than on abstract values, is the goal-conflict model proposed by Robbennolt et al. (2003). They offer four principles by which decision makers balance the pursuit of conflicting goals:

First, the principle of equifinality holds that some goals may be alternately satisfied through multiple pathways …. Second, the principle of best fit holds that pathways may sometimes fulfill some goals better than others …. Third, the principle of multifinality holds that a particular pathway may accomplish multiple goals simultaneously. Some of the decision makers’ objectives may be consistent with each other and may be achieved concurrently …. Finally, the principle of goal incompatibility holds that some objectives will inevitably conflict and, thus, be difficult or impossible to satisfy concurrently.

(Robbennolt et al. 2003, p. 1128)

Robbennolt et al. (2003) suggest that these principles are implemented in a type of connectionist model called a constraint-satisfaction network. A constraint-satisfaction network consists of a set of cognitive elements connected by positively or negatively weighted interconnections, which are then adjusted until the best-fitting compromise among constraints is found (Read et al. 1997, Simon 2004, Thagard & Verbeurght 1998). “The central feature of constraint satisfaction mechanisms is that the mental model will reconfigure itself until the constraints settle at a point of maximal coherence” (Simon 2004, p. 522).

Medical marijuana provides a potentially effective psychological tool for resolving the conflicting values and goals involved in deliberations about marijuana regulation. From the standpoint of Tetlock’s model, medical marijuana provides a way to be soft on marijuana users without endorsing either recreational use or recreational retail sales. We can also interpret this in terms of Schwartz’s (1992) model of basic human values, which suggests core values can be arranged in a circle. Marijuana politics triggers conflict between hedonism, stimulation, and self-direction (values in the 270–360° sector) and their polar opposites, conformity, tradition, and security (values in the 60–180° sector). Medical marijuana provides some compromise by triggering benevolence (in the 45–60° sector).

To say that medical marijuana eases the ambivalence is not to say that it eliminates it. It is interesting that the adoption of Prop 215 did not appear to bring about any immediate change in Californians’ attitudes toward full legalization (Khatapoush & Hallfors 2004). But it weakened the presumption that prohibition was the default policy response.

Generating an Evidence Base that Could Be Used to Downplay Concerns About Nonmedical Legalization

There has been a notable increase in the number of published studies attempting to understand how changes in medical marijuana laws affect public health and public safety. The vast majority of these studies use a differences-in-differences approach, which limits the ability of authors to make strong causal claims about the policy changes. There now seem to be new studies published on this topic on an almost-monthly basis, and there is no guarantee that the associations identified in the short run will persist.

That said, the results of aggregate-level medical marijuana studies suggest the sky has not fallen. Although there appears to be a positive association with adult cannabis prevalence, almost all of the peer-reviewed research published to date suggests that these laws are not associated with an increase in past-month marijuana prevalence for youth (Pacula & Smart 2017). Although the policy changes may have differential effects on prevalence and the amount of marijuana consumed, existing surveys rarely ask about the latter (which matters more when thinking about public health). The distinction between prevalence and quantity consumed is largely confined to researchers, leaving certain media outlets and proponents of marijuana legalization to argue that the available evidence suggests that liberalizing marijuana laws and increasing access to marijuana do not have any noticeable effect on youth.

The available research also suggests there may be some benefits in terms of a reduction in harms related to alcohol and opioids. The studies with findings that are consistent with an alcohol substitution (e.g., Anderson et al. 2013) have been criticized for using a binary measure of medical marijuana laws (Pacula & Sevigny 2014). As for opioids, most peer-reviewed studies suggest increased availability of medical marijuana is negatively associated with opioid-related outcomes⁶ such as opioid overdose deaths at the state level (Bachhuber et al. 2014), prescriptions for painkillers among Medicare Part D enrollees (Bradford & Bradford 2016), prescriptions for pain killers among Medicaid enrollees (Bradford & Bradford 2017), and opioid positivity among 21- to 40-year-old fatally injured drivers (Kim et al. 2016).⁷ That said, some of the peer-reviewed studies have been criticized for overplaying the causal mechanisms driving these results, which is very difficult to determine when working with aggregate-level data (Caputi & Humphreys 2016, Finney et al. 2015).

Methodological issues aside, those pushing for marijuana legalization can now refer to several peer-reviewed studies from some of the best journals in the world to argue that increasing access to medical marijuana does not increase marijuana use among kids and may yield significant public health benefits (aside from increasing access to medicine). For example, the lobbying arm of the Drug Policy Alliance (Drug Policy Action) produced a factsheet about California’s Proposition 64 titled “It Protects Youth,” and noted that

Youth Use of Marijuana Is Not Expected to Increase With the Passage of Prop. 64. Four states and Washington, D.C. have legalized the adult use of marijuana and 26 states have legalized the medical use of marijuana. Research shows that the wider availability of marijuana in states that have legalized medical use and adult use has not led to increased rates of use among teens.

(Drug Policy Action 2016, emphasis added)

In another example, the “Yes on 4” campaign in Massachusetts included a “marijuana and opioids” section on their webpage, with the first paragraph noting that

Recent scientific research provides significant evidence that affordable, legal access to marijuana can be an important tool in addressing the crisis of opioid addiction and overdose. Contrary to opponents of Question 4 who claim that marijuana use is a “gateway” to opioid use, multiple studies suggest that legal marijuana may be an effective substitute for opioid medications for patients who struggle with chronic, debilitating health conditions.

(Yes on 4 2016, emphasis added)

Whether these arguments were persuasive to some voters remains to be seen. The lack of hard evidence linking medical marijuana laws to an increase in opioid overdoses may also have been important to some voters who were previously concerned about marijuana being a gateway drug.

Creating a Visible and Active Marijuana Industry

Although marijuana products used to be sold in American pharmacies in the early twentieth century (Gieringer 2012), few voting on marijuana legalization have personal experience with that era. The proliferation of medical dispensaries in the 2000s introduced the public to the idea of stand-alone stores selling marijuana products (and not just to those residing in states that allow medicinal marijuana, as dispensaries are regularly featured in media stories across the country). Subsequent advertisements in alternative weekly newspapers and occasionally on billboards also exposed voters to this new quasi-legal industry.

Of course, not everyone is thrilled with the establishment of medical dispensaries and the related advertising, and there is tremendous variation in what states allow and how they are regulated (Pacula et al. 2015). Further, even in states that allow dispensaries, some local jurisdictions have decided to prohibit them. But for those who see these dispensaries on a regular basis, they not only have a sense of what legalization may look like but also may be desensitized to the idea of retail marijuana. In fact, some may conclude that allowing any adult to enter those stores instead of only those who have recommendations—which in some states are very easy to obtain (e.g., High Times 2016)—may not be a big change.

Medical marijuana laws also created an industry looking to expand its market beyond medical patients. Indeed, it is hard to imagine that some of those developing, e.g., brands, new methods of ingestion, and improved production methods did not have their sights on a larger market. Although those in the advocacy community criticized the industry for not donating more to the campaigns (e.g., see quotes from Ethan Nadelmann in Freedlander 2016), some with a financial interest did make contributions. Note that some in the medical marijuana industry opposed recreational legalization, as some worried about having to comply with onerous regulations and eventually getting pushed out by larger players. Indeed, the largest growers association in California stayed neutral on Proposition 64 (McGreevy 2016).

Revealing that the Federal Government Would Allow State and Local Jurisdictions to Generate Tax Revenue from Marijuana

The visibility of the medical industry makes it easier for voters not only to envision what the legal industry may look like but also to understand how taxes could be collected (e.g., at the retail level). But with marijuana being illegal at the federal level, there have always been questions about what federal agencies would tolerate at the state and local levels. Although Drug Enforcement Administration raids of medical dispensaries and medical grows were not uncommon during the Clinton and G.W. Bush administrations, the market was still quite visible, especially in California. In 2009, a memo released from Deputy Attorney General David Ogden indicated,

The prosecution of significant traffickers of illegal drugs, including marijuana, and the disruption of illegal drug manufacturing and trafficking networks continues to be a core priority in the Department’s efforts against narcotics and dangerous drugs, and the Department’s investigative and prosecutorial resources should be directed toward these objectives. As a general matter, pursuit of these priorities should not focus federal resources in your States on individuals whose actions are in clear and unambiguous compliance with existing state laws providing for the medical use of marijuana.

This was perceived by some to legalize dispensaries in states that created legal protections for them, but it did not give medical marijuana wholesalers and retailers full immunity. Indeed, raids still occasionally occurred, but in many cases it was not clear whether the raid was purely for marijuana or for some other type of offense. For the Fiscal Year 2015 budget, Congress passed an amendment to a budget bill that would “prohibit any Federal funds from being used to supersede State law in those States that have legalized the use of medical marijuana” [Commerce, Justice, Science, and Related Agencies Appropriations Act of 2016, H. Amdt. 335, 114th Congr. (2015–2016)]. This was also passed for Fiscal Year 2016.

Tolerance for marijuana commerce is not the same as allowing state and local governments to generate tax revenues from these sales. Just like the market participants, state officials who accept and deposit funds that are known to be from marijuana transactions are committing federal felony offenses. Although it may make sense that federal officials would want state and local government to have funds to offset the cost of regulating these markets, there are no federal guidelines on the appropriate level of taxes and fees that these jurisdictions should be collecting. There is also no public federal oversight for how these funds are being spent.

In the California campaign to legalize in 2010 and the 2012 efforts in Colorado, Washington, and Oregon, some proponents made strong arguments about the amount of tax revenue marijuana would generate for the state. Indeed, the voters in Washington State were told that in the first five years of legalization it could generate up to $1.9 billion in tax revenue (Henchman & Scarboro 2016). California, Colorado, and Washington already had a history of collecting sales tax revenues from medical dispensaries, and some local jurisdictions in California applied additional taxes on marijuana businesses. Because there was a precedent in these jurisdictions that the federal government would allow states to generate revenues from marijuana sales, it seemed plausible the Obama administration would allow these states to collect the fees and tax revenues from the legal regimes.

Of course, one could imagine that the federal agencies would be more likely to tolerate sales of marijuana that are loosely medical rather than those that are explicitly for recreational purposes. Indeed, after the initiatives passed in Colorado and Washington, no one was quite sure what the federal government was going to do. In August 2013, Cole released another memorandum indicating that the Obama administration would not block state efforts to tax and regulate marijuana as long as they had “strong and effective enforcement and regulatory systems” (Cole 2013). Whether future administrations continue this approach remains to be seen.

CONCLUDING THOUGHTS

With President-elect Trump selecting cabinet nominees as we finish this article, it is unclear how federal agencies in the next administration (and others after that) will approach state-level legalization. Although we do not speculate about that here, we are confident that other states will seriously debate legalization, and we would not be surprised to see some form of it on the ballot in multiple states in 2018 and 2020. There is also a possibility that some state legislatures will vigorously debate legalization in the near future. Although all of the successful nonmedical initiatives sans Washington, DC, involve commercial models, legalization does not have to involve profit-maximizing companies. For jurisdictions considering alternatives to prohibiting marijuana supply, there are several middle-ground options, some of which can still generate meaningful revenues for the state (Caulkins et al. 2015, MacCoun 2013).

It is probably uncontroversial to argue as we have that medical marijuana helped to enable the emergence of legalized recreational sales in the United States, but we think the story is an interesting and important one. That does not mean, however, that the two movements will remain intertwined. Indeed, it is by no means clear that the medical marijuana system will survive. Medical marijuana faces two threats. First, many in the recreational industry now see medical marijuana dispensaries as a threat because the latter have typically operated with minimal regulation and tax burden. Medical dispensaries and recreational shops are now competing for many of the same customers. Second, as federal barriers to research are lowered, the science may not back up the claims for medical efficacy, or nonintoxicating products may prove more useful and easier to administer through the traditional pharmacy model.

Can the same political strategy work for other illicit drugs? Cocaine and many opioids are already Schedule II drugs, meaning that they have recognized medical uses, but their medical distribution is far more tightly regulated than that for medical marijuana. There is also far less of a political constituency for facilitating their use. There is increasing attention to clinical uses of psychedelics (e.g., MDMA, psilocybin, ibogaine), and proven medical benefits might serve to destigmatize nonmedical use. Further, current clinical protocols for these drugs require closely supervised use in highly controlled settings.

In some places, the rhetoric of marijuana reform has largely shifted from a plea for public health and harm reduction to promises of state revenues, jobs, and tourist dollars. But because marijuana has been at the leading edge of drug policy reform, advocacy with respect to other drugs might be significantly weakened, giving new meaning to the phrase orphan drugs. But for some advocates on the other side, the “war on drugs” strategy is also wearing thin. If this continues, we may see drug policy lose much of its adversarial drama as a battlefield of the culture wars.