Harvard professor Jeff Miron is the go-to guy for pot tax quotes. The Miron Report is the mother of all poteconomics papers. Yet the author remains detached from his subject. For Miron, weed is “just another boring agricultural thing.” He defends his conservative estimate like it’s the Alamo. Then why do we like him so much? Find out, as we subject the longstanding activist to a merciless grilling.
A.Well if we take data on use rates, from any of the existing main surveys, there’s basically two main sources. One is conducted by the federal government, the National Household Survey of Drug Use (NSDUH), and then there’s something called Monitoring the Future (MTF), which is a survey of high school seniors, conducted by the University of Michigan.
So those give use rates for marijuana – reported use rates. Of course, those may be inaccurate in some ways. But those don’t show any dramatic changes up or down. They show some changes up or down, but it’s not as if they say we went from a 10% use rate to a 50% use rate or anything like that. So I think a lot of the fluctuations, whether the movies portray it one way or the other, is changing trends fast and not a reflection of changes in the underlying use. Or at least it’s plausible that it doesn’t bear much resemblance to the changes in underlying use.
I’ve stayed with the main estimate of the size of the market that I took from a report commissioned by the White House’s drug czar’s office, back in the late 1990s, published in 2001, and I’ve just taken that number, and I’ve updated it for inflation, and I’ve looked at whether maybe I should update it for use rates. I wanted to use their number for 3 reasons: first, I read the study, and it seemed like it made reasonable assumptions; second it’s an order of magnitude that seems plausible to me in the sense of how big a market they’re talking about compared to alcohol and tobacco. There are estimates around of the marijuana market that imply that it’s way bigger than the legalized alcohol and tobacco markets, and that doesn’t strike me as plausible.
I also wanted to use a number that is basically beyond criticism, OK, in particular beyond criticism from the White House drug czar office and the DEA, or other advocates of marijuana prohibition – that’s why I used their number, because I don’t want them to be able to say ‘you deliberately inflated the size of it to make a tax revenue estimates look bigger than they’re actually going to be.’ So, I think given that if anything I erred on the low side, I’m immune to that criticism.
A.Well, I think there are a couple of things to say. One is that in some sense there already is a lot of advertising about marijuana. There’s the movies and the TV shows and the books and all that, that create obviously an awareness, and there’s all the anti-drug education that goes on in junior high schools and high schools. So every single student that goes through the schools is made aware of the existence of marijuana. They may be dissuaded; they may in fact be enticed by that anti-drug education. But there’s no scarcity of knowledge about the existence.
Second, advertising is expensive. Budweiser spends a lot of money selling something that is not much more than water. Same thing for Coke. And they get people to spend quite a bit of money on a product where the raw ingredients are not all that costly. That means they have to charge a high price, because they’re spending all the money on that advertising in addition to the amount they spend on materials. So I think that advertising of legalized marijuana would mean higher prices and that tends to discourage consumption. That also is a factor that limits rather than enhances the market.
But right now, all that advertising and publicity is free. Normally corporations that want that level of advertising and publicity would have to pay a fortune for it. Right now, it’s sort of naturally occurred. So to me, that’s a naturally occurring free boost.
Well if they’re getting a free boost now, indeed, if it became legal and there were a relatively small number of big suppliers of marijuana like Miller vs. Budweiser, or Phillip-Morris vs. RJ Reynolds, they would love for someone to ban advertising because then they could save all that money on advertising. But as long as policy allows advertising, you’re going to end up in kind of an arms race between rival suppliers of marijuana that are going to spend tons of money claiming that their product is better than their rival’s product.
That means the prices will be higher and that will limit the size of the market. That’s one of the reasons I don’t feel the price is going to go down that much, and in fact it might go up if you allow legal advertising of marijuana.
The extent to which advertising is successful at getting people to smoke marijuana, some of that might come from reduced consumption of alcohol, but that means the tax revenues for alcohol go down, while the tax revenues for marijuana go up. That’s another reason to be cautious about thinking we’re going to get huge net increases in tax revenues from legalized marijuana.
A.My estimates do take account of all that. Given the size of the market that I estimate, and the dispute across different researchers who’ve looked at the tax revenue issue, I think it comes down almost entirely to how big the market will be, not to our assumptions about what the price rate would be, not to our assumptions about how the market would adjust. It’s what do you think the size of the market is compared to how it would change?
My estimates include that there would be standard business taxes, employment taxes, income taxes and all that, and we would collect a fraction of that that is equal, or a bit higher, than the fraction we collect from existing legalized industries. So 30% is a nice round number representing what tax revenue is, relative to economic activity, relative to the GDP. And I assume we have a special tax, an excise tax or a sin tax.
But I am taking account of all those factors. And I still think that they don’t add up to huge numbers, because I think that’s because it’s one product. If we look at the tax revenue we get from any product, no matter how popular, it’s maybe, maybe, tens of billions. But it’s not hundreds of billions.
A.No! I think we’re being from a tax revenue perspective a bit hopeful that somehow this is sort of a miracle commodity that’s going to overtake everything else. For example, you hear this statement all the time, that marijuana is the largest cash crop in California. I challenge anybody to discover the source of that statement. I’ve been hearing that statement for decades, and I bet whoever made that statement had zero data to back it up.
But even if there is, there are many many crops. And whether it’s a #1 cash crop or not is not the crucial issue, there’s wheat and corn and soybeans and all that stuff, and it just can’t be that big a deal cause it’s only one of zillions of things that any economy produces.
A.Oh, well, they sort of did the estimates in a slightly different way. Because I started with a dollar figure estimated by the ONDCP’s report, which in fact had been built up on data about grams, ounces, tons, etc. I never needed to use the data directly on the physical number of units.
I just had an estimate of the size of the market measured in dollars, then, using sort of standard microeconomics about the standardness of demand for product, as a function of the change in the price, could ask how much the price of the market could change. If the price went down, how much the revenue would be, if a certain tax were imposed then a certain tax rate. So those metric ton numbers are implicit in the approach that I’ve used. But given the source that I started with, I didn’t need to talk about them any further.
When you’re looking at the level of the country as a whole, it seems to be possible to have 50, 60, 80% sin tax rates, without driving the market underground. That might be a bit surprising. It’s certainly an incentive for the underground market to develop if that rate gets high, but it also appears people want to buy from a legal supplier, and therefore you can impose a reasonably high sin tax without driving the market underground. That’s what experience from alcohol and tobacco seems to suggest.
A.Well, I could take the point that you just made, maybe people in an informal way view alcohol as having a therapeutic value, not necessarily because it’s preventing heart attacks, but just because it relaxes you, and makes it easier for you to enjoy a pleasant evening after a hard day at work and things like that.
Second, for marijuana users, I think there’s a broad range of the way in which people are using it therapeutically. In some cases, it’s clearly very similar to the way people use prescription pharmaceuticals for a very special condition with extreme sort of negative effects. In other cases, I think it is very much like a martini or a glass of wine at the end of the day. It’s helpful for relaxation, for calming, for putting you in a better mood. And so I think the distinction is not so clear-cut. In any case, I don’t see why that would affect the nature of the calculation.
The way economists think about it, at least, is what matters is your demand for a good. If your demand for a good is there, then you’re going to want to purchase it if it’s more convenient to buy it from a legal supplier. If you get more reliability from a legal supplier, then that’s going to make you willing to pay some sin tax, in order to have that convenience.
Now if it’s right that some sin tax would keep some of the market underground, that’s one more reason for us to think that the tax revenue we’re going to collect is moderate, not huge. If a lot of the market stays underground, then we don’t collect nearly as much in tax revenue, and I think that’s another reason to be cautious.
A.I don’t think it’s a dichotomy. But there’s another side of it, which is, whatever the amount is, whether the tax revenue for the country is $10 billion, or $50 billion or $500 billion – whatever number you want to assume for a moment. None of that should be the main reason to legalize marijuana. The crucial reason to legalize marijuana should be: why not? Why should the government be prohibiting something, harassing people, jailing people, for producing, consuming, selling something, unless there’s a compelling case that in doing any of those things people are harming innocent third parties. That’s the attitude toward alcohol that the country accepts.
So we certainly try to prevent people from using alcohol in ways that harm others such as driving a car when you’re under the influence. But we don’t criminalize the whole industry. We don’t make it illegal to produce it, sell it or consume it. We accept that the vast majority of people use alcohol in a way that is mildly beneficial, maybe somewhat beneficial if the heart benefits are substantial, but in any case, for the most part, not harmful to themselves or others.
And that’s why it’s a legal good and we only go after the irresponsible use. That should be the attitude, and that’s the reason to legalize marijuana even if it didn’t give us a single dollar in extra tax revenue or save us a single dollar of extra expenditure. It should still be legal for reasons of freedom and liberty, never mind any of these sort of economic arguments. So I don’t want what I think are the compelling reasons, the liberty arguments, to wind up being undercut by inflated, excessive, and possibly contaminated estimates of the economic benefits.
I think that makes the movement seem somewhat disingenuous. We’re kind of trying to sell people on this being a free lunch. It’s not a free lunch. Some people will misuse marijuana. Of course many of them are already doing that. I think the increase in people misusing marijuana will be very small. But we shouldn’t think that we’re getting some great big goose-that-laid-the-golden-egg benefit out of this. The benefit is letting people consume marijuana in peace.
Q With all that said – and I heard every word you said, you’re preaching to the choir – is there a magic number that could induce politicians to get off the fence? Is there a magic number that if you were a Republican from a red state you could get past the Nancy Reagan programming that you have? Everything has its price. Is there a magic number?
A.I have no idea if there’s a magic number. For some politicians that are relatively liberal, they represent liberal districts, they don’t necessarily need a big number, and their constituencies are with them already. For people who are in hard corps red states or red districts, I think there’s lots of fiscal conservatives who couldn’t care less whether the number is $10 billion, $50 billion, $100 billion. They think it’s evil, and they won’t put a price of anything evil.
There clearly are people in the middle, there are politicians in districts which have more of a mix, and I certainly talk to a lot of politicians who say, ‘Look, I need cover. I agree with you philosophically or morally, but I need an excuse to tell my constituents that I’m going to vote for this.’ And so yes, for some politicians, a big number maybe is helpful. But I don’t think that’s a reason to make the number bigger than my objection analysis of the economics suggests.
A.I think that for the Democrats to take on this issue, especially outside of California, is risky, because they want people in the middle, and if you think of independents broadly, they’re a very mixed group. I mean some Democrats are sort of Libertarians who would be great with this issue. But many independents are pretty socially moderate, maybe socially conservative, even if they’re not sort of hard corps religious right.
So I think that any Democrat in moderate states – the purple, instead of red versus blue – they would feel very nervous about this issue. Now they don’t have to address it yet. For the moment, it’s basically a California issue with the legalization initiative that’s on the ballot in November. But if I were a Democrat, I would just hope that this issue would go away.
A.There’s a couple of reasons for the difference. First, I’m using a demand side estimate. There’s data from surveys on the amount that people use. He’s using that data and data from the supply side estimates of the amount seized, combined with some assumptions about what fraction of the market gets seized in a typical year – and those numbers turn out to be higher.
I think that the supply side numbers are very, very suspect, because they come from law enforcement, and law enforcement has a strong incentive to want to make it look as though they’re being very effective. They like to show up on the nightly news saying we seized 100 million pounds of marijuana. Frequently, when they estimate the weight, they’re looking at the entire plant, not just at the smokable part, so the plant is way big in terms of weight and they‘re looking at the wet weight, which is less relevant than the dry weight for estimating the size of the market.
So I don’t use the supply side stuff, and I think one should be very cautious on the supply side. When you get the demand side stuff, you get to a point where you have data on a number of users who say to the surveys “I am a daily user of marijuana.” And then you have to come up with an idea of how much daily use means. Does it mean smoking a puff a day, a joint a day, five joints a day?
I mean you have to make an assumption because you don’t have that data from those users. And, again, if I look at those assumptions that they’ve made, my guess is that many of the people who say they’re daily users are consuming way less than the amount they’re assuming is being consumed on average. So the demand-side estimates are higher than what seems plausible to me.
A.. . . from any commodity, there still are limits. If we add more and more kinds of taxes, that directly goes to the concern that we’re going to drive the market back underground. I get criticized from both ends. When I say we’re going to have this legalized industry and collect taxes, bunches of people look at that and say, “oh, everyone’s that producing illegally is just going to stay underground,” in which case there’s no tax revenue. So I think the right answer is in between, as long as the taxation is moderate, it will come mostly above ground.
Of course there are still some individuals who will grow in their basements or their back yards or wherever. But just saying that we could impose all these kinds of taxes does start to beg the question of how much can you tax it before you drive it underground again?
A.You’re not confined or restrained to not do new things. What you’re constrained by, if you want to still be an economist, is to use the methodology that economists accept. Meaning, in terms of how big the marijuana would be, we look at basically the degree in which changes in price affect the amount consumed. We try to look at how much the price is likely to go down based on evidence from comparing different places with different marijuana rations, and so our forecast of what the marijuana market would be is based on that, not on some hypothesis that if it were legal, there might be five times as many users.
Our hypothesis, our view is, our framework is the change in the amount consumed, the change in the number of users would be governed by the change in the price, OK, and that seems to imply, based on the kinds of data that we’ve looked at for other products, relatively moderate changes, not drastic changes.
A.First of all if he’d been in my class, I would have been discussing his progress with him all along the way, and I would have cautioned him that I thought his higher numbers were incredibly implausible, and asked him to do more sort of robustness checks and more investigation into the supply side numbers that underlie his big estimates.
And I would challenge him to go back and find the original sources of these amounts seized, and these estimates that come from various sources that he uses, and I’m pretty sure what he’d find out was that there was never any data underlying some assertions about the size of the market, it was just some DEA or other official saying ‘I think the market will be this big,’ based on nothing. Therefore I would tell him you should put relatively little weight on that. I guess I don’t want to say what grade I would have given him, because it would have been a different experience.
A.I don’t think there’s a lot of different theory. What’s in there is a lot of data and a lot of work. I mean there’s a lot of merit in his thesis, but I don’t think there’s a lot of merit in the numbers. He didn’t get instilled in him a healthy enough skepticism of data. I mean any sort of data that one hears about even less controversial things should make one ask, “How did somebody actually figure out that number? How could anybody actually know that number?”
And there’s an example related to the book Freakeconomics, which makes a claim that the wage of architects are lower than the wages of prostitutes or something like that. So an economist went out to try to check that, an economist name John DiNardo from the University of Michigan. And he basically found that he couldn’t get any data anywhere on the wages of architects or the wages of prostitutes. So how can anyone know which wages were higher or lower, if there were no decent on either of the two pieces? So the skepticism about data, that’s what was missing from Max’s thesis.
A.I think his figures are too high. He uses both demand side and supply side. The demand side pretty consistently comes out lower than the supply side. He takes an average. That’s why he gets a number that’s sort of in between my estimates and Chaiken’s estimates. So Chaiken is kind of extreme supply side, almost all the way on the supply side, Gettman is putting weight on an average of the two approaches, recalling that I’m using only the demand side, which always tends to be the lowest.
So that’s how those three wind up being in the way that they’re ordered. Again, I won’t use the supply side at all, that’s part of why I end up with a lower number. On the demand side, I think that the amount assumed per use, for each person who shows up in the demand side estimates, is probably too big. So that’s the difference between mine and Gettman’s, focusing only on the demand side. And there’s some more sort of details, some sort of wonky things that I don’t remember all the details of, but, basically, that’s the crucial number.
If you look at the California Board of Equalization estimates, it comes down to the same thing. They’re basically using a demand side approach. They have data on frequency of use broken down by daily users and annual users. They focus on people who are regular users, and they have an assumption about how much is being consumed by a daily user. But we don’t know what a daily user is. Is a daily user someone who has a puff or two before going to sleep each night, is it someone who consumes all day, every day, and that’s what daily use means?
And so, at a minimum, without knowing whether daily users are in one category or mainly in another category, you can clearly get a really big range. But they in particular assume an amount, which is sort of unlikely to be the case.
A.Sorry. Elasticity is how much the quantity demanded responds to changes in the price. What we find for most commodities is that when the price goes up, the quantity consumed for purchase goes down, and in this case we’re talking about a decline in the price from legalization. Exactly how big a decline we don’t know for sure, but we have some information.
In some ways, I’ve actually assumed things that would make my tax estimates even smaller, because depending on what that elasticity is, we may get different numbers than I came up with. So I have to make assumptions too. Lots of assumptions, but I try to minimize them, and I certainly tried to err, if I was erring, on the cautious side, but I didn’t make that kind of choice where I thought I was erring by a factor of ten to one. I made the choice that maybe I was erring by a factor of 10 to 20%.
Take the underreporting of use by people responding to surveys. One can easily imagine that if you pick up the phone and someone says, “I’m from the federal government and I want to ask you how often you use marijuana,” some people might say never, when, in fact, the truth is that they do. A few people may actually lie in the other direction and claim they use it when they don’t, who knows, but it’s perfectly plausible that use is underreported in these surveys.
But we have a lot of studies that have compared survey measures of use to objective measures of use, and have found that the surveys are low by 10 or 20%, not by 50% or 80 or 90%. Sometimes you can ask people if they’ve used drugs, but you’ve also given them a urinalysis test. So you can check if they’ve used drugs and compare it against what they said. So the degree of honesty is actually sort of surprising. It’s not 100%, but it’s not dramatically less than 100% either.
So I don’t make an adjustment for that. If you said that the ONDCP numbers are underreported by 10%, that would imply a bigger number. But that’s $11 billion instead of $12 billion, it’s not a radical change. It’s not getting me to $100 billion or $300 billion.
A.It’s nothing to do with creativity. It has to do with accuracy. So let’s take an example. Tons of cities decide to subsidize baseball stadiums, football stadiums, convention centers, and so on and so forth. They also commission some local consulting firm to do an economic impact statement. Lo and behold, because those studies were actually commissioned by the group that wanted to get the stadium built, they always wind up implying that there’s some sort of unbelievable benefit from building a new baseball stadium in Boston.
Now in the case of a state, there is an argument that there might be some significant benefit, which is maybe the city of Boston has a much better Fenway park which has 50,000 seats instead of 35,000 seats, we’ll get more fans from Rhode Island, from Maine, or some Yankees fans from New York. So there will be money coming into Massachusetts from out of state. On the tickets, on the hotels, on the meals, etc. And it turns out that zillions of economists have evaluated the benefits of adding stadiums, adding convention centers, and things like that. They universally find that there’s almost no net gain and often a big loss. Why?
Because first of all, almost anyone who goes to Fenway Park is already a Massachusetts resident. Second of all, there’s tons of double counting in these studies. When Massachusetts, if it gets more people to come and go to baseball games, most of all, it’s because those people weren’t going to a football game or weren’t going to a movie or weren’t going to NASCAR or something else. So it’s just a shift in their dollars. For the country as a whole, if Alaska has cannabis tourism, that’s just money that’s not being spent on some other form of tourism, which almost for the most part was going to be spent within the United States, so there’s no net benefit. It’s just shifting dollars.
A.I would say what I sort of hinted at earlier. I think we have a better chance of convincing the country to change policy, and staying that way over the long haul, if we focus on the arguments other than the budgetary, taxation arguments. I think they’re item number 6, 7, or 8 on the list of reasons. The crucial reasons are enhanced welfare of current marijuana users. The second reason is the enhanced welfare of people who would like to use marijuana but who are abstaining because they’re worried about the legal penalty. The third crucial benefit is reduced crime, reduced corruption. Also increased civil liberties, because people don’t have to be worried about punishment because of the color of their skin, not being afraid of having their door knocked down in the middle of the night and things like that.
And so, the fact that we will spend less money on police and prosecutors is certainly a plus, the fact that we collect some tax revenues is certainly a plus. But to me, these other arguments that I just mentioned are even more important. And those are the ones that we should focus on when arguing for legalized marijuana. So that’s my message.