About that RAND ‘study’

There have been a lot of media reports in the past few days talking about the new RAND study that shows how California legalization will result in as much as an 80% decrease in marijuana prices and doubling of marijuana use.

Except, of course, that the RAND report doesn’t really say that at all.

Altered State? Assessing How Marijuana Legalization in California Could Influence Marijuana Consumption and Public Budgets. By: Beau Kilmer, Jonathan P. Caulkins, Rosalie Liccardo Pacula, Robert J. MacCoun, Peter H. Reuter

It’s a 55 page report with lots of interesting stuff in it, but when it comes to an actual projection of change in marijuana use with legalization, they have absolutely no idea.

Here’s what they’re pretty sure they know:

Consumption will increase, but it is unclear how much because we know neither the shape of the demand curve nor the level of tax evasion (which reduces revenues and the prices that consumers face).

So much for the doom reports about the entire state being stoned all the time.

However, that doesn’t make a very interesting report, so Rand decided to make up some shit.

Thus, in the absence of marijuana-specific information, we multiply our participation elasticity of –0.3 by 1.75 to proxy the total elasticity. After accounting for possible income effects, we settle on a baseline total price elasticity of –0.54.

Add to this even more uncertainty…

Once again, readers should not interpret our use of these two particular demand curves, the 25-percent evasion rate, or the $50-per-ounce excise tax as signaling what we think the most likely scenario will be. The purpose of Figure 4.2 is only to demonstrate how much additional uncertainty there is about revenue and consumption estimates, above and beyond that already illustrated in Figure 4.1, when one recognizes that none of the other model parameters is known with certainty.

So, where do media get the notion that the RAND study says that marijuana use will double?

A central question about legalization is whether it will make marijuana consumption go up a little or a lot. The central point of Chapter Four is that it is hard to answer that question because there is great uncertainty about how much consumption will increase. However, it is also hard to answer that question because different people have different thresholds for distinguishing between “a little” and “a lot.”

For the sake of exposition, we will consider a doubling in consumption as a bright line for defining whether consumption changes are small or large.

Yep. They just made up that figure for the purpose of argument.

Now, they probably should have stopped there, but no, they decided to apply these completely imaginary numbers to real world situations and estimate, for example, how many more people would need treatment for marijuana if legalized (based, of course, on nothing).

At points, this discussion gets completely surreal. Let’s take a look at the discussion on drugged driving. It starts out fine…

While driving under the influence of marijuana or any other intoxicating substance can be risky, a question remains about whether marijuana use impairs individuals sufficiently to cause crashes and fatalities. While there is significant experimental literature suggesting a diminished effect on response rates and performance under very strictly controlled conditions, evidence from epidemiological studies has been less conclusive (Ramaekers et al., 2004; Blows et al., 2005).

OK, that’s good. According to actual facts and studies, there’s no real significant causal evidence between pot smoking and car crashes. And, as we know from above, there’s no real evidence to show a significant increase in pot use with legalization, let alone stoned driving.

So?… (Check out the gymnastics in this next bit.)

However, a simple calculation suggests that, if someone believes that marijuana is causally responsible for many crashes that involve marijuana using drivers, legalization’s effect on crashes could be a first-order concern for them. […]

There is no empirical evidence concerning an elasticity of fatal accident rates with respect to marijuana price, prevalence, or quantity consumed, and, as we have underscored repeatedly, there is enormous uncertainty concerning how legalization might affect those outcomes.

However, 50- or 100-percent increases in use cannot be ruled out; nor can the possibility that marijuana-involved traffic crashes would increase proportionally with use. So it would be hard to dismiss out of hand worries that marijuana legalization could increase traffic fatalities by at least 60 per year…

If someone actually said the above to me in person, I would accuse them of being high and having one of those weird free association monologues. How does an academic actually find the guts to say something that absurd? It’s embarrassing.

What if I were to say: “We have no evidence to show that stepping on a crack in the sidewalk will break your mother’s back or how often people step on a crack, yet if someone believes crack-stepping leads to back-breaking, then it would be hard to dismiss out of hand the concern that an increase in sidewalks could increase the number of disabled matrons by at least 60 per year.” People would think I was crazy.

Again, remember that RAND at least tells the truth in the report about not knowing anything. Which is good. But given how the press loves fresh meat, it appears that they then had to go ahead and give projections that they knew were pulled out of their asses, and that they probably knew would be misused in the press.

RAND tries to pretend that it’s a nonpartisan and unbiased research center, but as long as it employs people like Rosalie Pacula, who has acted time and again like the RAND Drug Policy Research Center is her own personal tool for opposing legalization, it is an organization with absolutely no credibility.

She couldn’t even resist putting in her own personal opinions in the AP story about the RAND study.

Rosalie Liccardo Pacula, who co-directs the RAND Drug Policy Research Center, acknowledged that as a California voter, she was uncomfortable with “the lack of specificity” in both the ballot measure and the bill that would have put pot in the same regulatory category as alcohol.

“Neither was sufficient for us to get an idea of what the effect of this was, and as a voter that was disturbing to me,” she said.

At least the other authors of the study, while often legitimately criticized for being afraid to give proper consideration to prohibition alternatives, have the integrity to act like researchers. Pacula is a serious blight on the organization.

This entry was posted in Uncategorized. Bookmark the permalink.

41 Responses to About that RAND ‘study’

  1. Wow. Very good work here Pete.

  2. allan420 says:

    I’m with David. Wow…

    That’s not just gymnastics, that’s unsophisticated slight-of-hand with a dash (just a dash mind you)(*snark*) of perfidy.

  3. claygooding says:

    The Rand group have been the goto statistics group that the ONDCP has used every since their inception,for skewed statistics showing success in the WOD. They were at the table in support of the ONDCP at the committee hearing in April and could not give any statistical proof of success
    in stopping drugs or at least they did not offer any.
    They are allegedly a non-profit organization and we would really like to be able to find out who actually owns it,
    because they do most of the statistical studies for our government,on many issues and government policies and the government probably owns it.

  4. Pingback: Feinstein opposes pot legalization measure – San Francisco Chronicle

  5. Nick says:

    Thanks for the information Pete.
    Thanks for saying “shit”.
    Now what the fuck did they just say?

  6. Pingback: Tweets that mention About that RAND ’study’ - Drug WarRant -- Topsy.com

  7. Pingback: California: An Altered State? - FOXNews (blog) | USTaxAid Services

  8. Rev. Run says:

    Wow, Pete. Thank you SO much for this. I admit that I had fallen for a “reasonable sounding” press release and I just assumed (never do that) that RAND was only utilizing the highest standards in scholarship.

    What a stinking turd they’ve produced.

  9. Jon says:

    There are good explanations for each of the steps you quote out of context.

    E.g., total elasticity exceeds the participation elasticity because it also reflects the conditional elasticity. For MJ the participation elasticity has been measured directly, but not the total elasticity, and total elasticity is what is needed for this sort of analysis. Since for tobacco and alcohol, the total elasticity is roughly 1.5 – 2 times the participation elasticity, it is not an unreasonable starting point for sensitivity analysis to adjust the participation elasticity for marijuana by the same amount. Making no adjustment would implicitly assume that the conditional elasticity was zero. That (having a not insignificant participation elasticity, but zero conditional elasticity) would be rather odd.

    Your next paragraph focuses on the projection that is conditional on the $50 per ounce and 25% evasion rate. The forecasts are given conditional on tax rate and evasion rate because those parameters are not ones that can be forecasted from the scientific literature. Rather, the analysis creates a tool that lets any person who has their own private beliefs about what the tax rate and tax evasion rate would be to convert those judgments into an understanding of the range of plausible effects on consumption. Note: Those two values tend to make consumption increases smaller not higher, so if your “complaint” is that the media picked up the consumption projections from that scenario, then you really shouldn’t complain; there are other scenarios with lower effective tax rates and bigger increases in consumption.

    Read the entire text surrounding the “bright line” and you’ll see that it has nothing directly to do with the projection of the effect on consumption; rather, it is an attempt to help people understand what a doubling would mean. In particular, the central point of that passage is that we’ve already seen something along those lines in 1978.

    RE the final comment — Neither the Ammiano Bill nor the RCTC spells out in much detail what the regulatory structure would be. They are better thought of as empowering certain organizations (ABC and localities, respectively) to create a regulatory structure. E.g., the RCTC does not even specify a tax rate, and the rate could be different in each jurisdiction. That creates uncertainty for modelers/analysts and for voters. I’m not clear why it is wrong for someone to dislike uncertainty.

    • Pete says:


      Thanks so much for responding and your good explanations of those particular points of the report. Regarding the first three paragraphs of your response… Keep in mind that in those sections of my post, I was not criticizing the methods of the research, but merely pointing out, as the report itself did over and over again, that, despite the detailed analysis, there was no possible way of knowing if any projection was right, or even in the ballpark.

      Whether the press got it from one point or another in the report matters little. The fact is that the researchers went out of their way to point out that they didn’t know, and yet, surprise, surprise, the media went ahead and implied much more certainty. The question I had was whether the researchers should have known that would happen and do a better job of forestalling it.

      Everything in that first part of the report was correct, as far as I could tell, but people just needed to be made aware of the level of uncertainty.

      Where I started to really take issue with the RAND report itself was when it took all these uncertainties and started applying them to things like dependency and drugged driving. I notice that you didn’t respond to that. It was at that point where the report lost its way, in my mind.

      Regarding Rosalie’s comments to the AP… She didn’t say:

      “As a researcher, it’s important to note that the proposed bill empowers localities to create different structures, and that means that any attempt to make specific projections is really impossible.”

      Instead, the co-director of the Drug Policy Research Center that had just completed a report said:

      “…and as a voter that was disturbing to me.”

      She didn’t say “as a researcher,” but “as a voter.” She expressed her own personal bias against the measure, while releasing a report. At its most charitable, that’s pretty bad timing. At worst, it signals someone who was prepared to do a hatchet job if necessary.

      Again, thanks for stopping by. I value your response greatly.

  10. Don Frost says:

    Thank you Pete. Very well done. I just have to say that it has been proven by my team of government paid scientists that; Stepping on a crack will break a mothers back, but… It just may not be your own mother, it may be the mother of someone you know, or some other mother you may not even know. Sidewalks are dangerous. If it were not for sidewalks, we would have less abductions. Studies show that kids get abducted from sidewalks. I propose a new law banning sidewalk use unless it is for medicinal purposes.

    Again, Thank you Pete. This was a good follow up to the RAND report. Legalize it, the time is now. PEace

  11. Cannabis says:

    Whenever a reporter/stenographer says that it’s hard to convey shades of gray in a story and headline I will point to this post. It’s really not, but only if your editor & publisher will let you. The bias is intentional, even if it’s not obvious to the casual reader. Good work, Pete!

  12. claygooding says:

    With all the disclaimers in the report it becomes apparent that the report was released for the prohibitionists to hold up and point out the worst case scenario and claim as proof of marijuana’s “dangers”.
    The Rand group,being a government supported non-profit organization was paid to do this study by someone and my money goes to it being the ONDCP because they are still,after 40 years of failure,trying to scare America into keeping marijuana/hemp prohibited.

  13. Pete says:

    There’s one thing I find somewhat ironic in this report, and that’s the participation of Peter Reuter, who was the co-author of the AEI report An Analytic Assessment of U.S. Drug Policy just five years ago, where it was stated:

    Nor do we explore the merits and demerits of legalizing drugs, even though legalization is perhaps the most prominent and hotly debated topic in drug policy. Our analysis takes current policy as its starting point, and the idea of repealing the nation’s drug laws has no serious support within either the Democratic or Republican party. Moreover, because legalization is untested, any prediction of its effects would be highly speculative.

    Yep, no point in considering legalization because there’s just no political support for it… and yet, five years later, here he is doing exactly that — and guess what? It turns out that legalization has serious support among the people.

    The second point is also notable — that any prediction of its effects would be highly speculative. Well, that didn’t stop the group now from speculating.

    So when Jon in comments above defends Rosalie’s personal vendetta against legalization with “I’m not clear why it is wrong for someone to dislike uncertainty,” he ignores the fact that the entire RAND report is constructed of uncertainty and would be so regardless of the details of the legalization proposals.

  14. Rob MacCoun says:

    Interested readers should go to http://www.rand.org and examine the full report and the 9 supporting background papers (all available for download for free) if they want to reach their own conclusions about what we actually said and did.

  15. Cannabis says:

    And read the press release, too. If you ignore all of the stories in the press it’s all good, right?

  16. Pete says:

    Thanks, Rob.

    Yes, please do. There’s lots of interesting reading in the reports. You should never take the word of the press reporting on anything like this (or my word, for that matter).

    There are probably some useful calculations that may shed light on some aspect of legalization that I didn’t discuss. Although it’s unlikely that you’ll find anything other than the uncertainty about projecting future use (that both I and the report emphasized), or that you’ll find any real justification for including the crack-stepping drugged driving nonsense.

    If you do, please let me know.

  17. claygooding says:

    I find it great that Rob MacCoun is here to clear the air on what they studied or how they got their figures but I am still baffled as to why anyone would even do the statistics while filling the report with disclaimers,that the press and prohibitionists ignore,put it out as statistics.
    And then release the price drop in the US and the increase in users in England.
    It just sounds like propaganda to me.

  18. claygooding says:

    And Mr MacCoun,can you tell us who paid for this cracks in the sidewalk statistical study?

  19. David Marsh says:

    Figures lie and lairs figure. The report is meaningless statistically. A 100% increase in the use of a substance that is safe only means that more people will not be harmed when cannabis is legalized.

  20. Rob MacCoun says:

    As noted in the report: “This project did not have an external sponsor; the time used to conduct the work was either donated by the authors or internally funded by two of the RAND Corporation’s units: RAND Health and RAND Infrastructure, Safety, and Environment.” The results are you are discussing are not “statistics” (although there are lots of statistics in the report); they are model projections under the assumptions of the analytic model and the parameter values, which were either varied systematically (to show how various assumptions about unknowns would change the projections) or repeatedly sampled from triangular distributions using monte carlo simulation. The conclusion about uncertainty followed from the modeling; the model projections show that the likely consumption and revenue consequences vary dramatically depending on which assumptions one makes about, for example, the shape of the demand curve.

  21. claygooding says:

    If it was an internal exercise,why put it out. The Rand org is non-profit,so out of curiosity or heart felt civic duty they volunteered their time and efforts in an endeavor for the good of the country.
    Geez,thanks fellows,I’ll sleep better knowing we have ppeople like you worrying about the future price of marijuana or how many people will be doing it.
    If the Rand Org wants to estimate something,give us an estimate on how many more billions of dollars the government will spend trying to keep marijuana prohibited.

  22. claygooding says:

    PS marijuana/hemp

  23. Bruce Buds says:

    The press release sounds a lot more cetain than the study!
    I think legal cultivation will probably lead to lower prices (Yay!), but I’m not sure how it will increase the number of smokers. A recomemdation for medicinal weed is VERY easy to get here in Cali (I got mine to alleviate the constant shoulder pain I get from working at a computer) and by now the green stuff is ubiquitous – If you want to get stoned here you can already.
    Perhaps RAND is concerned that legalization will take cash out of the hands of big Pharma.

  24. allan420 says:

    I’d certainly love to see Mr MacCoun and friends spend their time doing a similar study of our drug policies. And no fancy statistical projections required…

    How does such an org conduct a study such as this and not consider the effects of reduced legal restrictions on cannabis in countries like the Netherlands and Portugal and extrapolating with those known trends?

  25. Rob MacCoun says:

    What Can We Learn from the Dutch Cannabis Coffeeshop Experience? — 2010
    MacCoun RJ

    * * *
    MacCoun, R., & Reuter, P. (1999). Does Europe do it better? Lessons from Holland, Britain, and Switzerland. The Nation, September 20th. http://conium.org/~maccoun/nation99.html

    * * *
    MacCoun, R., & Reuter, P. (1997). Interpreting Dutch cannabis policy: Reasoning by analogy in the legalization debate. Science, 278, 47-52.

    * * *
    MacCoun, R. J. (2001). American distortion of Dutch drug statistics. Society, 38, 23-26. http://conium.org/~maccoun/Society2001_DistortionDutchDrugStats.pdf

    * * *
    MacCoun, R., & Reuter, P. (2001). Drug war heresies: Learning from other vices, times, and places. Cambridge University Press.

  26. Pat McGroin (Irish) says:

    Ohh I thought it was Any Rand at first. Legalization might be the only way out of California’s debt crisis. Look for the feds to block it at every turn. Is that hopey changey workin’ out for ya?

  27. David Marsh says:

    Do your figures lie? Or are you a liar figuring? Modeling probabilities is like predicting the weather, it will either rain or it will not rain. The quality of the output is defined by the number of data points, the reliability of the data, the directional quality of the inquiry and the temperature of the coffee consumed the night before the project is due. What is the full spectrum of the model? When politics determines the direction of inquiry the validity of the conclusions derived from the model and subsequent policy decisions must be questioned. You have a moral obligation to the modeling process to model accurately and publish the full data set. There are serious consequences to the publishing of one sided conclusions. We here are not afraid of the data, extreme to extreme. Will it rain, or will it rain? The answer is yes, where and how much remains the question you have failed to answer. What is the best case and worst case scenario for the legalization of cannabis in your model?

  28. claygooding says:

    Mr MacCoun,since you organization does the statistical and impact studies for the drug policy,could you give us the total dollars spent since 1972 for studies done through or for NIDA on marijuana?
    And what percentage of the trillion dollars our drug warriors have spent on the war on drugs were used enforcing marijuana laws and interdiction.

  29. allan420 says:

    c’mon Rob, those are a decade old.

  30. Will legalization increase consumption? Let’s hope so.

  31. Solidad says:

    Thanks for your analysis!

    All of this pretty much ignores the fact that [at least from my perspective in the U.S. Northeast] marijuana is already “legal” in California. Anyone with a burning desire to smoke marijuana “legally” can easily find a doctor willing to issue a prescription and can buy it at a medical marijuana clinic. [Also it ignores that it is pretty much available on the illegal side too.]

    The current price of “legal” [or “illegal”] marijuana is quoted by some as about $300/oz. At this price one can get a significant high on for about the cost of a six pack of craft beer ($7.50). This high last perhaps 2 to 3 hours, whereas your six pack will last somewhat less. So “price” is not a real impediment, even now.

    Another factor that the RAND study ignores is usage rates where marijuana is pretty much legal: Holland & Portugal. These actual, versus hypothetical, trials suggest that marijuana use does not increase significantly with availability.

    With regard to driving accidents, we had dire predictions when Massachusetts “decriminalized” [i.e. made it a simple $100 fine for less than 1/oz] of a spike in casualties. However, this has not materialized. In fact there has been no change in the crime rate except for eliminating many simple “possession” arrests.

  32. BluOx says:

    ‘R’esults’A’re’N’ever’D’ubious. RAND… RIGHT!

  33. denmark says:

    This is your basic chest thumping. Now, what would be worthy of chest thumping would be the TRUTH. Too much to ask I suppose.
    Thanks Pete, good one.

  34. Pingback: There’s going to be some uncertainty - Drug WarRant

  35. Robert says:

    Use among baby boomers would go up slightly, kids might have a harder time buying it after the cartels are put out of business, but I doubt overall population use would change much. The Dutch have half the rate of marijuana use of the US despite de facto legalization. I don’t see cost being a factor, it’s still cheaper than alcohol regardless of enforcement. If there is a small increase in marijuana use, it would undoubtedly be matched by a corresponding decrease in alcohol and prescription pharmaceutical use. That’s a good thing. And that’s a fact. No crunching of arbitrarily assigned numbers needed. You’re welcome.

  36. Pingback: Why Mark Kleiman is a crock - Drug WarRant

  37. TrebleBass says:

    The purporters of the study publicly argue that few taxes would be collected and that users would double. Well, if few taxes were collected, then that would mean that big growers would not be operating, which means that the price would never drop to 38 dollars, which means that use would not rise the way they say it would. So either plenty of taxes are collected and use possibly doubles, or less taxes are collected and use does not rise that much. They shouldn’t be trying to scare people with both things at the same time.

  38. I have wrtten about this so many times I am getting kind of sick of it. It is safe to say, based on what has happened in other countries, that consumption will go up for a while, then settles at a slightly higher number than where we are today. At the same time tobacco and alcohol sales will go down. With more growers coming online EVERYDAY and some of the old growers are increasing their crop in anticipation of passing prop.19 in November, there will be a glut of cannabis, which will force the price down. I see a base amount at $80-$120 an oz based on best guesses and marketing. They want a stable price that is sustainable. After the big glut, there will be many who shut down production and just buy it, like alcohol is now sold. This is a weed that grows anywhere without needing anything but sun and water(if outdoors). It will be far easier to grow it than brewing your own in the basement, with less equipment. Eventually the price will drop further once other states come online. Compitition will be fierce. That will drive down the price. We will have much more money to spend because our taxes will be lower, that means more cash to spend. That is my peek at the future.

  39. Pingback: Rand Study: Marijuana Legalization Would Markedly Cut Mexican Drug Cartel Profits | quoteainsurance.com

Comments are closed.