How to legalize cannabis

With a little help (ok, actually quite a lot of help) from my friends Lowry Heussler, Jon Caulkins, Keith Humphreys, and Beau Kilmer, and my sister Kelly, I produced an op-ed for the Financial Times on the design of a post-prohibition cannabis control regime.

Here’s the punchline:

A large increase in problem use might be a price worth paying to rid ourselves of the many ills attendant on prohibition. But it is not a price we have to pay. Smarter policies could lead to better outcomes.

Many thanks to Kesewa Hennessy, Deputy Comment Editor at the FT, for a superb job of copy-editing.

One key point, omitted to save space: the system of user-set quotas proposed for cannabis could also apply – should, in my view, be applied – to alcohol and gambling.

How to legalize cannabis

Debating *whether* to legalize pot is pretty pointless. The important debate now is *how* to legalize it. Some notes toward an essay on that topic.

Debating whether to legalize pot is increasingly pointless. Unless there’s an unexpected shock to public opinion, it’s going to happen, and sooner rather than later.

The important debate now is how to legalize it. The results of legalization depend strongly on the details of the post-prohibition tax and regulatory regimes. In the current situation, continued prohibition might be the worst option. Full commercial legalization on the alcohol model might well be the second-worst. But that’s the way we’re heading.

I’m preparing an essay about designing a post-prohibition regime. After the jump is a set of topic sentences and paragraphs for sections of that essay, not yet in a well-defined order. (UPDATE: Numbers inserted to facilitate comments.)

Substantive comments are welcome. Rant and snark will be ruthlessly zapped.

Continue reading “How to legalize cannabis”

Statistics on the Prevalence of Drug Users Can Be Misleading

A new paper shows how drug use prevalence statistics can skew perceptions of drug policy

Drug policy development and analysis often focuses on the prevalence of users, most commonly the proportion of the population that has used one drug or another in the past 30 days or past year. These are the data that are typically being cited when people say things like “Drug use is down/up by 10%”. But the prevalence of people who have used a drug in the past 30 days or year is a crude statistic that papers over extraordinarily diversity in these users’ drug consumption and consequences thereof.

For example, if everyone who uses cocaine once a month starts using it once every five weeks instead while everyone who uses it 2-5 days a month starts using it every single day, a politician could point to the changed prevalence of past 30-day cocaine users and claim a big drop. But the total amount of cocaine being consumed would have risen sharply, as would the number of people who were getting addicted or experiencing other damage from cocaine use!

Among U.S. policymakers, The Obama Administration is I believe the only one in history to embrace this reality by setting no goal for the prevalence of drug users within the adult population. Rather, the goals of the drug strategy for people over the age of 25 are to reduce the number of chronic heavy users of hard drugs and the prevalence of drug-related morbidity and mortality.

In the research world, the wizards of drug policy at RAND Corporation have taken on the issue of simple prevalence measures in a new journal article that is available for free here. In a series of helpful examples, they show how reliance on these measures can skew perceptions of a range of drug policy issues. The example below comes from their paper.

The left two columns of the table tell a familiar story: African-Americans make up a much higher proportion of drug offense arrestees than they do past-year drug users. From this one might draw a range of conclusions, for example that decriminalizing possession would reduce the African-American arrest rate more than the White arrest rate, or, that interventions to reduce racial discrimination in policing (e.g., stop and frisks that lead to drug arrests) would particularly lower the African-American arrest rate. However, these conclusions are based on the assumption that all past-year drug users are the same, which the right column of the chart shows is false.

With the same crudely-defined prevalence category (past year drug use, yes or no) exists substantial diversity on another dimension: Frequency of making drug purchases. Whites account for a smaller proportion of purchases than their proportion of past-year users would lead one to expect, whereas African-Americans show the reverse pattern. There are many reasons this could be so — Whites buying drugs in bulk more often, Whites more commonly sharing drugs in a social network with only one purchaser, dealers being more available in some neighborhoods than others — but whatever the mechanism, the data in the third column suggest new interpretations that were hidden when the measure was simply past-year use. If arrests are mainly about purchases for example (and they do track them remarkably closely), decriminalizing simple drug possession might not affect African-Americans disproportionately after all.


Does this mean that prevalence of users statistics are inherently deceptive? Only if they are used to mean something that they don’t. But even when they are interpreted correctly, it still leaves observers in the dark about much of the reality of drug use in a society. That’s why we need other measures such as volume of consumption, purchasing pattern and prevalence of addiction.

Cannabis, alcohol, and public safety: Knowing what you don’t know

Cannabis might be a substitute for alcohol. Or they might be complements. Right now, we just don’t know.

[See update and correction below. Apologies to Adam Nagourney and Rick Lyman.]

The main problem with journalism, in a world full of unknowns and uncertainties, is that faith is news and doubt isn’t. 

Mark Anderson of Montana State and Daniel Rees University of Colorado at Denver presented an interesting paper at this year’s meeting of the International Society for the Study of Drug Policy, using some moderately sophisticated statistics to investigate the relationship between cannabis availability, alcohol consumption, and traffic accidents. The paper moved my opinion about whether cannabis and alcohol are substitutes or complements from “Not a clue; could go either way” to “There’s a little bit of evidence for substitution” (that is, for the proposition that making pot more available or cheaper would lead to less heavy drinking).

Net substitution would be a big deal, if true, because alcohol does so much more harm than cannabis that a small reduction in the alcohol problem would, in social-cost terms, outweigh even a big increase in the cannabis problem due to legalization.

I say “a little bit of evidence” because the change in cannabis availability due to medical-marijuana laws is poorly measured; because the changes in cannabis availability due to medical marijuana laws are small compared to the changes that would result from full legalization; and because the short-term effects might not only be larger or smaller than, but might not even be in the same direction as, the long-term effects. (Remember the guy who jumped off the Emprire State Building. Being asked as he passed the fiftieth floor, how the experiment was going, he replied, “So far, so good!”)

Unfortunately, Anderson and Rees decided to build on their finding, and other findings from the literature, to make a strong claim: “We expect that the legalization of recreational marijuana in Colorado and Washington will lead to increased marijuana consumption coupled with decreased alcohol consumption. As a consequence, these states will experience a reduction in the social harms resulting from alcohol use.”

Even more unfortunately, Adam Nagourney and Rick Lyman  of the New York Times decided to take that claim at face value, ignoring the papers in the same issue of the same journal by Rosalie Pacula of RAND and Eric Sevigny of the University of South Carolina arguing that the Anderson and Rees claim is more than the current data can support.

Pacula and Sevigny didn’t make the opposite claim: that legalization will increase heavy drinking. Had they done so, normal journalistic practice would have been to cite the two conflicting opinions. But since they brought in the Scotch verdict of “Not Proven,” Nagourney and Lyman simply ignored what they had to say. That let them use the Anderson and Rees claim to support the theme of the story, well-captured by the headline “Few Problems with Cannabis for California.”

[Update and correction:

The above paragraph is inaccurate about the sequence of events. Adam Nagourney writes:

You were wrong to assume I was aware of the dissenting studies and chose to ignore them. I was not; if I was, I most certainly would have mentioned the concerns. Professor Rees  gave me an advance copy of their study; I did not have an advance copy of the journal. Again, if I had, I most certainly would have raised the fact that other people question their assertions.
As someone who has been doing this a long time, I can tell you that this: “The main problem with journalism, in a world full of unknowns and uncertainties, is that faith is news and doubt isn’t” is not true, at least for me or the place I work. One of the great things about the New York Times is that there is room for nuance and subtly. A story is stronger journalistically if it addresses both sides of an argument.

This time I have to give myself an F in Journalism 101; I should have checked with Nagourney before criticizing his work.

Very sorry, sir! Won’t happen again, sir!]

In order to know how cannabis legalization would affect heavy drinking, you’d have to know:

1. The extent of the price decline.
2. The extent of the change in other conditions of availability and use (convenience, marketing, perceived product quality and variety, stigma, legal consequences, social customs).
3. The responsiveness (“elasticity”) of demand to price changes, consisting of two components: the “participation elasticity” measuring changes in how many people use cannabis at all, and the “conditional elasticity” measuring changes in consumption-per-user.
4. Ditto for responsiveness to changes in non-price factors.
5. The “cross-elasticity” between cannabis and alcohol.
6. For all of these, you’d need estimates about the long-term effects of lasting changes, not the immediate effects of transient changes.

All Anderson and Rees can show are reasonable but not perfect estimates of the short-term effects of relatively small and poorly measured short-term changes in price and availability on various outcomes. They’re not within a million miles of being able to offer the sort of prediction on which one ought to base public policy.

For example: One likely impact of legalization, as Anderson and Rees note, is increased cannabis use among minors. Even if it were known that cannabis and alcohol are, at a given point in time, substitutes for minors, that wouldn’t show that increased pot-smoking by sixteen-year-olds might not increase their risk of becoming heavy drinkers at twenty-one. And the Anderson and Rees methods don’t even look for such long-term effects.

“It’s not what you don’t know that hurts you,” said Will Rogers. “It’s what you know that ain’t so.” Knowing what you don’t know doesn’t get your name in the newspaper. But it saves you having to eat your words. Since I think that cannabis legalization of some sort (1) is probably a good idea and (2) very likely to happen in any case, I devoutly hope that Anderson and Rees turn out to be right. But it seems clear to me that Pacula and Sevigny have the stronger argument. We just don’t know.

Public opinion on cannabis: Is it “Game Over”?

Not quite. But close. Time to focus on policy details.

Public opinion on cannabis legalization has been trending positive since sometime in the late 1990s. By a couple of years ago, it was just about break-even. But in the latest Gallup poll, the “pro” side has moved to an almost-20-point advantage, 58 to 39.

Of course this calls for some caveats.

1. The question, as asked, was about legalizing use, not production and sale.

2. With a sample size of only 1000, some of the movement might be noise rather than signal. The Galston-Dionne analysis of Pew Center data from earlier in the year shows a more closely divided country; again, it’s an open question whether that difference is measurement error or real change over a few months.

3. Bad outcomes in Washington or Colorado might reverse the trend.

All of that said, the new numbers are enough to move my probability distribution. For a while now, I’ve been predicting national legalization sometime in Hillary Clinton’s second term, but without much conviction and with the caveat that the trend in public opinion would have to continue in order to make that prediction come true. Now I’d have to say it will come true unless the trend were somehow to reverse.

Note also that at some point positive feedback takes over. Being for pot legalization has a different social meaning when it’s the clear majority view. The polling results will therefore influence the behavior of some survey respondents and voters directly. Moreover, politicians and other opinion leaders shift their announced (and perhaps actual) views in response to perceived shifts in voter attitudes, that in turn will shift the beliefs of some voters.

If the question of whether to legalize now seems largely settled, that makes the much-less-debated question of how to legalize even more topical. Some of the smarter opponents of cannabis have figured this out, and are now looking for ways of limiting the increase in drug abuse likely to follow legal availability. However, career and ideological interests and group ties are likely to lead the majority of the active drug warriors to keep fighting what now seems like an unwinnable battle, telling one another that legalization is sure to be such a disaster that the public will demand re-prohibition. By doing so, the warriors will help to ensure that the legal system that eventually arises will be over-commercialized, under-regulated, and under-taxed.

This would simply repeat the mistake they made in opposing the medical use of cannabis. While the warriors kept chanting “Cheech and Chong medicine,” the pot advocates rolled right over them. Rien appris, rien oublié.

Update Andrew Sullivan strikes a triumphal note. Hard to fault him for that. But goddammit, “less harmful than alcohol” and “not harmful to most of its users” do not add up to “harmless.” Adolescents spin out on cannabis and wreck their academic careers. People of all ages do stupid things while stoned, including driving their cars into trees and other cars. Cannabis now follows only alcohol as the primary drug of abuse reported by people voluntarily entering drug treatment.

Why take the perfectly reasonable case that cannabis should not be illegal and ruin it with the silly claim that the stuff is harmless?

Cannabis taxes will wind up too low, not too high

The big threat to cannabis legalization isnt high taxes; it’s low taxes and loose regulation.

Legal cannabis will naturally be much, much cheaper than illegal cannabis. A joint is the same sort of item as a teabag: the dried flowers of a plant in a wrapper. A fancy teabag costs a dime at the supermarket; the marijuana in an average joint costs about $4 (0.4 gram of sinsemilla flowers @ $10/gram) on the current illicit and quasi-medical markets. The combination of not having to worry about law enforcement and the economies of mass production will inevitably drive the joint price down close to the teabag price. (Generic tobacco cigarettes can be manufactured for about two cents each; the remainder of the price is marketing expense, quasi-rent on brand names, and taxes.)

Now that the federal government has made it clear that state-licensed production in Washington and Colorado will mostly get a pass from federal law enforcement, and now that Washington has decided to allow outdoor growing, avoiding the production bottleneck that might have resulted from the lags in local land-use approval for growing facilities, I’d expect to see much lower-than-current prices in Washington State’s commercial stores no later than next fall. (If I had been running things, I would have started with lower tax rates to speed the transition to the legal market, and then raised taxes to offset the fall in market prices, but the tax rates were in the legislation the voters passed.)

Even at current prices, cannabis is a remarkably cheap intoxicant. A  drinker who hasn’t built up a tolerance might need about $5 worth of mass-market beer to get sloshed; a similarly fresh cannabis smoker could float away on half a normal joint: call it $2 worth. Colorado medical dispensaries already offer their “weekly special” strains of sinsemilla at $5/gm., with volume discounts, and vaporization seems likely to lower the effective cost substantially.  Anyone who’s worried about the price of cannabis is spending far too much time stoned.

Taxation, even if it is very heavy on ad valorem (percentage-of-price) basis, won’t change that picture much; Washington state will collect something like 40% of total retail prices in tax, but 40% of “damned near free” isn’t very much. Colorado’s taxes will be even lower.

The illicit markets may start out with a price advantage over taxed and regulated commercial markets for a year or two. Even then, the quality/reliability/ legality advantages enjoyed by the legal stores would be expected to quickly push the illicit business to the margins, as legal alcohol has done with moonshining. That will be especially true if the states that legalize make a vigorous law-enforcement push against purely illicit activity.

The untaxed and (in Washington State) unregulated quasi-medical markets may also enjoy a price advantage; if people with medical recommendations can buy their cannabis tax-free, we should expect a certain amount of arbitrage. How best to rein in the out-of-control medical-recommendation systems is a challenge that Washington and Colorado (and California, which hasn’t legalized for non-medical use but which has more “medical marijuana” outlets than it has Starbucks) will confront over the coming months and years.

So the biggest worry about legalize cannabis would be a big upsurge in heavy use, and that worry would be exacerbated to the extent that the growth in heavy use is among juveniles. (Provision to minors is by definition illegal, but, as with alcohol, it seems unlikely to generate a distinct illicit market; younger smokers who can’t buy in the stores will be supplied by older siblings, older friends, their parents’ supplies at home, and straw purchasers.

How big that problem turns out to be depends in part on unknowns and in part on policy choices. If cannabis prices are allowed to fall to something like their free-market levels, a very large increase in heavy use would be the likely result. Preventing that will require heavy specific-excise taxation (perhaps on a per-milligram-of-THC basis) and enough enforcement to prevent the evasion of that tax.

The other approach to limiting the increase in heavy use and use by minors would be on the information side: limits on marketing, required vendor training, aggressive consumer information both at point of sale and in the community.

Naturally, true-believing libertarians insist that cannabis legalization be done in the way likely to generate bad outcomes. Taxes BAD! Regulations BAD! “Commercial speech” is SACRED! The free market FOREVER! And of course drug abuse is a merely imaginary problem, so cannabis is just an ordinary commodity that the market will handle perfectly.

It’s possible that they’ll get their way, and that as a consequence the results of cannabis legalization will be just about as bad as the drug warriors keep predicting: the reproduction of our very bad, no good, awful alcohol markets. It’s barely possible – this is what my drug-warrior friends hope – that the results will be so awful that the voters shy away from legalization altogether.

Jacob Sullum has elevated the temporary risk that relatively high taxes starting will slow down the migration from the illicit to the licit market into an existential threat to the legalization project, and is more or less encouraging Colorado pot fans to double-cross the rest of the voters by rejecting the taxes that were the premise of last year’s legalization push. “From a consumer’s perspective, something has gone terribly wrong if legal marijuana prices do not end up being substantially lower than black-market prices.” That’s true, if by “consumer” you mean someone who smokes multiple joints per day. For anyone else, the cost of weed is way down in the rounding error of a personal budget; a weekly smoker might now be paying something like $100 per year for cannabis, even at current prices.

Andrew Sullivan seems to take this seriously. Talk about solving the wrong problem!

What did the Washington Liquor Board do about cannabis concentrates?

Why did the Washington State Liquor Control Board backtrack and allow the sale of cannabis concentrates?

A friend who follows drug policy is puzzled by this line in a Reuters story about the semi-final rulemaking for Washington’s regulated cannabis market.

Responding to concerns of fueling a black market, the board also clarified that highly potent marijuana extracts,which have gained in popularity in recent years, may be legally sold so long as they are adulterated with at least trace amounts of an inert substance, such as vegetable oil.

Yes, it’s rather hard to figure out what that means, but the reporter – working within space constraints – can certainly be pardoned for not explicating fully. Here’s the story as I understand it:

The Liquor Board faced a legal question and a policy question.

The legal question is whether the initiative-passed statute the Board is supposed to implement, which permits the sale of “useable marijuana” (defined as flowers) and of “marijuana-infused products,” does or does not allows the sale of concentrates such as hashish or “hash oil.” The lawyers for the Board decided that it didn’t. But – as I among others pointed out – that created a definitional puzzle. Clearly, a drop of butane hash oil in a gallon of olive oil would constitute a “marijuana-infused product.” But how about a drop of olive oil in a gallon of BHO? At what magical point does diluted BHO become infused olive oil?

The policy question, at first blush, was whether concentrates are more dangerous than the herbal form. My guess is that the answer will turn out to be “It depends,” – e.g., on the difference between controlled-dose vaporization and dabbing – but we don’t actually know much about either the relevant pharmacology or the relevant consumer behavior. We do know that the number of trips to the emergency room involving cannabis is up about 50% over the past decade, and now runs just under half a million visits a year, so the question of how to prevent cannabis-induced panic attacks does need some attention, and there’s reason to think that preparations and modes of administration that get more THC to the brain more quickly are associated with more risk.

But here again there was a complexity. The Board could ban the sale of concentrates in the stores it regulates, but not in the unregulated quasi-medical market or the purely illicit market. So a ban on concentrates might have slowed the migration of consumers into the taxed and regulated market without much reducing the actual use of concentrates.  (That’s the part of the story the reporter was trying to tell.)

Still more complexity: The law allows the sale and possession of up to an ounce at at time of “useable marijuana,” but up to 72 ounces of “marijuana-infused products in liquid form.” That makes sense for, e.g., cannabis-laced root beer, which is mostly water. But 72 ounces of pure BHO could be $100,000 worth, in an easy-to-smuggle package. So if concentrates are indeed going to be sold, the quantity limit ought to be a fraction, rather than a multiple, of the quantity limit for herbal cannabis.

In the end, the Board decided to use the legal ambiguity to punt the policy question. The “one-drop” rule will apply, allowing concentrates to be sold as long as they have some admixture, no matter the how trivial, of something else. In January, the legislature will be asked to amend the law to explicitly allow what can’t be prevented, while imposing a reasonable quantity limit of perhaps five grams.


If I sometimes seem just a tad impatient with those who are “for” or “against” something they call “marijuana legalization” based on abstract general principles, the concentrates story shows why. This sh*t is complicated, people!

Why Biological Tests of Driving Impairment are Superior to Behavioral Performance Tests

Biological tests of driving impairment are lousy until you compare them to every other kind of test

Andrew Sullivan has had a series of posts recently about “driving while stoned” (Examples here, here and here). He and a number of the people he quotes argue that driving impairment from drug consumption does not correlate perfectly with biological measures (e.g., blood alcohol level as measured by breath test, concentration of cannabis and its metabolites in drawn blood or urine), and that such tests are therefore inferior to assessing impairment directly.

Sometimes in the early morning, when I drive to work on a narrow two-lane road, I see headlights coming toward me. At those moments I hope that if the oncoming driver is alcoholic that he’s had a few belts before he got into the car, because badly shaking, sweaty hands probably impair an alcoholic driver more than does somewhat elevated blood alcohol concentration (BAC). That’s an example of the broader reality to which Sullivan points: More drug consumption does not always mean worse driving. Why then don’t we toss biological tests aside and instead rely solely on behavioral tests of driving impairment, for example reaction time tests done on a laptop, proprioception and coordination-based physical tests (e.g., touching one’s nose with eyes closed), verbal demand tests etc?

The case for such a change at first seems strong, until one recognizes that an ideal driving impairment assessment method would not only perfectly measure impairment, but would also be reliable across testers and testees. Behavioral impairment measures can’t hold a candle to biological measures in this regard.

If like me you have no life and have therefore watched many videotaped driving impairment tests done by police officers, you will have observed how hard it is to administer them in a consistent fashion. Consider the old “walk in a straight line, heel-to-toe” behavioral test. When this test is administered to a driver, both the driver and the police office are generally on the side of the road. Sometimes the ground is flat and sometimes it’s not. Sometimes the road is wet or icy and sometimes it’s dry and clear. Sometimes a strong wind is blowing and sometimes the air is still. Sometimes there is traffic zooming by which distracts both the tester and testee and sometimes the road is quiet. Those are only a few of the variables (none of which influences the result of a blood or breath test) that make it hard to administer the test reliably.

Those variables all introduce measurement error even when the best of capacities and intentions are present…and they aren’t always present. If the officer doesn’t like you for some reason, or is exhausted after a long day, or is all hyped up from the violent crime scene s/he just left, or had an argument with the spouse that morning, the behavioral performance test will be even less reliable in determining the officer’s final judgment of whether the driver is impaired. In contrast, a breathalyzer in every one of this situations will give virtually the same answer.

The reliability of biological tests is a facilitator of equal treatment under the law. Fans of behavioral impairment tests should consider whether, for example, a young black male who has had three beers and has to do a series of performance tests under the watchful eye of white cop is likely to do as well as a white, middle aged guy who has had the same amount to drink. The former driver is more likely than the latter to have his performance on behavioral impairment tests decremented by nervousness or fear of the police officer. Likewise, the immigrant who doesn’t speak English well enough to precisely understand the officer’s instructions will also be at a disadvantage relative to the native speaker. To again draw a clear contrast, if you give a blood or breath test all those factors are eliminated from the measure of impairment.

Which leaves me agreeing that perfectly reliable behavioral impairment tests would be better than what we have now, in the same sense that I agree that universal, high-quality health care at low cost would be better than what we have now. But what of it? Winston Churchill (who was probably a better driver at .08 BAC than at .00) said that “democracy is the worst form of government except all the others that have been tried”. The same could be said of biological assays of driving impairment.