As legal marijuana spreads across America, mostly for medical use, anxiety about its side effects is spreading with it: What other changes will it bring? Campaigns against loosening the law tend to focus on its unknown and possibly dangerous repercussions—a surge in pot smoking, perhaps opening the door to increased use of harder drugs and to associated spikes in crime and other societal ills.
Amid the heated debate, a small amount of hard data is starting to emerge. And among the most intriguing findings is a recent study suggesting that Massachusetts could enjoy an unexpected boon from last November’s vote to legalize medical marijuana: fewer deaths on our roads and highways.
A team of economists who specialize in health and risk behaviors looked at the link between marijuana laws and traffic deaths, and found that roadway fatalities dropped significantly in states after they legalized medical marijuana. On average, deaths dropped 8 to 11 percent in the first full year after the law went into effect, and fell 10 to 13 percent by year four. Five years out, the results grew more varied, and faded in some cases.
The study doesn’t include Massachusetts, whose medical-marijuana law just went into effect in May, well after the researchers had finished collecting and analyzing their data. But applied to Massachusetts’ most recent traffic fatality statistics, the study’s findings would roughly translate to about 35 lives saved per year.
The notion that loosening the restrictions on a drug—one that’s hardly known for improving reaction times—might actually improve traffic safety is surprising on the face of it, and the researchers are careful to say that there’s nothing safe about driving under the influence of marijuana. But as they try to unpack what might be making the difference, it is becoming clear that the knowledge emerging from America’s new experiments with marijuana law could significantly change the public conversation—giving us new data about the effects of drugs on society, and landing a familiar debate on unfamiliar new ground.
For more than four decades, starting in 1970, a complete prohibition on pot was the law of the land, both federally and in every state. But in 1996, California cracked the door to legalization by allowing medical marijuana, and 19 states have followed. Two states, Colorado and Washington, have fully legalized marijuana for recreational use, both last year. Meanwhile, it remains illegal under federal law to buy, sell, use, or possess pot anywhere, in any amount.
The state-level legalization trend has been so rapid that there are thus far relatively few definitive studies on its effects. For instance, while medical marijuana laws seem to increase pot smoking generally, there are conflicting findings over whether it increases use among teenagers. A scattering of contradictory, often localized, studies have also been done on changes after legalization in crime, emergency room visits, and the use of other drugs. Obviously, each of these categories is complicated, with numerous factors at work.
Daniel Rees, a University of Colorado economist, and his colleagues decided to look at one major but narrow public-health statistic: state-by-state data on traffic fatalities compiled by the National Highway Traffic Safety Administration. They gathered the numbers and controlled for other factors, such as the overall nationwide decline in traffic deaths, and states individually lowering their legal blood-alcohol limits. It didn’t take long to see a pattern: Medical marijuana laws coincided with less roadway carnage.
The bulk of the team’s work, published this spring in the Journal of Law and Economics, was spent trying to figure out why. After parsing the statistics, the researchers themselves chalk the drop in deaths up to “substitution”—the idea that more pot-smoking means less booze-swilling. (It is assumed by most drug researchers that some medical marijuana leaks into the general population, so it’s not just patients who have more access to the drug.) The counter-argument, often used as a case against legalization, is that cannabis and alcohol are “complementary,” meaning that increased use of one spurs more consumption of the other. Once again, studies of this issue have conflicting results, because it’s tough to get precise consumer data about an illegal product. But Rees and his team say a deeper analysis of their data points to lower alcohol use as the likely mechanism for the drop in traffic fatalities.
For one thing, medical marijuana laws had a smaller impact on the number of deaths in crashes where alcohol was not a factor—a 7 percent drop on average, compared to a 13 percent drop in deaths where alcohol was implicated. In addition, the drop in deaths was more robust among young adults (between 20 and 40), especially young men, and it was stronger on nights and weekends. All of that lines up with what’s known about drinking and driving.
When it comes to traffic safety, says Rees, “the uncomfortable conclusion is that you’d rather have young adults smoking marijuana instead of drinking alcohol. Even I’m uncomfortable with it. But that’s where the logic takes us.”
The researchers offer two possible explanations for why more marijuana use could lead to less drunken driving. One is that pot smoking takes place in different circumstances than drinking. Drinking is legal, and drinks are served in many places that can only be reached by car. People drink at bars, restaurants, ball games, picnics, concerts, and just about any adult social gathering; then they drive home. Because recreational marijuana is still illegal in all but two states, it’s used in a much less open range of environments. In other words, people go out and drink, but stoners tend to stay home. (This is one factor that may start to change if legalization takes hold: In early 2013, the first “pot bars” opened in Colorado and Washington.)
The other possible explanation is straightforward, if definitely not something you’re likely to hear from your local chapter of DARE: It could be that pot availability leads to drunk drivers being replaced with stoned drivers, and that stoned drivers are, on average, safer. In fact, while studies indicate that pot is just as bad as alcohol for distance perception, reaction time, and hand-eye coordination, it appears to be less of a danger in simulated and real-world driving tests. Driving high is by no means safe: A meta-analysis by the British Medical Journal early in 2012 found that drivers who were high on marijuana had nearly double the risk of a serious crash compared to sober counterparts. But driving drunk is worse, causing a tenfold increase in accident risk for drivers with a blood-alcohol concentration at the legal limit of 0.08, or a forty-eightfold increase at the old legal limit of 0.1. The researchers also point out that drivers under the influence of marijuana may “engage in compensatory behaviors” such as driving slowly, avoiding sudden, risky maneuvers, and staying well behind the car in front of them. Perhaps they are just more cautious than a drunk person would be, even though they are still impaired.
Marijuana legalization advocates may be eager to trumpet these results, but the research case is far from closed. Rosalie Pacula, an economist at the RAND Corporation specializing in drug policy research, says medical marijuana laws are far too varied from state to state to draw any broad conclusions about the effects of fuller legalization. (In Massachusetts, the law’s patient-registration requirement places it on the stricter side, though its allowance for up to 35 dispensaries suggests fairly wide distribution.)
In work she’s presented at academic conferences but has yet to publish, Pacula reanalyzed the same crash incident data and found that the drop in traffic deaths was strongest in states that restrict spillover into recreational use by requiring patients to sign on to a state registry, as Massachusetts does. This muddies the case for “substitution,” since presumably those effects would be strongest when pot was most easily obtained. Along the same lines, Pacula’s analysis found that the decline in deaths was offset when marijuana dispensaries were allowed to operate and advertise their services openly under state law.
“I think they have a really interesting finding,” Pacula says. “But this is just the tip of the iceberg. It’s not, aha, we have it!”
What does seem clear is that as more data become available and pot prohibitions continue to fall, America’s approach to marijuana policy will have to get a lot more complicated than “just say no.” The legality of alcohol means that we have both solid information and precise laws about drinking and driving; now, as better data starts to trickle in about marijuana, what we learn will no doubt influence a variety of health and safety measures.
Rees and his collaborators continue to look at the effects of medical marijuana laws. In a forthcoming paper for the American Journal of Public Health, they have found correlations between medical marijuana laws and declines in suicides, and they’re also looking into a range of other effects.
Even if these results support the substitution theory argued in their traffic fatality study, with marijuana substituting for alcohol and perhaps mitigating some of its harmful effects, they acknowledge that there may be other social problems that pot makes worse than booze ever did. “It’s a possibility,” says Mark Anderson, a Montana State University economist and Rees’s primary collaborator on the marijuana studies. “I think that’s where we let the data tell us what’s going on.”
The one certainty is that drug policy is rife with tradeoffs. As we learn more about the experience of states that relax marijuana restrictions, the fallout will certainly be more complicated than just “good” or “bad.” America’s public experiment with looser drug laws has only just begun to tell us what we’ll need to know.Chris Berdik is a journalist in Boston. His book, “Mind Over Mind,” was published in 2012 by Current, an imprint of Penguin.