Search About Donate
An employee at a medical marijuana cultivator in Ohio works with a marijuana plant.
Q&A

How Dangerous is Marijuana, Really?

A Marshall Project virtual roundtable.

On Jan. 7, The Marshall Project published an interview with Alex Berenson, a former New York Times reporter and author of "Tell Your Children: The Truth About Marijuana, Mental Illness, and Violence," which warns that the rush to legalize the drug has obscured evidence of its dangers. The interview stirred up a storm on social media, so we decided to enlarge the discussion.

What follows is a conversation, conducted by email and moderated by Bill Keller, editor-in-chief of The Marshall Project. Berenson is joined by three other panelists. Maria McFarland Sánchez-Moreno is the executive director of the Drug Policy Alliance, a non-profit that advocates ending the war on drugs, including the "responsible regulation" of marijuana. Its donors include companies in the legal, for-profit cannabis industry, whose gifts, the group reports, made up less than 1 percent of the alliance’s 2018 revenue. Keith Humphreys is the Esther Ting Memorial Professor at Stanford University. He has been deeply involved in drug policy as a researcher and White House advisor. Mark A.R. Kleiman is a professor of public policy at the New York University Marron Institute of Urban Management, where he leads the crime and justice program. He is also chairman of BOTEC Analysis, which advises Washington State and Maine on cannabis regulation.

The discussion has been lightly edited for length and clarity.

The Marshall Project: This first question is for all of you. Let's start with the core question Alex set out to answer in reporting his book: What do we know about the connection between marijuana and mental illness? What would you say is established medical science, and what is still unresolved?

Alex Berenson.

Alex Berenson: Okay, I'll jump in. Marijuana causes psychosis. This is an established medical fact, not open to debate. It can cause temporary psychotic episodes even in healthy people. It worsens the course of schizophrenia and provokes severe relapses in people with schizophrenia whose disorder was controlled.

The mainstream literature and the physician-scientists who have done the most work on the issue also believe it is responsible for some cases of schizophrenia that otherwise would not have occurred—that is to say, that it can cause schizophrenia, especially when used regularly to heavily by adolescents. The advocacy community focuses on that part of the issue and overstates the uncertainties surrounding it, which after 30 years of research are relatively minor, and probably relate more to risk ratios than anything else. But in doing so, advocates elide an equally significant issue. Even if marijuana did not produce a single de novo case of schizophrenia, its use frequently causes temporary episodes of psychosis. To say otherwise is to ignore the reality of emergency rooms all over the United States.

Keith Humphreys: I am going to start by making a distinction, which I will rely on throughout this discussion, between "old cannabis" and "new cannabis".

By "old cannabis" I refer to the drug that was dominant in the U.S. throughout the 20th century. It was a plant, almost always smoked, that had single-digit percent THC content and significant CBD as well. Its user base included many people who smoked only occasionally, once a week or once a month. [THC is the chemical in pot that gets you high. CBD is a different chemical that can buffer the effects of THC.]

By "new cannabis" I refer to the drug that has been become dominant in the past 5-10 years. It is very high in THC (average 20 percent in Washington's legal market), lower in CBD and is consumed in a range of ways beyond smoking. Its user base includes many people who consume it every day or nearly every day.

Keith Humphreys.

It's very unlikely that the effects of old cannabis and new cannabis on population health are the same. This is critically important because almost all our research is on old cannabis; by definition any study asking "what does cannabis consumption do over 10 years" is a study of old cannabis. I therefore feel much more humble speculating about the impact of new cannabis than old cannabis.

So, for Bill's questions.

(1) Does cannabis cause acute psychotic reactions?

For old cannabis, the risk was not zero but it was incredibly low. When users had acute toxic reactions to old cannabis and came into the emergency room, it looked much like a panic attack. Unpleasant, to be sure, but the person had not lost touch with reality as in a psychotic reaction. In the age of new cannabis, the risk for acute psychotic reactions is much higher—we can see that in E.R. data both in the U.S. and Europe.

(2) Does cannabis cause schizophrenia? The data from old cannabis is consistent with two explanations, (a) that cannabis has a very low risk of causing schizophrenia (SCZ) in people who would not get it otherwise or (b) in people with vulnerabilities to SCZ, use of old cannabis can make the disorder emerge earlier or have a worse course. If I had to bet my house on one of those two, I would bet on (b). As for the answer to this question for new cannabis, unlike data on acute effects (psychotic reactions, car accidents) we can't know yet about SCZ; it will take years to tell. I hope that when we know, we will not find out that explanation (a) has become true—that cannabis causes SCZ in people who would not otherwise get it, but that is definitely possible in my opinion.

Mark A.R. Kleiman: Keith's point is central: Most of what we thought we knew about cannabis is now open to question because the material (and, I would add, techniques of self-administration, such as dabbing) has changed so much. So we should treat strong assertions that cannabis is safe with caution.

Alex's initial contribution brings out, I think, the two different kinds of issues that have arisen around his book and around the journalistic coverage of it. There are scientific questions—about what is known, or believed on good evidence—and there are rhetorical questions, about how claims are communicated.

Mark A.R. Kleiman.

Start with Alex's first sentence: "Marijuana causes psychosis." As Keith says, there's a sense in which it's true: consumers of "new cannabis" sometimes part company with consensus reality for a period typically lasting hours. But a psychotic event—something which will happen to about 20 percent of any population at least once during a lifetime—is not at all like psychosis (and schizophrenia in particular) as a persistent mental-health condition, which is one of the worst things that can happen to anyone. The sentence as written—and specifically the use of "psychosis"—is deeply ambiguous: excellent as an attention-grabbing headline, not so good at informing a reader who is thinking about cannabis policy, or the reader's own cannabis use, about the risks.

In addition to wanting to know what sort of "psychosis" cannabis might cause, a sensible reader would want to know how often those bad things happen: both what fraction of psychosis is attributable to cannabis, and what the probability is that any given pattern of cannabis use will lead to psychosis, however defined.

Dr. Ziva Cooper, an associate professor of clinical neurobiology in the psychiatry department at Columbia and a member of the National Academies panel that wrote the report Alex's book cites for that claim, writes (in this Twitter thread):

"...we did NOT conclude that cannabis causes schizophrenia. We found 1) an association between cannabis use and schizophrenia and 2) an association between cannabis use and IMPROVED cognitive outcomes in individuals with psychotic disorder."

She adds: "Since the report, we now know that genetic risk for schizophrenia predicts cannabis use, shedding some light on the potential direction of the association between cannabis use and schizophrenia."

Note how central that last point is: if the people genetically more vulnerable to schizophrenia are also more likely to use cannabis, then the finding that people who use cannabis are (modestly) more likely to be diagnosed with schizophrenia doesn't tell us anything about whether cannabis use increases the risk. But since "new cannabis" is so different from "old cannabis" that's only moderately reassuring.

So much for the science. Now back to the rhetoric.

Imagine that we were discussing vaccination instead of cannabis, and someone started the conversation with "Vaccination causes paralysis." We'd all groan, and say "Oh no, we've got an anti-vaxxer."

That would be the case even though vaccination does, in fact, cause paralysis. As we discovered during the Swine Flu scare, about one per million injections leads to Guillain-Barre syndrome, which involves paralysis. But that doesn't make "Vaccination causes paralysis" a sensible statement in that context.

Likewise, “Tell Your Children,” as German Lopez writes in Vox, “while a compelling read written by an experienced journalist, is essentially an exercise in cherry-picking data and presenting correlation as causation."

That's part of the reason for the negative reaction that Alex's writing, and the writing about it, have generated from people who have made a career out of studying this stuff. That criticism looks to Alex like "denial" and "advocacy," but I think the problem is more the difference in professional practice between journalism and science.

Maria McFarland Sánchez-Moreno: Marijuana has been in wide use in the U.S. for many decades, but due to marijuana prohibition, research on its potential harms and benefits has been limited. What the available evidence indicates is that the main mental health risk of marijuana use is cannabis use disorder (CUD), which is characterized by a continued problematic pattern of use despite negative consequences. CUD affects about 9 percent of users.

Researchers have also found an association between marijuana and rare, temporary incidents of paranoia or panic following over-consumption of marijuana; medical providers have labeled these incidents “marijuana-induced psychosis.” But contrary to Alex’s over-the-top claim that marijuana “frequently causes” psychosis (which most readers will understand to mean something very different from these incidents), most scientists are reluctant to make sweeping statements about causality. The fact is that the overwhelming majority of people who use marijuana do not develop a mental disorder as the simple result of using it.

Maria McFarland Sánchez-Moreno.

What research does indicate is that some people with certain genetic markers for schizophrenia and other psychiatric disorders may be more likely to use marijuana and other drugs. There is also research to suggest that individuals with certain markers may be more likely to experience psychosis or develop psychiatric disorders, including schizophrenia, due to substance use.

But the development of any mental illness is complex and multifaceted, and we know relatively little about how the many variables that may be relevant interact. If we’re in the business of banning anything that is associated with psychosis or schizophrenia, then we should ban alcohol, tobacco, all other stimulants (including coffee), the passing on of certain DNA, homelessness and possibly growing up with cats.

Conversely, researchers have recently found an association between marijuana use and improved cognitive outcomes in individuals with psychotic disorders. They have also found that cannabidiol (or CBD), a cannabinoid found in marijuana, improves outcomes in patients with schizophrenia when given as an adjunct medicine.

Keith rightly notes that variations in marijuana’s potency (in both the legal and illegal markets) may lead to different effects and that there have been increases in marijuana-related E.R. visits in some states post-legalization, though the total number remains exceedingly low. Whether that increase is due to a rise in potency or use, or simply the fact that users now feel safe to seek help when they don’t feel well, has yet to be established. Regardless, the effects of evolving marijuana products will have to be carefully studied to ensure effective regulation of potency, dosage, labeling and advertising, as well as to ensure effective consumer education. Under prohibition, of course, such regulation and education is impossible—it’s the illegal market that sets the potency and the consumer knows little about the risks or how to minimize them.

TMP: I'd like to drill down a bit on the relationship between cannabis and psychosis. Alex states flatly that marijuana "causes psychosis." Keith says that "In the age of new cannabis, the risk for acute psychotic reactions is much higher—we can see that in E.R. data both in the U.S. and Europe." Maria is more dismissive, asserting that when people speak of "psychosis" in this context they really mean "rare, temporary incidents of paranoia or panic following over-consumption of marijuana." So what do we mean by "psychosis," and is there a meaningful difference between saying marijuana "causes" it and saying marijuana is a risk factor?

McFarland: Psychosis is a phenomenon that can involve delusions, paranoia and/or hallucinations. Psychotic episodes can be a contributing symptom towards a diagnosis of schizophrenia, or they can be stand-alone experiences resulting from various biological, psychological, or environmental factors.

Nobody has established that marijuana causes psychosis. Even Ziva Cooper, an author for the National Academy of Sciences report that Alex keeps citing, recently corrected him and said that they did not establish a causal link between marijuana and schizophrenia. Why Alex keeps repeating this patently false claim about causation, in the face of all the science, is a mystery.

Humphreys: I agree with Maria that we don't know if cannabis use causes schizophrenia in people who would not otherwise have it or not (a long-term psychotic illness). As I said before, under old cannabis, it is possible that the association emerges because people with vulnerability to schizophrenia have earlier first psychotic breaks or more protracted disease course if they use cannabis. That would still be bad for them of course, and for what it's worth I always tell people with a family history of psychotic illnesses that marijuana (as well as stimulants) are more dangerous for them than the average person. But at the same time it doesn't mean we can say that old cannabis causes the disease of schizophrenia.

Whether new cannabis will cause schizophrenia is, as I said, something we cannot know, it might do so now that the drug is stronger and used so intensively by many users. We run a lot of experiments on our population. When television was introduced, we didn't do a randomized trial we just exposed everyone to it. A mix of good and bad things followed, but it did not destroy us as some said it would at the time. When we introduced lead into gasoline and exposed everyone to that, we didn't realize how much damage we did to the brain development of children for a long time, but it now appears the damage was massive. New cannabis could be like television or lead in terms of how it affects population mental health, but we will not know that for years.

Psychotic reactions are different than the onset of lifelong illness—they are an acute effect that doesn't take years to evidence themselves. For those we can say that new cannabis causes more psychotic reactions, we can see that already in emergency room data. One could argue that other factors are often involved in toxic psychotic reactions and in seeking help for them and this is true, but if that means it's a non-issue then automatic weapons have no role in mass killings because other factors are involved there as well.

One intermediate outcome (i.e., something that emerges more slowly than psychotic reactions but faster than schizophrenia rates will if in fact new cannabis changes them) that I find intriguing is how many people present for treatment for cannabis addiction. There is a nice study by Michael Lynskey and colleagues of cannabis potency and treatment of cannabis use disorder in the Netherlands over 16 years, where cannabis has long been legally available through the cafes and there is no legal pressure to seek cannabis treatment. The curve of potency goes up and up to the point that the government got concerned and regulated down potency. The curve for admissions to cannabis treatment is remarkably similar in shape to the potency curve, following along behind it by a few years upward when the potency was rising and then going back down when it falls. So it's not just whether it's safe to come forward—it was always safe to come forward in The Netherlands, it's the drug itself and how it's regulated.

So far we have barely regulated new cannabis. I could be wrong but I don't think a single state caps potency or anticipated that price collapse feeds heavy use. That is producing more acute psychotic reactions immediately, should produce more cannabis use disorder treatment seeking in the next few years, and may or may not raise schizophrenia rates over the next decade (depending on whether new cannabis has the power to induce SCZ in people who would not otherwise get it; obviously I hope it doesn't but I don't think we can tell yet if this will be like television or lead in terms of how it affects population mental health).

Opening Statement

Sign up for our daily newsletter covering the best in criminal justice news.

Berenson: I am surprised Maria is choosing to take this angle of argument.

The reasons for the complexity of the science around cannabis and psychosis are both hard and easy to explain. They are hard because psychosis and schizophrenia are incredibly complicated disorders, and our understanding of them is so crude even now. Some top schizophrenia researchers now even believe we should retire schizophrenia as a classification and only refer to psychosis—because schizophrenia is a classification of exclusion and outcome more than anything else. Here's what I mean. Diagnosing cancer is relatively simple. You go to the physician with a lump in your breast, or a chest X-ray finds a spot on your lung. A biopsy returns cancerous cells. You are then checked to determine how far in your body it has spread. Based on the staging and type of cancer, you are presented with various treatment options. And if you choose, you can easily find out the odds of five-year survival and cure for your type of cancer.

Diagnosis of schizophrenia is an entirely different process. There's no blood test for it, no brain scan. Psychiatrists make it based on the way a patient is acting and how he or she says he is feeling—after excluding other organic causes such as brain tumors or dementia, along with drug use. But what are the odds of recovery? We don't really know. Why do anti-psychotic drugs work in some patients and not others? We don't really know. If a patient recovers from schizophrenia and is able to function without anti-psychotic drugs after a year or two—and that does sometimes happen—did he really have schizophrenia at all? We don't really know. Defining a disease based on its result doesn't seem to make a lot of sense.

At the same time, while so-called positive symptoms such as hallucinations, paranoia, and thought disorders are usually the reasons that someone winds up diagnosed with schizophrenia, the disorder also encompasses negative symptoms such as apathy and disorganization and cognitive symptoms such as a loss of IQ. Those can sometimes be more crippling than the positive symptoms. I will never forget—this was when I worked at the Times and was writing about Zyprexa—Dr. Thomas Insel, at the time the head of the NIMH, telling me, people can get through life with a lot of strange ideas if they can get up in the morning and go to work. Anti-psychotic drugs rarely treat those symptoms, and can even make them worse.

So schizophrenia is both less and more than the positive symptoms of psychosis. And it is no wonder that we don't understand at the cellular level how cannabis affects psychosis or the course of schizophrenia, when there are so many unknowns about the basic biology of the disease itself.

But that complexity does not mean a connection between cannabis, psychosis, and schizophrenia cannot be made, or proven to a reasonable degree of scientific certainty. And it has been. Not just through epidemiology. We know that cannabis and THC can cause temporary psychotic episodes. Even in laboratory settings, when volunteers have explicitly been screened for pre-existing psychosis (because ethical researchers will no longer knowingly expose people with pre-existing psychosis to THC), there are case reports of volunteers suffering psychotic symptoms after being given THC. We know that people who use cannabis receive a diagnosis of schizophrenia at a younger age than those who do not—and younger age of diagnosis is a crucial negative factor in the course of the disease. We know that people whose schizophrenia is controlled can suffer quick and severe relapses if they begin to use again. We know that the synthetic cannabinoids, which bind to the same CB1 receptor as THC, though more strongly, can cause severe and even permanent psychosis in people after only a few uses.

The fallacy that cannabis users are simply self-medicating has been debunked over and over. As early as 2002, the British Medical Journal wrote in an editorial, "these findings strengthen the argument that use of cannabis increases the risk of schizophrenia and depression, and they provide little support for the belief that the association between marijuana use and mental health problems is largely due to self-medication."

The National Academy of Medicine report from 2017 is even more clear. In the first sentence in its chapter on cannabis's effects on mental health, it writes, "Cannabis use is likely to increase the risk of developing schizophrenia and other psychoses; the higher the use, the greater the risk." In the Jan. 12, 2017 press release discussing the report, NAM wrote, "The evidence reviewed by the committee suggests that cannabis use is likely to increase the risk of developing schizophrenia, other psychoses, and social anxiety disorders, and to a lesser extent depression." In the two years since the report was released, as far as I know, no one seriously disputed its findings.

To be clear: if cannabis use increases the risk of developing schizophrenia, it is causing cases of schizophrenia that otherwise would not occur. If a population of 1,000 non-users has a 0.5 percent risk of schizophrenia, 5 people will eventually develop schizophrenia (on average). If all 1,000 instead use cannabis regularly in adolescence, and regular cannabis use in adolescence triples that risk to 1.5 percent, 15 of those 1,000 (on average) will develop schizophrenia—10 additional cases. There is no way around that math.

And now we get to the easy part of the reasons for the confusion—the advocacy community simply refuses to acknowledge the weight of the scientific evidence, as Maria's email shows. It is worth noting that the psychiatric risks of cannabis are unrelated to its legal status. We can know that cannabis can cause psychosis and still believe it should be a fully legal drug. That's a political debate. But the Drug Policy Alliance is not doing anyone—including users—any favors by refusing to acknowledge the risks here. People are going to use cannabis whether it is legal or not. And we need to make sure that they know these risks and give them simple and clear messages:

TMP: Maria, I'm curious: do you take issue with the five warning messages Alex recommends?

McFarland: It’s puzzling that Alex in the same response talks about the complexity of mental illness and the many factors associated with it, and then goes right back to his irresponsible claims that we “know” that marijuana “frequently causes” psychosis. That’s just not true. Even if you don’t believe Dr. Cooper, the NASEM report author who corrected Alex, the report itself states that “the issue of comorbidity [substance use disorders and mental health disorders] directly affects the ability to determine causality and/or directionality in associations between substance use and mental health outcomes.” We also know that the majority of people who use marijuana will not develop any of these issues. Interestingly, contrary to Alex’s suggested message about a lack of medical benefits, the NASEM report also found “conclusive or substantial evidence that cannabis or cannabinoids are effective for the treatment of chronic pain in adults, …chemotherapy-induced nausea and vomiting, [and] …multiple sclerosis spasticity symptoms.”

As for messaging around marijuana risks, it’s critically important—as I’ve said before—that we offer sound and sober public education around the real risks and benefits of drugs, rather than hyperbole. That’s why the Drug Policy Alliance has developed an entire high school curriculum, called Safety First, around drugs that’s modeled after modern sex education: equipping kids with the information they need to make safer choices. We also have for many years had a useful handbook for parents on how to talk about drugs to their kids, and reduce the harms of both drug use and drug prohibition. Sadly, much drug education in the past has been overwhelmingly focused on scaremongering, similar to Alex’s distorted claims about marijuana leading to psychosis, violence and homicides. And that has been a disservice to kids, who end up not trusting adults, and not having the information they need to make the best possible choices for themselves.

Berenson: I'm sorry. Help me out. Is that yes or no to none, any (and which ones), or all of those five specific potential warnings?

McFarland: I don’t think it’s productive to spend much time here quibbling about the details of Alex’s proposed messages. So, very briefly: the first message is part of our Safety First curriculum, where we tell young people that the teenage brain is uniquely susceptible to harmful changes after alcohol or other drug use, and that while abstaining from any use is the safest choice, delaying use will reduce harm. As for #2, of course, if anyone’s anxiety or depression is worsening, for any reason, they should speak with a professional—that’s not unique to marijuana. The third message would need to be rewritten more precisely to be accurate. Alex’s fourth message is off; it’s unclear where he’s getting this. Finally, as I stated earlier, the NASEM report that Alex keeps misquoting makes clear that there are established medical benefits to cannabis, so part of message #5 is wrong.

More importantly, these more restrained messages are inconsistent with Alex’s own alarmist claims that cannabis “frequently causes” psychosis in otherwise healthy individuals. By shifting the focus to these messages, he seems to be trying to avoid having to defend the deeply problematic core of his book: his repeated insistence that cannabis use causes psychosis; that in turn, this supposed psychosis causes violence; that legalization increases cannabis use, psychosis and violence; and that therefore, legalization is a bad idea. We’ve discussed the first of these claims, but I hope we can start to dig into the rest.

TMP: One thing I suspect all of you will agree on is that we should lift the ban on FDA research into the health effects. All of you see a need for more science. Given the magnitude of the potential harm of "new cannabis," does it make sense to legalize now rather than wait for some more definitive research? What if, in Keith's lovely analogy, "new cannabis" turns out to be, not a benign innovation like television, but a serious menace, like lead paint?

McFarland: I’d say no to waiting. Any thoughtful discussion has to weigh the reality of how things work under prohibition, and the full scope of harms flowing both from marijuana use and from its prohibition. The reality is that under prohibition, marijuana is widely available—millions of Americans use it—and there are no regulations in place to protect health or safety: no requirements around labeling, dosage information, product quality and testing, etc. In that sense, responsible regulation and education is a much more effective way to protect public health.

Also under prohibition, hundreds of thousands of people are arrested every year in the U.S. just for possessing marijuana for personal use, facing the risks of incarceration and deportation and the devastating collateral consequences of a conviction, such as loss of employment, housing, education and the right to vote. A grossly disproportionate share of those arrested are black and brown, meaning that communities of color are far more likely to bear these terrible burdens. Criminalization and incarceration themselves have terrible public health consequences, and are associated with increased mental health problems. But Alex’s analysis mostly dismisses these grotesque, concrete, clearly documented harms.

TMP: I have one question for Mark, and then I’d offer all of you a chance to share any additional thoughts. Mark, I interviewed you back in 2013 when you were advising the state of Washington as it tried to build a legal marijuana marketplace. You had come around to the view that while pot was harmful, a well-regulated market was the best hope of minimizing the harm. Now that several states have some experience of legalized cannabis, would you say your faith in the regulated marketplace has been vindicated?

Kleiman: Quick version: We don't really have a choice about legalization, because prohibition is broken beyond repair.

But we can legalize in ways more or less friendly to public health, and what the legalizing states have chosen so far is on the "less" side of that choice. We want high prices, extremely strict limits on marketing and aggressive consumer education about the risks of cannabis and techniques for maintaining moderation. That should include point-of-sale delivery by professionally trained and licensed "bud-tenders," functioning more like pharmacists than like audio-store salespeople.

TMP: I'm curious, has any jurisdiction come close to the kind of market you recommend? And if not, why not?

Kleiman: Canada may get reasonably close. Everywhere else, the laws were written by legalization advocates, who were so focused on telling voters that cannabis was harmless they didn't worry about preventing its actual harms.

Humphreys: I worked on Mark's team advising Washington State on how to implement legalization and was on the steering committee of California's Blue Ribbon Commission on Marijuana Policy with then-Lieutenant Governor, now Governor, [Gavin] Newsom, which advised on our state's legalization options. Drawing on those experiences and my discussions with policy makers and activists around the U.S., I think every single legalizing state has missed significant opportunities to protect public health and to bring people of color into the legal industry (it is also worth noting that the racial disparity in cannabis-related arrests is as large as ever in legalization states). The newly legal industry looks a lot like the tobacco industry: An under-regulated, under-taxed, politically connected, white-dominated corporate entity that generates its profits mainly by addicting lower income people to a drug (85 percent of Colorado’s cannabis for example is consumed by people who did not graduate from college). We could do a lot better, but we haven't yet. I am still hoping for a U.S. state to legalize in a way that prioritizes public health and racial justice over profit, but am not optimistic that this will happen. I agree with Mark that Canada (and I would add Uruguay) will almost certainly implement cannabis legalization in a more sensible fashion than will the U.S.

McFarland: As the only Latina in this group, I feel compelled to highlight the jarring insensitivity Alex’s writing reveals toward prohibition’s impact on people of color. The worst example is this passage in the book:

“Further, the civil rights issues around marijuana legalization are far more complicated than the media or politicians would like them to be. Yes, marijuana arrests disproportionately fall on minorities, especially the black community. But marijuana’s harms also disproportionately fall on the black community. Black people are more likely to develop cannabis use disorder. They are also more likely to develop schizophrenia—and much more likely to be perpetrators and victims of violence. Given marijuana’s connection with mental illness and violence, it is reasonable to wonder whether the drug is partly responsible for those differentials.”

In other words, Alex is proposing that one reason that police arrest black people three to four times more than white people may be because marijuana supposedly makes black people psychotic and therefore more violent. And here we get to what may be the ugliest outcome of Alex’s book: a bio-chemical excuse for both racially biased policing and for marijuana prohibition. Conveniently, Alex has ignored the fact that black and white people use marijuana at similar rates; the reason for the higher rate of arrest of black people is the over-policing of communities of color in the name of enforcing prohibition. This statement reeks of the crack baby and super-predator myths of the 1990s, which much of the media uncritically repeated and which inspired a dramatic escalation of sentences nationwide, including such horrors as life without parole for juveniles. If we want to avoid repeating these grave mistakes, it’s critical that we guard against the sort of alarmist propaganda that Alex’s book is spreading.

Berenson: The misery of psychosis, and the violence psychosis causes, afflict people of all races. In the United States, most people with diagnosable psychotic illnesses are white. I never use the term super-predator in my book, and I never would. That said, the differentials I mention in that paragraph are real. Large studies show that African-Americans are about twice as likely to have diagnosable cannabis use disorder as white people. They also show that African-Americans are three to four times as likely to be diagnosed with psychotic disorders as white people. And black people are far more likely to be both victims and perpetrators of violent crime, including homicide. I am sorry if Maria and the Drug Policy Alliance find this discussion of facts unpleasant. I do as well. I wish they weren't the case, just as I wish that people with psychosis were not more likely to commit violence. But as Dr. Sheilagh Hodgins told me, "The best way to deal with the stigma is to reduce the violence." To pretend reality is not reality, or to argue that discussing reality represents "jarring insensitivity," helps no one, least of all the victims of violence—whether those victims are black, white or Hispanic. And in fact black Americans are less, not more, likely than white Americans to believe that marijuana should be legalized. I don't think anyone of any race should lose his or her future because of a marijuana possession arrest. Very few people are in prison for marijuana possession in the United States, so we may as well drop the fiction of enforcement that those arrests create and simply decriminalize the drug. It is true that decriminalization will not end all arrests or racial disparities in law enforcement, but neither will full legalization. Even in states that have legalized, black people continue to be arrested more frequently than white people. Marijuana-related arrests are a result, not a cause, of racial disparities in law enforcement and even larger societal problems. As for Mark's and Keith's discussion of the pros and cons of various forms of legalization, they underestimate the problem that home-grown marijuana represents. Any legalized regime that includes expensive licensing requirements, “high prices” (presumably through high taxes), and limits on extracted products (as in Canada), will be undercut by home-grown cannabis. Because heavy users of cannabis consume so much of the product—as Mark notes in his book—price differentials of even a few dollars an ounce are enough to ensure a thriving black market. And heavy users want THC extracts; the black market will supply those if the legalized market won’t. This problem simply doesn’t exist for alcohol, where black market products generally can’t compete. So the legalization scheme that Mark proposes may lead to both a group of for-profit companies that will inevitably promote their product and lobby lawmakers and a violent black market—the worst of all possible worlds. I will confess that I am a bit confused as to why Mark insists on the necessity for full national legalization of cannabis. It is not as if the rest of the world has legalized, and the United States is an outlier. Indeed, we are an outlier, but in the other direction—European countries have lower rates of use than the United States and are for the most part nowhere near recreational legalization or even widespread "medical" authorizations. One of the facts that surprised me most when I researched “Tell Your Children” was that between 2000 and 2005, the United States and Britain had roughly the same levels of cannabis use. Since then, a steady drumbeat of pro-marijuana messages has driven up use in the United States, while tenacious and sustained efforts—led by Sir Robin MacGregor Murray, one of the world's leading experts on schizophrenia—to educate Britain about the potential mental health effects of marijuana have helped lower use there. Today, Americans use cannabis nearly twice as much as people in the United Kingdom. The United States also has a far worse opioid crisis than the UK, and while suicide rates are soaring in the US, they are falling in the UK. Education matters. Knowledge of risk matters. Legalization sends the wrong message. Maybe, instead of pursuing a policy that is all but guaranteed to encourage even more use of this drug—if the experience of the states that have already legalized is any guide—we should instead do everything possible to communicate its potential harms and discourage people of all ages and races from using it.

McFarland: I’m sure one could have a fascinating discussion about why clinicians more often diagnose schizophrenia among people of color than among whites. In fact, the study Alex cites points to a variety of potential factors that may influence the rate of diagnosis, including possible clinician bias and the under-diagnosis for major depressive disorders among black people, concluding that, “Because clinicians are the diagnosticians and misinterpretation, bias or other factors may play a role in this trend [and] caution should be used when making inferences about actual rates of psychosis among ethnic minority persons.”

Alex’s argument also makes no sense, as rates of marijuana use are roughly the same for black and white people. If rates of schizophrenia diagnosis are starkly different, but rates of marijuana use are more or less the same, then that undermines any claim that marijuana use is what drives the differentials. Another problem with the quote in Alex’s book: the overwhelming majority of arrests for marijuana possession are not of people exhibiting violence; therefore, there’s no reason to attribute racial disparities in arrests to disparities in violence or psychosis.

Legalization is not a panacea for racial injustice in our society, and racial disparities in the criminal justice system will continue despite legalization. However, as we’ve seen in the states that have legalized, the total number of arrests will drop dramatically. And that matters, especially to the communities of color that are disproportionately impacted, and that overwhelmingly bear the brunt of prohibition. If Alex doesn’t believe me, maybe he should talk to Bernard Noble, who recently served seven years of a 13-year sentence in a Louisiana prison for possessing two marijuana joints. Or maybe he should talk to Bernard’s seven kids, who grew up without their father. Or maybe he should talk to the thousands of people in California who showed up to our expungement fairs, to try to get their records cleared of old marijuana convictions post-legalization—because arrests and criminal records can mean jail time, deportation, loss of jobs, housing and the right to vote.

Keith is right that more needs to be done to ensure that legalization furthers racial justice. I’m proud that my organization is fighting to ensure that legalization in New York, New Jersey and New Mexico includes provisions for record expungement, reinvestment in the communities most impacted by prohibition and market access for those communities. It’s unfair for black and brown people to get punished under prohibition, and then excluded from the benefits of legalization. Unfortunately, there are sectors of the marijuana industry that have opposed such measures.

As in any market, with marijuana there will be industry actors who pursue their own profit, sometimes at the expense of what’s best for society. States have the responsibility to regulate them effectively to protect public health and further racial justice, and we as advocates will seek to ensure that legalization is as effective as possible at reducing both the harms of marijuana use and the harms of prohibition.

All of this brings me to one of the key problems with federal marijuana prohibition: it keeps states from experimenting with state control or forms of legalization that don’t rely entirely on private actors. If we want to be able to explore reforms like Uruguay’s or Canada’s, which might limit the influence of the industry, then the answer is getting rid of federal prohibition.

TMP: Thanks to all of you for an illuminating exchange. I don’t know whether it will change minds in either direction, but it will challenge easy answers.