As many of you know, I hosted an FDL Book Salon this weekend with the authors of “Marijuana Legalization: What Everyone Needs to Know.” I’ve been very clear all along about the fact that the actual facts about marijuana are, as a whole, extremely well presented in the book, but I have serious issues with the emphasis on uncertainty upon which the authors focus. I understand that their stated purpose is get “both sides” to realize that there are a lot of unknowns, and I realize that there are unknowns because of the lack of full legalization examples in the modern world.
But there’s more to it.
One of the questions I asked during the salon:
This book focuses extensively on the fact that we cannot know for certain what will happen with legalization. And yet, all public policy (and, in fact, all of human endeavor) involves uncertainty. Every day we act without knowing all the consequences with certainty.
Henry N. Pollack, author of “Uncertain Science… Uncertain World,†said: “Frequently, ‘scientific uncertainty’ is offered as an excuse to avoid making important policy decisions. We must recognize, however, that delaying decisions because of uncertainty is an implicit endorsement of the status quo and often a thinly veiled excuse for maintaining it. It is a bulwark of the take-no-action policy popularly known as ‘business as usual.’â€
Couldn’t the book be considered a ringing endorsement for inaction?
The response from two of the authors was that they were not trying to endorse inaction. However, it doesn’t matter–as academics in drug policy, their emphasis on uncertainty itself acts as an endorsement for inaction, whether they intend it or not.
Now, to many of us on the reform side, any uncertainties about what will happen with legalization are generally irrelevant to the decision. If you believe in the liberty argument, then it doesn’t matter what happens with legalization; legalization is the only option. Then, if there are problems, figure out how to handle them without using the unproductive and corrupt sledge hammer of prohibition.
Just like with freedom of speech, where the courts have ruled that the internet, for example, cannot be forced to dumb down to 12-year-old levels merely because some communications may be offensive to 12-year-olds, the liberty argument says that we don’t “dumb down” life merely because some people are unable to use drugs responsibly. (And the liberty argument isn’t solely a libertarian argument, since it doesn’t necessarily preclude a variation with government intervention specifically for those who abuse.)
For other people, however, who may not accept Mill or similar views, but rather believe in the value of nanny-state approaches to societal well-being, the issue of uncertainty regarding the potential outcomes of legalization can have a powerful impact on their views.
So… just how certain are these uncertainties?
Let’s take a look at one tiny example. In an otherwise mostly good section about marijuana smoking and lung cancer, the authors managed to squeeze in a fair amount of uncertainty:
Marijuana smoke contains carcinogens. What is not known is whether exposure is great enough to cause cancer. […]
What is lacking is clear epidemiological evidence from population studies showing that groups who smoked marijuana had higher rates of cancer than otherwise similar groups that abstained. [very curious wording] The published research shows mixed results. […]
In part, the answer to the cancer question depends on the level of proof one demands. [Another bizarre statement. Wouldn’t that also be true about the whether-the-earth-goes-around-the-sun-or-vice-versa question?]
When it came to the data they used to demonstrate the uncertainties, red flags went up immediately in my head.
I asked this question during the book salon:
For our authors: there are literally thousands of studies on marijuana that have been done all over the world. How do you find and select your data for the book? Obviously, you can’t include it all.
I’m curious as to why, for example, when discussing marijuana and cancer (and you listed several studies), you chose to include the extremely small (79 cancer cases, and 324 controls) New Zealand study that is often touted by our government, yet didn’t reference the 2006 NIDA-funded UCLA study by Donald Tashkin (1,200 cancer cases, 1,400 controls) showing no evidence of a lung cancer connection, even among heavy smokers. http://www.washingtonpost.com/wp-dyn/content/article/2006/05/25/AR2006052501729.html
There were no responses.
It’s a valid and important question when it comes to evaluating the certainty of their uncertainty. After all, why wouldn’t the Tashkin study be included (and why weren’t the controversial limitations of the New Zealand study mentioned)? If the authors hadn’t heard of the Tashkin study, that makes their research suspect; if the authors had heard of the study and chose to use the New Zealand study instead because it served the uncertainty argument better, that makes their entire uncertainty argument throughout the book suspect.
[Update: Beau Kilmer, one of the authors, has since emailed me to say that it was merely a matter of having already listed a couple of earlier studies that were positive and not feeling the need to add another one in the interest of space. I understand that, but still question the thought process behind the inclusion of the New Zealand study.]
It’s easy to say that outcomes are uncertain. If I leave the house today, I might get hit by a bus, or get mugged, or step on a crack and break my mother’s back. Uncertainties without context are useless. Proper analytic approach puts uncertainties in perspective so that they doesn’t make me stay home.
Yet the uncertainties in the book often have no reasonable evaluation of the certainty of their uncertainty, which makes them worthless.
One of the critical areas in the book has to do with marijuana legalization and alcohol abuse.
The authors have often referred to this section. In the intro:
for example, no one knows whether increased marijuana use would lead to less heavy drinking, or to more.
Let’s look at the actual section. Starts out OK…
It might seem intuitive that making marijuana more available would tend to decrease alcohol use; [yes, it would] as competing means of altering one’s mood, one drug can substitute for the other. No doubt if cannabis were legal some of today’s alcoholics would be daily pot-smokers instead; that would, on average, make them and those around them better off.
But then they go off the rails with one of the worst metaphors ever written…
But two drugs can also be mutually complementary. When two commodities are economic complements–like cell phones and cell phone apps–making either one cheaper or more available increases demand for the other.
Marijuana : Alcohol :: Cell Phones : Cell Phone Apps
At that point, they start really speculating (!)
On the other hand, if the effect went the other way–if doubling marijuana use were to increase alcohol abuse and dependence by 10 percent–it’s hard to see how any of the gains on the marijuana side could balance out the harms from increased heavy drinking. And yet, based on what is now known, it’s not possible to rule out even bigger changes, in either direction.
The list of things that it’s also not possible to rule out is literally infinite. And meaningless.
For their “data,” the authors note:
Economists have tried to estimate what they call the “cross-price elasticities of demand” between marijuana and alcohol […] Alas, different studies reach opposing conclusions. […] there is truly no scientific basis for any confident assertion about what would happen to heavy drinking if marijuana were legalized.
However, they don’t list or reference any particular studies, so there’s no way to ascertain any grounds to support their uncertainty (which already seems pretty ridiculous in their own words).
The little bit of research I did immediately found some pretty good references regarding marijuana being a substitute for alcohol. Finding studies showing marijuana and alcohol as complementary was much tougher and the results were not even remotely compelling; most seem to be fishing expeditions conducted by RAND’s Rosalie Pacula (who has a history of anti-pot posturing).
What should probably be a point strongly in favor of legalization (assuming a proper analysis of the comparative studies) ends up being another point of uncertainty, arguing for inaction.
Make no mistake, there’s a place for uncertainty. But uncertainty for the sake of uncertainty is useless. And false uncertainty contributes to the destruction of the drug war.