In November 2012, Colorado and Washington approved ballot initiatives that legalized marijuana for recreational use under state law. Since then, nine additional states (Alaska, Oregon, California, Nevada, Maine, Vermont, Massachusetts, Michigan, and Illinois) plus the District of Columbia have followed suit, either by ballot initiative or legislative action. Voters in four other states (New Jersey, South Dakota, Arizona, and Montana) approved state ballot measures legalizing marijuana for personal use in the November 2020 election.

Supporters and critics make numerous claims about state‐​level marijuana legalizations. Advocates suggest that legalization reduces crime, raises tax revenue, lowers criminal justice expenditures, improves public health, increases traffic safety, and stimulates the economy. Critics argue that legalization spurs marijuana and other drug or alcohol use, increases crime, diminishes traffic safety, harms public health, and lowers teen educational achievement.

In previous work, we found that the strong claims made by both advocates and critics are substantially overstated and in some cases entirely without support from existing legalizations; mainly, state legalizations have had minor effects. This paper updates previous work to account for additional years of data and the increase in the number of states with legalized marijuana. Our conclusions remain the same, but our assessments of legalization’s effects remain tentative because of limitations in the data. The existing data nevertheless provide a useful perspective on what other states should expect from legalization or related policies.

In November 2012, Colorado and Washington approved ballot initiatives that legalized marijuana for recreational use under state law.1 Since then, nine additional states (Alaska, Oregon, California, Nevada, Maine, Vermont, Massachusetts, Michigan, and Illinois) plus the District of Columbia have followed suit, either by ballot initiative or legislative action.2 Four additional states approved marijuana legalization in the 2020 November elections (New Jersey, South Dakota, Arizona, and Montana).

Supporters and critics make numerous claims about state‐​level marijuana legalization. Advocates suggest that legalization reduces crime, raises tax revenue, lowers criminal justice expenditures, improves public health, increases traffic safety, and stimulates the economy. Founder and executive director of the Drug Policy Alliance Ethan Nadelmann, for example, asserted in 2010 that legalization would help end mass incarceration and undermine illicit criminal organizations.3 Former New Mexico governor and Libertarian Party presidential candidate Gary Johnson has also advocated for marijuana legalization, predicting it would lead to less overall substance abuse because individuals addicted to alcohol or other substances would find marijuana a safer alternative.4 Even some law enforcement officials agree legalization lowers crime; Denver police chief Robert White, for example, said in 2014 that violent crime dropped almost 9 percent.5

Critics argue that legalization spurs marijuana and other drug or alcohol use, increases crime, diminishes traffic safety, harms public health, and lowers teen educational achievement.6 Colorado Gov. John Hickenlooper, a Democrat, opposed initial efforts to legalize marijuana because he thought the policy would, among other things, increase the number of children using drugs.7 Former U.S. Attorney General Edwin Meese III, who is now the Heritage Foundation’s Ronald Reagan Distinguished Fellow Emeritus, and Charles Stimson, also with the Heritage Foundation, have argued that violent crime surges when marijuana is legally abundant and that the economic burden of legalization far outstrips the gain.8 Kevin Sabet, former senior White House drug policy adviser in the Obama administration, called Colorado’s marijuana legalization a mistake, warning that potential consequences may include high addiction rates, spikes in traffic accidents, and reductions in IQ.9 David Murray, a senior fellow with the Hudson Institute, and John Walters, a former director of the White House Office of National Drug Control Policy and president and CEO of the Hudson Institute, claimed in 2014 that “what we saw in Colorado has the markings … of a drug use epidemic” and argued that there was a thriving underground marijuana market in Colorado and that more research on marijuana’s societal effects should be completed before legalization is considered.10 John Walsh, the U.S. attorney for Colorado, defended the targeted prosecution of medical marijuana dispensaries located near schools by citing figures from the Colorado Department of Education showing dramatic increases in drug‐​related school suspensions, expulsions, and law enforcement referrals between 2008 and 2011.11 Denver District Attorney Mitch Morrissey pointed to the 9 percent rise in felony cases submitted to his office from 2008 to 2011, after Colorado’s marijuana laws had been partially liberalized, as evidence of marijuana’s social effects.12

Reviews of the literature on the first wave of marijuana decriminalizations in the 1970s note that marijuana use did not change in response to relaxed restrictions.13 Analysis of the recent U.S. state legalizations is more limited, but broader research suggests little to no effect of decriminalization on drug use.14

In previous work, we assessed these claims based on data from states that had legalized the recreational use of marijuana by mid‐​2018. In this paper, we update our earlier work to account for an additional two years of data, both from those initial states and from others that have since legalized marijuana.15

Our earlier conclusion was that the strong claims made by both advocates and critics are substantially overstated and in some cases entirely without real‐​world support. At the time, our data showed that state‐​level legalization of marijuana had generally minor effects. One notable exception was the increase in state tax revenue from legalized marijuana sales; states with legal marijuana markets have collected millions of dollars in state tax revenues. As of July 2020, all but two jurisdictions with legalized marijuana had opened the door for retail sales. Although both Vermont and the District of Columbia officially allow marijuana consumption, neither permits the substance to be bought or sold on the market.

New data reinforce our earlier conclusions. Even with two additional years, however, the data available for before‐​and‐​after comparisons are limited, so our assessments of the effects of legalization remain tentative. Nevertheless, the existing data provide a useful perspective on what other states should expect from legalization or related policies.

Until 1913, marijuana was legal throughout the United States under both state and federal law.16 Beginning with California in 1913 and Utah in 1914, however, states began outlawing marijuana, and by 1930, 30 states had adopted marijuana prohibition. Those state‐​level prohibitions stemmed largely from anti‐​immigrant sentiments and particularly from racial prejudice against Mexican migrant workers, who were often associated with the use of the drug. Prohibition advocates attributed terrible crimes to marijuana and the Mexicans who smoked it, stigmatizing marijuana use and the purported “vices” that resulted from it.17 Meanwhile, film productions, such as the 1936 movie Reefer Madness, presented marijuana as “Public Enemy Number One” and suggested that its consumption could lead to insanity, death, and even homicidal tendencies.18

Starting in 1930, the Federal Bureau of Narcotics pushed states to adopt the Uniform State Narcotic Drug Act and to enact their own measures to control marijuana distribution.19 In 1937, Congress passed the Marihuana Tax Act, which effectively outlawed marijuana under federal law by imposing a prohibitive tax; stricter federal laws followed.20 The 1952 Boggs Act and the 1956 Narcotics Control Act established mandatory sentences for drug‐​related violations; a first‐​time offense for marijuana possession carried a minimum sentence of 2–10 years in prison and a fine of up to $20,000.21 While those mandatory sentences were mostly repealed in the early 1970s, President Ronald Reagan reinstated them under the Anti‐​Drug Abuse Act of 1986. The current federal legislation controlling marijuana possession, use, and distribution is the Controlled Substances Act, which was published in 1971 and classifies marijuana as a Schedule I drug. This category is for drugs that, according to the Drug Enforcement Administration, have “no currently accepted medical use and a high potential for abuse” as well as a risk of creating “severe psychological and/​or physical dependence.” 22

Despite this history of increasingly draconian federal action against marijuana (and other drugs), individual states have been backing away from marijuana prohibition since the 1970s. Eleven states decriminalized the possession or use of limited amounts of marijuana between 1973 and 1978, including, in chronological order, Oregon, Alaska, California, Colorado, Maine, Minnesota, Ohio, Mississippi, New York, North Carolina, and Nevada.23 However, not all states followed such a straightforward path toward marijuana liberalization. Alaska, for example, decriminalized marijuana use and possession in one’s home in 1975, but in 1990, a voter initiative recriminalized possession and use of marijuana. A second decriminalization wave began when Nevada defelonized marijuana possession in 2001; 19 more states and the District of Columbia have since adopted similar reforms.24 By the mid‐​1990s, amid mounting scientific evidence pointing to marijuana’s potential medicinal benefits—including treating chronic pain, glaucoma, Alzheimer’s, Parkinson’s, epilepsy, and other medical conditions—various states began to legalize medical marijuana but restricted access only to patients who satisfied strict criteria.25 Over the past two decades, 33 states and the District of Columbia have legalized marijuana for medical purposes, significantly expanding the number of patients eligible for medical marijuana prescriptions. In some states, these medical regimes approximate de facto legalization.26

The most dramatic cases of states undoing earlier prohibitions and departing from federal policy have occurred in those states that have legalized marijuana for recreational as well as medical purposes (Colorado, Washington, Oregon, Alaska, California, Nevada, Maine, Massachusetts, Illinois, Michigan, and Vermont). Nearly every state that has legalized marijuana thus far has done so through citizen‐​driven ballot initiatives. After formally legalizing marijuana, states normally take one to two years to set up regulatory regimes, establish licensing guidelines, and impose marijuana taxes; only then can the first marijuana shops open.

In the 2020 elections, more states’ ballots included measures to liberalize their marijuana laws. New Jersey, South Dakota, Arizona, and Montana passed ballot measures legalizing marijuana for recreational use. Mississippi and South Dakota voters likewise approved ballot measures legalizing medical marijuana. As of November 2020, the Marijuana Policy Project listed 23 states with bills to legalize marijuana, 14 with bills to decriminalize marijuana, and 12 with bills to create medical marijuana programs.27

Although states’ paths differ in some ways, most follow a pattern of first decriminalizing, then medicalizing, and then legalizing. One exception is Michigan, which did not decriminalize marijuana statewide prior to legalizing medical marijuana—although many cities had adopted local decriminalization laws by that time.28 Another is Vermont, which legalized medical marijuana in 2004, nine years before decriminalizing it in 2013.29 For states following the usual decriminalize‐​medicalize‐​legalize pattern, their experiences with decriminalization and medical legalization inform the expected effects of total legalization, since these partial measures often serve as steps toward that end.

To determine the effects of legalization and other policy changes on marijuana use, we examine the trends before and after the changes. We focus on recreational marijuana legalizations, because earlier work has covered other marijuana policy modifications, such as medicalization.30

The specific statewide legalizations we consider are Colorado (2012), Washington (2012), Oregon (2014), Alaska (2014), California (2016), Nevada (2016), Maine (2016), Massachusetts (2016), Vermont (2018), Michigan (2019), and Illinois (2020).

Our analysis examines whether the trends in marijuana use and related outcomes changed substantially after these dates. We consider trends in alcohol and drug use, suicides, crime, traffic fatalities, and economic conditions. Any observed changes may, however, be due to other factors and do not necessarily implicate marijuana policy. Similarly, an absence of changes does not prove that policy changes had no effect; a confounding variable operating in the opposite direction might have approximately offset the policy change.

One of the most important potential effects of marijuana legalization is increased marijuana use. If increases are minimal, then the other effects of legalization are also likely to be minimal since ancillary effects depend on use.

Figure 1 displays the trends in prevalence of marijuana use in eight states in the 12 months prior to the National Survey on Drug Use and Health from the Substance Abuse and Mental Health Services Administration (SAMHSA). The data are from people aged 12 and older. These prevalence data derive from self‐​reports in the SAMHSA surveys of drug and alcohol use. The vertical line in the graph marks the year of legalization in the states. Use in states where marijuana is legal tends to be higher than use in the United States overall, but this difference mainly pre‐​dates legalization. Among the 11 states that have legalized it, marijuana use rates in 2011—prior to any legalization—averaged 15 percent compared with the national rate of 11.6 percent. Only Illinois’s was lower, at 11.4 percent.31

In many states, use increased modestly in the years leading up to legalization. For example, Maine’s participation rate hovered around 12–13 percent between 2003 and 2009; it then increased to 14 percent in 2011, 16 percent in 2013, and 19 percent from 2014 through 2016. After legalization in 2016, the increase continued to 22 percent in 2017 and almost 24 percent in 2018. Similarly, marijuana use in Massachusetts began increasing in 2012, several years prior to its legalization in 2016. Maine and Massachusetts track the pattern previously seen with early legalizers (Colorado, Washington, Alaska, and Oregon) of increases in use prevalence in the few years leading up to legalization. California’s pre‐​trend is less pronounced, and Nevada’s is flat. Vermont, Michigan, and Illinois demonstrate a similar increase pre‐​legalization, but data for the years following legalization are not yet available. Legalizing states display higher and increasing rates of use prevalence, but these patterns existed prior to legalization.

Much of the concern surrounding marijuana legalization relates to its possible effect on youth. Many, for example, fear that expanded access—even if legally limited to adults age 21 and over—might increase use among teenagers, with negative effects on cognitive development, educational outcomes, or other behaviors. For instance, Madeline Meier and others analyzed a large sample of individuals tracked from birth to age 38 and found that those who smoked marijuana most heavily prior to age 18 lost an average of eight IQ points, a highly significant drop.32 However, other studies have found results that rebut such claims. Claire Mokrysz and others examined an even larger sample of adolescents and, after controlling for many potentially confounding variables, discover no significant correlation between teen marijuana use and IQ change.33 Deborah Cobb‐​Clark and others show that much of the relationship between marijuana use and educational outcomes is likely due to selection, although there is possibly some causal effect in reducing university entrance scores.34 Evidence from Daniel McCaffrey and others supports this selection explanation of the association between marijuana use and educational outcomes.35 M. Christopher Roebuck, Michael T. French, and Michael L. Dennis suggest that chronic marijuana use, not more casual use, likely drives any relationship between marijuana use and school attendance.36 Olivier Marie and Ulf Zölitz estimate grade improvements are likely due to improved cognitive functioning among students whose nationalities prohibited them from consuming marijuana.37 Jan C. van Ours and Jenny Williams concluded that cannabis may reduce educational outcomes, particularly with early onset of use.38 Other studies discuss additional evidence on likely negative effects of early onset of use.39

Figure 2 in the Appendix shows self‐​reported youth marijuana use in the 30 days prior to the survey date, using data from the Youth Risk Behavior Surveillance System, a survey of health behaviors conducted in middle schools and high schools. Surveys are conducted in odd‐​numbered years. Washington and Oregon do not participate in this survey. Nationally, the trend is toward fewer youth reporting marijuana use. Youth participation rates are reportedly higher in legalizing states than in the United States as a whole. Of the six states with post‐​legalization data, in four—Maine, Massachusetts, Alaska, and Colorado—adolescent use reportedly decreases in the years immediately prior to legalization and then returns roughly to prior use rates. The available data show no obvious effect of legalization on youth marijuana use.

The high and increasing rates of marijuana use prior to legalization (shown in Figure 1) might provide evidence for a cultural explanation behind the recent swell of legalizations: as marijuana becomes more commonplace and less stigmatized, residents and legislators become less opposed to legalization. In essence, rising marijuana use may not be a consequence of legalization but a cause of it.

Consistent with this possibility, Figure 3 in the Appendix plots data on perceptions of risk from monthly marijuana use, collected between 2002 and 2018.40 All states that have legalized marijuana fall below the average U.S. risk perception. This is consistent with the view that attitudes toward marijuana fostered both policy changes and increasing use rates. In some states, risk perceptions rose around the time of legalization. This rise may have resulted from public safety and anti‐​legalization campaigns that cautioned residents about the dangers of marijuana use.

Data on marijuana prices may also shed light on marijuana use rates. Before legalization, advocates in some states hypothesized that marijuana use might soar post‐​legalization because prices would plunge. For example, Dale Gieringer, director of the California branch of the National Organization for Reform of Marijuana Laws, testified in 2009 that in a “totally unregulated market, the price of marijuana would presumably drop as low as that of other legal herbs such as tea or tobacco—on the order of a few dollars per ounce—100 times lower than the current prevailing price of $300 per ounce.” 41 A separate study by the RAND Corporation estimated that marijuana prices in California would fall by 80 percent after legalization.42 These analyses consider legalization at both the state and federal levels, which would allow for additional avenues for lower prices such as economies of scale but also for additional avenues for higher prices because of federal taxation and advertising.