Sunday, February 28, 2016

Overpenalization of Drug Use


In this post I’ll examine the economic consequences of “overpenalizing” drug use. Refer back to my previous post (here) for the starting point of this discussion. This post is a continuation of the ideas in that one.

If we embark upon a policy of drug prohibition, we risk penalizing a harmless activity. The public’s views of drug-related dangers are distorted, and mistaken or cynical politicians respond to these distorted perceptions of risk. The average voter believes that harmful drugs are far more dangerous than they actually are, and they wrongly believe that some harmless drugs are harmful. I won’t specify which drugs have exaggerated harms or describe in detail the degree of the public’s distorted perceptions. We need not agree on the particulars. We can still agree that there is an enormous risk of harm if we overpenalize.

Supply and Demand Curves

The graph below shows the supply and demand curves for drugs with exaggerated harms. The marginal private benefit is actually equal to the marginal social benefit, but the voting public mistakenly believes there are big social costs to drug use. The dotted black line represents the public’s distorted perception of drug-related harms; a fully informed public would see this line as overlapping the solid black line. (Reiterating a point from the previous post: The red line, the “marginal social cost,” is the supply curve. It represents the costs of supplying the drugs to the consuming public.)

The World As It Should Be

The chart below spells out the welfare implications if policy is as it should be. If we don’t penalize drug use (as we shouldn’t in this example), total welfare is simply the areas of A’ + B’. Triangle A’ is the consumer’s surplus, and triangle B’ is the producer’s surplus. You can take this chart to represent a world where the drug really doesn’t have any externalities, or you can take it to represent a world where a drug with externalities is appropriately penalized with an ideal tax. The welfare implications are the same. (See previous post for details.) The next section considers adding an *additional* penalty on top of that.

The World As It Is

Now we look at a world where the public irrationally overpenalizes drug use, which is much closer to the world we live in. There is a prohibition regime, and the MPB is lowered so it overlaps the (public’s misperceived) MSB. The consumer and producer surpluses have both shrunk. The area C represents the penalty applied to drug users. It is a pure loss. Drug users are harassed (jailed, imprisoned, beaten, etc.), and there is no offsetting benefit here, as there would be if drug users were taxed instead. Triangle D is a deadweight loss associated with the loss of consumer and producer surplus. Total welfare in this case is A + B. If we hadn’t penalized drug use, the total welfare would have been the full area A + B + C + D (A’ + B’ in the graph above). Society loses D + C. There is quite a lot to lose from overpenalizing drug use.

(Edited 3/1/2016. I had made an error in my initial analysis, in which C was lost twice. I should have suspected an error. I believe it's correct now.)

If we’d over-taxed rather than overpenalized, it wouldn’t be quite so bad.

Overtaxing vs Overpenalizing

Suppose we deter the unwanted drug use with a tax rather than a penalty. The welfare consequences aren’t *quite* as bad, because in this case somebody at least collects the taxes paid by the drug users. Total welfare is A + B + C. Area C is tricky. Drug users enjoy a consumer surplus equal to C, but they pay a tax equal to it's value, but society collects that tax. It's a surplus, plus a pure transfer. So C - C + C = C. Triangle D is still a deadweight loss. So social welfare is A + B + C + D in an untaxed world, but A + B + C in a taxed world. The loss is D. It’s not as bad as losing C + D, but still unfortunate.

(I've edited the paragraph above to correct my initial over-counting. I initially thought that society lost C in the overtaxing regime, but I believe I was mistaken.)

The Cost, In Plain English

Whether we deter drug use with a tax or a penalty, the implications are the same (in kind if not in magnitude). Some thrill-seekers are denied a harmless pleasure for no good reason (the loss of consumer surplus). And some perfectly respectable jobs and businesses are destroyed, again for no good reason (the loss of producer surplus). Suppose you can’t bring yourself to care about the interests of drug users or producers, *even though* I’m specifying that we’ve misidentified a harmless drug as harmless. Fine, then imagine for a moment that we’re talking about penalizing, say, country music or video games or ceramic Siamese cats, or whatever guilty pleasure you happen to indulge. Surely you understand that the government shouldn’t go around deterring popular-yet-controversial hobbies willy-nilly.

When we evaluate drug policy, we need to factor in this potential cost of overpenalizing. If we misidentify a harmless drug as harmful, that might sound like an inexpensive error. One might dismiss it with a “better safe than sorry.” But it isn’t free, and it isn’t even cheap. Some people will continue to use the banned drug, and those remaining users still face the stiff, dehumanizing penalties. The people who are successfully deterred are denied a cheap thrill. I think it’s benighted to treat these costs as trivial. To blow this off is to blow off every encroachment of government into our private affairs. Policy makers need to fully consider the possibility that they have exaggerated a problem, and the costs of an inappropriate policy response needs to be given the same consideration as any other risk. Government overreaction is a serious problem. A bit of epistemic humility would temper it a great deal.

Saturday, February 27, 2016

Drug Prohibition or Drug Taxation?

Intro to Supply and Demand Curves

I’ll formalize the economics of drug prohibition in this post. First a review of some concepts (or an intro if you’ve never taken any economics).

With price on the y-axis and quantity on the x-axis, there is a downward-sloping demand curve and an upward sloping supply curve. The slope of the demand curve indicates that people become satiated as they consume more and more; this is true for *any* kind of good or service. The very first few “units” are worth quite a lot; imagine the most devoted uber-fans at the very front of the line, who will derive the greatest enjoyment from the first few units produced. As more is consumed, desires are satisfied, and at some point additional units made available for consumption bring negligible pleasure to the consumers. The upward-sloping supply curve represents the costs of production, which should increase as more is produced. For the first few units, you can use the very most productive fields (for a product that is farmed) or the most productive machinery or the most skilled and dedicated workers. As those factors of production get employed, producers start reaching for less productive means, and at some point it becomes too costly to employ any more resources in production.

These curves intersect at a point that marks the market price and the quantity produced, the y and x coordinates respectively. The area between the supply curve and the price line represents the “producer’s surplus” and the area between the price line and the demand curve represents the “consumer’s surplus.” Both areas summed together represent the “social benefit” of that particular good. People can and do argue about whether this is a valid approach to judging social welfare, but it’s a good starting point for analysis. The only alternative is inane blather. Let’s at least start with some formalism. (MSC = Marginal social cost, MSB = Marginal social benefit)

Considering Costs to Third Parties

When discussing something like drug use or pollution, there are potential harms to third parties (“externalities” in econospeak) that need to be considered. The idea here is that someone who uses illegal drugs doesn’t bear the full cost of his actions. In this case there is a private benefit and a social benefit. The social benefit is *lower* than the private benefit for any given quantity of drug use. This is bad, because it causes people to consume more than the socially optimal amount. The quantity consumed will be at the intersection of the private benefit and social cost curve, when we’d like it to be at the intersection of the *social* cost and *social* benefit curve.

If you are a paternalist, meaning you like the idea of protecting people from themselves, this analysis is still the right approach. You can think of a person as having multiple “selves.” One such self really wants to use drugs and thinks the costs are worth the benefits, and the other self thinks the first guy is nuts and would rather not be injured by him. The divergence between the social benefit and the private benefit can represent this self-harm, or it can strictly speaking represent the costs imposed on third parties, or it can be some combination. I will use the same graphs in any case.

If everyone is free to choose whatever level of drug use they please, assuming there are significant externalities from drug use, there will be too much of it. So presumably government policy can curb this excess use. There are two ways to do this. I can tax you, or I can beat you. I’ll call these options the “tax” and “prohibition” regimes. In either case, the private benefit curve drops. If consuming drugs comes with a beating or a hefty tax, it becomes less pleasant to consume for any given quantity, so this shifts the demand curve down. Ideally, you would set the tax or the severity of the bludgeoning *just right* to push the private cost curve to perfectly overlap the social cost curve. (Of course, prohibition doesn’t involve literal beatings…usually. I’m using “beating” as short-hand for all the other nasty things we do to people who use illegal drugs, such as imprisonment, social shaming, harassment, removal of children, and, yes, the occasional beating.)

Another approach is to shift the supply curve, which is the effect of targeting drug suppliers. Supply-side policies actually represent the bulk of drug control efforts, at least in the US. But I'm not going to treat them in this post. I'm rather taking the supply curve as a given and asking what happens when you target the demand side. 

Social Welfare Consequences of Drug Policy

We can determine the welfare consequences of different policy responses by drawing areas on the supply and demand curves. (“Welfare” means something like “total summed benefit to society.” It has nothing to do with the word's more colloquial meaning, being programs that help the poor or transfer payment programs. Incidentally, those programs can and should also be treated with an analysis like this one.) The chart below shows the social cost curve (the supply curve) and the private and social benefit curves (the demand curves). Triangles A and B are the consumer’s surplus and the producer’s surplus, as described above, for the scenario when drug consumption is at the social optimum. The quadrilateral C represents a “transfer”; drug users benefit from consuming more, and producers benefit from the higher price, but their gains are society’s loss. Without getting into a theological discussion of whose interests and benefits should count, let’s just call C a “wash” for now. Triangle D represents the true social cost of drug use. (Keep in mind this formalism works for any externality; the graph would look the same if I were examining air pollution.) Triangle D is the area between the marginal social cost curve and the marginal social benefit curve, integrated over the range that separates the social optimum and the private optimum. In the case of *no* drug policy, the total social welfare is equal to A + B + C – C – D. The C cancels, so I’m left with A + B – D.

If I *tax* drug use, and if it’s an optimal tax that moves the private benefit curve down to the social benefit curve, then I end up with something like the very first graph in this post. Consumption drops, and so does the price. C disappears for both the consumers and the producers of drugs, so they lose out on this benefit, but society collects C worth of taxes. The “losses” to users and producers of drugs are society’s “gain.” And the deadweight loss in triangle D goes away. So we end up with a total social welfare of A + B. In this imaginary scenario, society is better off by an amount D. (We should keep in mind that setting the optimal tax is *not* so easy, and in fact many drugs that are relatively harmless have been given a bad name by an overzealous media and an over-credulous society. So it’s worth thinking about the harm done in a scenario where the social and private benefits are *actually* the same, but the public believes they diverge. In that case you would punish a perfectly harmless hobby, and you’d end up with a net loss of social welfare. This surely describes what’s happening for at least some illegal substances. That might be a good exercise for a future post.)

A prohibition regime, on the other hand, has nastier welfare implications. We still shift the private benefit curve down to overlap social benefit curve, but in this case no tax is collected. The full area of C is lost, and this loss does not become “society’s” gain. Unless the agents we employ to harass drug users are particularly gleeful sadists, the harm imposed on drug users doesn’t get offset by a corresponding benefit to some other party. It is simply lost. The deadweight loss D is also gone, but the cost of eliminating it is unnecessarily high. Total social welfare in this scenario is A + B – C. If the area in C is actually greater than D (which depends on the exact shape of the supply and demand curves), then it’s possible that prohibition is better than *no* drug policy. But it’s strictly worse than a tax. 

This approach might seem odd for a number of reasons. Should we really consider the benefits to drug producers and drug consumers in our welfare calculation? Shouldn’t the social optimum quantity be *zero*, at least for certain drugs? To the first question, I’d say we’re comparing *different* regimes, so the total social welfare doesn’t matter. What matters is the difference between, say, laissez-faire and prohibition (a difference represented by C – D), or between prohibition and taxation (represented by C). Likewise, if I say, “The welfare transfer C from society to drug consumers and producers is immoral. I’m not going to count it as a benefit in my social welfare calculation.” That’s fine. In this case the welfare difference between taxation and prohibition is *still* C, because you still have in one case a tax that is collected and in another a punishment that is not “collected” by anybody.  

Edit 2/29/2016: I believe I made a minor goof above. Area C should be split into two sections. Split it by drawing  a vertical line upward from the price-quantity intersection point (at roughly (8,7.5) on my graph) to the MPB (solid black line). One section, a parallelogram, is a transfer (under a tax policy) or a pure loss (under a prohibition policy). The other section, a triangle, is a loss of consumer and producer surplus. I'll show what this looks like in a later post. I think this is a relatively minor correction.

To the second question, chart where the social optimum is *zero* drug use looks like the chart below. There’s no “A” or “B”. Of course in this case, it doesn’t actually matter whether you drive the private benefit curve down to the private benefit curve with a tax or a punishment. No tax will be collected nor will any punishment *actually* be imposed if consumption is literally zero. But such implausible scenarios aren’t terribly interesting. 

Other Considerations

There are other practical and moral considerations here. I’m not pretending that this approach to cost-benefit analysis “solves” the problem and tells us the ideal policy regime. As a practical matter, it’s actually quite difficult to deter drug use. Demand for drugs is quite inelastic (discussed here). Users will pay very high prices, so it’s difficult to deter them with either a tax or a penalty. If you believe the estimates for the number of heroin users from the SAMHSA surveys, and if you believe the CDC’s estimates for the number of heroin overdoses, you get something like a 1.5 – 2.5% mortality rate *per year* for heroin users, just from overdoses. (Depending on which year you look at.) This is huge. When economists and actuaries monetize the risk of death, they get values in the range of $50,000 to $120,000 for a 1% chance of death. Basically, these are the sums people will pay to avoid a 1% chance of death. So heroin under the current regime faces a very heavy implicit penalty, and we still have just shy of a half million users in this country. You’d have to resort to inhuman levels of cruelty (err…more so) to deter heroin use any further. If you look at drug use surveys that span several decades, you see up and down movements of usage rates that have nothing to do with official policy. So it’s pretty hard to claim we’ve achieved a significant level of deterrence. The deterrent effect of drug prohibition is surely real, but it’s more like a rounding error than a driving force.

There are moral considerations, too, and these are very important. But this post is already long. I won’t delve into theology right now. I’ll just say that there are no moral trump cards. There are still trade-offs, even if they are moral trade-offs. If we have more cocaine users under laissez-faire drug policy, but we have more gang members murdering each other under prohibition, we might consider both of these costs to be on a higher moral plane than mere monetary considerations. Maybe we consider cocaine use and murder “just evil” and refuse to monetize them, and we stoutly refuse to compare them to the frivolous pleasure reaped by drug users. But this still presents a trade-off to be managed.

Thursday, February 25, 2016

Demand Response to Vice Prohibition

I’ve made the point before (here and here) that drug prohibition is a losing battle. In order to deter drug use, you have to raise the cost faced by the consumer of drugs. In this post I want to make a simple scaling argument. My conclusion will be that the costs of prohibition most likely scale faster than the benefits.

Supposedly drug users irrationally harm themselves with their nasty habits, so we have to stop them. That’s a typical rationale for drug prohibition: People don’t fully account for the cost of their vices, so it would be good to stop them from indulging. But then these supposedly irrational drug users are presumed to respond rationally to legal sanctions. It’s akin to believing, “People don’t respond rationally to the [pharmacological] risks of drug use, but they do respond rationally to the [legal] risks of drug use.” Say the sentence without the words in square brackets to see the obvious contradiction. Maybe we can be more precise than this hand-waving.

Let’s suppose a potential drug consumer faces a price “X”. This X is the full economic cost to the user. It factors in the money paid to obtain the drugs, the time spent acquiring and using them, the potential health risks of using, the potential shame and social sanctions if his use is discovered, potential harassment by police, jailing, prison, etc. You can express X in any units. “Dollars” would be the most obvious: monetize the time value and the health and legal risks and roll it all up into a clean dollar figure. Or you can express it in “man-hours of labor,” or “% risk of completely ruining your life,” or “expected years (or months or days or hours) spent in misery.” It doesn’t matter for our purposes. I’m not going to try to estimate what this cost is, or even specify which drug I’m talking about. What matters is that you have this price called X, and to get you to stop using drugs I have to increase X. This is how drug prohibition works.

How strongly do you react when I raise X? Maybe there is a simple inverse relationship, whereby if I increase X by some factor Z, rates of drug use drop by a factor of 1/Z. So, for example, if I double the cost, I cut the amount of drug use in half. Taking that example, we have half as many users, but they are all paying twice the cost! It’s a wash! (By the way, I am leaving out *many* of the other costs of drug prohibition.) The chart below shows what this relationship looks like (simple inverse-relationship chart here). It is impossible to achieve “success” in this world. Hammer the users as hard as you like. The total cost will stay flat. This is a losing game, especially when you factor in the cost of waging a drug war.

(The “quantity consumed” is in arbitrary units. It doesn’t matter whether “5” means “5 metric tons per year” or “5 million users,” and again it doesn’t matter what drug I’m talking about or what the real-world value is. I’m simply making a scaling argument here, so the actual values don’t matter. These are *not* supply-and-demand curves from economics 101, which traditionally cross and tell you the market price and quantity supplied. Although the “Quantity Consumed” curve *is* actually a demand curve. If it’s confusing that I’m recycling the y-axis for three different kinds of values, don’t get hung up on it. My point is to show how total costs scale up with the price faced by the user.)

But maybe there’s reason for hope! Maybe it’s not a 1/Z relationship, but more like a 1/Z^2 or a 1/Z^10 relationship between price and total use. If this is the case, we do added harm to the remaining users, but those remaining users drop off quickly. Quickly enough that the total cost falls. Anything with a 1/Z^(anything greater than 1.0) relationship will show this falling cost pattern. The chart below shows what a 1/Z^1.5 relationship looks like.

Success! The cost falls when you hammer drug users with harsher penalties. (Once again I’m ignoring a whole host of other relevant costs.)

Is the above scenario reasonable? Not really. It assumes that drug demand is fairly elastic. “Demand elasticity” is usually expressed as a (% change in demand) / (% change in price). If you increase the price by 1% and demand falls by 2%, the elasticity is 2.0 (demand is fairly elastic). If you increase the price by 1% and demand falls by 0.5%, the elasticity is 0.5 (demand is fairly inelastic). Very roughly, inelastic demand means people are willing to pay quite a bit more to keep buying the product; elastic demand means that people buy quite a bit less when the price rises significantly.
Demand for drugs is actually quite *inelastic*. You can raise the price by a lot and you get a small demand response. Measured elasticities for common addictive substances (tobacco, alcohol, heroin, cocaine) are typically in the -0.3 range. So the relationship between cost and quantity probably follows more like a 1/Z^0.5 (anyway, 1/Z^(something-between-0-and-1) relationship, probably closer to 0 than 1). A 1/Z^0.5 relationship gives you a picture like the following:

The total cost *increases* as you increase X. There are fewer users, but they each face a much greater cost. Doubling the price doesn’t quite cut the using population in half, so you have *more than* half as many users each facing twice the cost. Here prohibition is clearly a losing game. It does you no good to hammer the users with stiffer sanctions because the costs scale up *faster* than the benefits.

Quite a lot is left out of this analysis. Harm to third parties is ignored, but then many of the costs of *prohibition* (as opposed to the costs imposed by the users themselves) are born by third parties. It’s not at all clear that accounting for the costs to third parties swings the analysis in favor of prohibition. Indeed, including those costs would most likely strengthen the case *against* prohibition. Also consider that many of those “costs to third parties” involve behaviors that are already sanctioned. Economic crimes to acquire drug money, pharmacologically induced violence (which despite a few scare stories is grotesquely exaggerated by the popular media), child neglect, and driving under the influence are already criminalized. If the goal is to deter these behaviors, we should crack down on them specifically. It makes no sense to focus our law enforcement resources on a behavior that kinda sorta sometimes leads to these other social problems. Address those problems directly if they are problems worth solving. If someone wants to do a more thorough accounting of these third-party costs, I’d love to see it. But please don’t dismiss my argument because I left out this consideration.

I want to clarify that my above argument does *not* apply to all crimes, but it does apply to all crimes in which the criminal harms primarily himself.  The prohibition approach makes the following deal with such “criminal”: “I’m going to harm you if you indulge in self-harming behavior X.” The vice-criminal already faces an implicit tax on the vice, in that such a vice has intrinsic harms. ("Vice-criminal" contains a wonderful double-meaning; a vice-criminal is not really a criminal, in the same way that the vice president is not really the president.) A real criminal is someone who primarily harms a third party, but whose crime imposes minimal cost/risk to himself. Sanctioning a crime raises the price from something negligible to something large. Sanctioning a vice raises the price from “already substantial” to “somewhat more substantial.” If you’re trying to deter a vice with inelastic demand, it’s hard to gain any ground. We should approach vices with “harm reduction,” not a prohibition approach. I don’t’ doubt that there are real harms associated with many vices.  These are costs to be managed, not problems to be “solved.”

Monday, February 22, 2016

Drug Prohibition Requires Implausible Assumptions to Work

The notion of deterring drug users with criminal sanctions is quite silly. Of course most drug policy actually concentrates on the supply side rather than demand side, but the argument I’m about to make applies to both kinds of drug control policy. Supply interdiction raises the cost of obtaining drugs by raising the price the buyer pays for those drugs (both the monetary price and the time “price” imposed by searching for a product in a black market). Sanctions against the users themselves increases the price by threatening users with police harassment or jail time (possibly including violent forced-entry raids on residences and the removal of children from a parent’s care).

Consider the price that a user pays for their drugs, first in the absence and then in the presence of prohibition. This will be an all-inclusive price, not just the money paid for the drugs. It will include the money, the time spent acquiring and using the drugs, and the various risks inherent to drug use (overdose, social disapproval, risk of losing control and committing a crime under the influence, etc.). Measure it in whatever units you like for whatever quantity of drugs you wish to consider, and call it X. Maybe X is “a 1% chance of completely ruining your life” or “20 hours of labor” or “$1,000”. It doesn’t really matter what units we’re using; we just want some number that we can compare to some other number. X will stand for the price of drugs in the absence of prohibition, and Y will stand for the price of drugs in a regime of fairly strict drug prohibition (as the United States currently has).

Y is greater than X. That is the point. That is how drug prohibition works. Increase the cost of using drugs, and (presumably) people use less of them. But that is not nearly enough. You must weigh the “benefits” of decreased drug use against the costs of the drug prohibition regime. That is what goes missing in the vast majority of commentaries on drug prohibition. It doesn’t count as success to merely decrease the number of users. To declare “success” you must achieve this end (with quantified and well-accounted for benefits of such a reduction) at an acceptable cost. I’m arguing here that “success” is implausible.

Suppose a not-unreasonable relationship between price and consumption such that doubling X cuts in half the amount of drug use. (And a tripling cuts it to a third, and a 1.2-fold increase cuts it to 1/1.2, and a tenfold increase cuts it to 1/10, and a Z-fold increase cuts it to 1/Z.) A prohibition regime increases the price from X to 2X and cuts the number of users in half. But those users each bear twice the cost, so the total harm is the same. (Actually greater, because we haven’t factored in the cost of the resources used to fight the drug war.) It doesn’t do you any good to say, “Well, increase the cost the user faces to 3X, or 10X!” You get the same answer: less drug use, but a much greater cost per user. This is a dubious goal anyway, considering that most people are actually risk averse. (Risk aversion means we’d rather face a high probability of a smaller cost than a smaller probability of a larger cost, assuming the average cost stays the same in both options.)

Maybe the scaling relationship is different, but it can’t be all that radically different. It might be a useful exercise to see how much the cost-benefit trade-off changes if, say, a Z-fold increase in price results in a Z^(1.5)-fold decrease in use rates.  But first two caveats. 1) You don’t get to propose something implausible (like a Z^5 scaling law or some other extreme relationship). And 2)  your proposed relationship must be consistent with empirical estimates of “elasticity”. Demand for drugs is actually highly *inelastic*, so the real relationship is probably more like a Z^(0.5)  or Z-raised-to-some-other-fraction-less-than-one scaling law, such that doubling the price gets you *less* than a halving of drug use.

The enterprise of drug prohibition looks implausible on a cursory analysis. An absolutely thorough analysis would explicitly draw supply and demand curves and measure consumer and producer surpluses per those neat Economics 101 diagrams, but my overall point will stand. You might achieve a small deterrence at a small cost, or a moderate deterrent at a moderate cost, or a great deterrence at a great cost. But the cost of drug prohibition scales at least as fast as the benefit, and most likely (injecting what is known about the inelasticity of drug demand) faster. We don’t get an enormous benefit just by declaring drugs illegal. It’s not an on-off switch that we can simply switch to “off.” It’s a throttle. It’s a continuously adjustable lever. If drug warriors are hoping to achieve a large deterrent effect at a low cost, they are deluding themselves. To convince someone to stop using, you can’t just wag your finger and say “No.” To achieve real deterrence, you must pre-commit to some form of harsh punishment (comparable in harshness to the inherent dangers of the drug itself). Some people will continue to use, which means you must pre-commit to tracking down those individuals and carrying out the horrible punishment.

Let me give a little flavor to what I’m talking about. Phrases like “cost-benefit analysis” and “raising the price from X to Y” sound a little dry and technocratic. The real-world consequences of drug prohibition are soul-wrenching. “Raising the cost from X to Y” involves forcibly removing children from their parents, throwing victimless drug users into prison with hardened criminals, subjecting intravenous drug users to the dangers of wildly varying heroin potency and unknown adulterants (like fentanyl), denying those same IV drug users clean needles (subjecting them to an epidemic of blood-borne pathogens), harassing dozens of innocent motorists or pedestrians for each “guilty” one, denying sick people the only medicine that makes them feel well, and (let’s not forget) denying a perfectly safe thrill to those drug users who never cause any harm to themselves or others. That’s hardly a full list of the negative consequences of drug prohibition. I could go on for pages, but you get the point. Inject some humanity back into the above calculus, and the war on drugs isn’t just contradicted by some dry cost-benefit analysis. It’s unconscionable. It’s a moral abomination. Given that there are perfectly serviceable alternative policies, often referred to as “harm reduction,” our current prohibition regime is particularly wicked and foolish. It’s time to try something else. 

Thursday, February 11, 2016

Campaign Finance and Impressionable Voters

Consider two possibilities.
  1.   Campaign spending has a negligible effect on election outcomes.
  2. The voting public is so ignorant and impressionable that campaign spending can readily sway an election.
Empirically 1) is more defensible; the “vast empirical literature” supports the conclusion that there is no effect of campaign spending on election outcomes. But let’s assume 2) for the moment. In that case, the problem is far too deep to be solved by “campaign finance reform.” Assuming that an election can be bought under free-wheeling “campaign finance policy 1” but reverts to the other candidate under a more restrictive “campaign finance policy 2”, it’s not at all clear that the more restrictive regime is better than the freer one. It’s not clear that the 1 gives you better policies, or that the results of 1 are imbued with greater moral virtue. If “the will of the people” is swayed by such a trivial influence, we need to doubt whether that’s even a meaningful concept. Dumb, impressionable voters pick bad policies. It’s no consolation that they weren’t influenced by money. Remove the influence of money, and some other meretricious influence will take its place.

Consider several ways that politics is unfairly biased in a way that can lead to bad policy:

Money influences politics, far beyond the actual merit of the policies it wins.
Public sector unions influence politics, far beyond the actual merit of the policies they win.
Sympathetic interest groups influence politics, far beyond the actual merit of the policies they win.
Economic populism influences politics, far beyond the actual merit of the policies it wins.
(Similar parallel structures for religion, defense hawkishness, nationalism, etc.)

Money is not the only thing that exerts an undue influence over politics. It’s not even the most destructive biasing influence.

Monday, February 1, 2016

CDC Drug Overdose Data: Patterns and Data Quality Issues

I went digging around in the CDC’s mortality data at the individual record level and it pretty much confirmed my suspicions about the recent surge in overdose deaths. I had done several data-pulls from CDC’s Wonder database, but that only allows you to pull grouped data with various different cuts and summaries. The CDC actually has the individual details on every single death that occurred for every year since 1999 (data files here and data layout here). Unfortunately, the data files are big (~1.5 GB or so) and have an alien file extension that nothing on my computer would open. Fortunately, I have a generous friend who is more tech-savvy than myself. I obtained the full list of the ~47,000 “drug poisoning” deaths, including every listed cause of death (up to 20 are possible, and most list a few), age, gender, marital status, autopsy indicator (was one done, Y or N?).

I was correct to suspect that most of these deaths are drug interactions, not single drug overdoses. I slightly suspect that part of the increase in recent years (tripling from 17,400 in 2000 to just over 47,000 in 2014) is not real, but is an artifact of more thorough toxicology screenings. I can’t prove that based on these data sets (I’d have to know what they *would have* looked like in a counterfactual world where this reporting bias didn’t exist), but it’s hard to explain some of the data trends without assuming there’s some kind of trend in reporting bias. Anyway, I won’t say much about the reporting bias in this post, because I’m only looking at one year’s records. I was probably right to suspect that many of these deaths were wrongly attributed to a drug that happened to be present; I think I have some good evidence that the medical examiner didn’t really know what they were doing (based on the lack of an autopsy, or the assignment to a clumsy “catch-all” cause of death code T50.9, “Other and unspecified drugs, medicaments and biological substances”).

The ~47, 000 figure often given for the total number of drug poisonings is a collection of very different causes of death. It includes ICD-10 codes X40-X44, X60-X64, X85, and Y10-Y14. Codes X40-X44 are unintentional drug poisonings; assuming the individual death certificates were coded correctly, they unambiguously belong in the total. Codes X60-X64 are intentional poisonings, suicides. These are very different classes of social problems. It’s not clear that adding together intentional suicides and accidental overdoses gives you a meaningful sum, just because the chemical mechanism behind the deaths was the same. Codes Y10-Y14 are “undetermined intent”, so many of them are likely suicides and many are likely accidental, but it’s impossible to determine the mix. It seems deceptive to include these in the sum without some sort of caveat. Code X85 is murder; there are only a few of these, but it clearly doesn’t belong in the sum. Alcohol poisonings are coded separately; the 47,000 figure doesn’t include alcohol poisonings unless it’s a combination of alcohol and other substances (which is actually fairly common). Of the 47,196 drug poisonings, 5,447 were suicides, 2,862 were “undetermined intent”, and 82 were murder. 38,841 were unintentional drug poisonings, codes X40-X44. So it’s not as though the ~47,000 figure is off by an order of magnitude. If your intent is to quantify the drug problem by counting overdose deaths, it’s off by about 20% or so. The vast majority are still “unintentional drug poisonings.” But it reeks of bad faith when someone overstates the magnitude of a problem in this way. To be fair to the CDC, some of its reports clearly specify the breakout (like this one). But it’s not always so forthcoming. I have often seen articles or blog-posts use the 47,000 figure as a headline or attention-grabbing block-quote.

I can’t quantify the following consideration, but it’s worth noting that some suicides could be miscoded as “unintentional poisonings.” The family might have religious, insurance, or other reasons not to mark the death as a suicide. Unfortunately, nothing in the data file allows me to quantify this kind of miscoding.

Multi-Drug Interactions

The codes mentioned in the above paragraph are broad cause of death categories; there are also “multiple cause-of-death” codes (mostly in the T40 through T50 range) that list specific causes of death. A few of these (T40.1 for Heroin, T40.5 for cocaine, T40.3 for methadone) are codes for a specific drug; others (T40.2 for “Other Opioids”, T42.4 for “Benzodiazepines”, T40.4 for “Other synthetic narcotics”) are for broad categories of drugs. It might be useful to know specifically if a death coded T40.4 was, say, a fentanyl poisoning from spiked street heroin or a synthetic prescription opioid painkiller that was carelessly mixed with alcohol; the cause of death codes unfortunately are not granular enough to reveal the true cause. A drug poisoning death is coded by one of the codes in the above paragraph (X40-44, 60-64, 85, and Y10-14); such a death will also typically include one or more of the “multiple cause of death” codes (the T40-T50 codes, including others related to organ damage, alcohol, etc.). This way, you can look at a death record and answer the two questions “Was it a drug poisoning?” and (with the multiple cause of death codes) “What drug or class of drug or combination of drugs was involved?” I found that most drug overdose deaths involve multiple substances. See the following table, and notice how rarely a single substance is implicated:

This doesn’t even include alcohol. I was a little unclear about which cause of death codes were alcohol, and apparently so are the medical examiners and nosologists who assign the cause of death codes. Sometimes a death involving alcohol is listed as T51.0 (“Ethanol”), sometimes X45 (“Accidental poisoning by and exposure to alcohol”), sometimes T51.9 (“Alcohol, unspecified”). Sometimes it’s a combination of these codes, as if the coder didn’t know which one to write down. I chalked this inconsistency up as another data quality issue. Anyway, when we include alcohol, even fewer of these deaths are single-substance overdoses:

Take heroin as an example, an archetype for a drug that’s easy to overdose on (and I’m not at all claiming that it isn’t). Only 29.5% of heroin deaths involved *just* heroin and no other drug; 70.5% of heroin poisoning deaths have other substances on the death certificate. Only 1.4% of benzodiazepine deaths have no other substances on the death certificate. This is not an inherently dangerous category of drugs, but people who take them need to be starkly warned not to take them with alcohol, opioids, and certain antidepressants. I thought the single-drug overdoses from methadone would be higher; the stuff stays in the system so long it can build up to dangerous levels for a naïve user. Still, it appears that the culprit is often a deadly combination, not methadone by itself.

I hesitate slightly in saying all the above. It’s possible that the medical examiners are simply writing down every goddamn thing they find on the toxicology screening, in the digestive track, or mentioned in interviews with family. I’m sure if you scoured the records you’d find death certificates with combinations of substances that aren’t really dangerous in combination, which would cast doubt on the notion that every drug mentioned on the death certificate contributed to the death. But this cuts the other way, too. If medical examiners are promiscuously listing every substance in a person’s body whether it contributed to the death or not, then no doubt some of these deaths are inappropriately categorized as drug poisonings. Unfortunately, this is a suspicion that I can’t quantify with the given data. If the death is miscoded, nothing in the CDC data file allows me to correct that error.

Why do I make such a big deal out of the multi-drug interactions? Because the solution to this problem is different from the solution to the problem that the public *believes* exists. The public thinks that people become enslaved by drugs, and then compulsively pop pill after pill (or snort after snort, or injection after injection) until they drop dead. This doesn’t really describe how most drug poisonings happen, or even how most single-drug overdoses happen. If a typical drug poisoning involves a bad combination of substances, then the solution is to make the users avoid *some* of those substances, not necessarily *all* of them. The solution is education and good PSAs: “Don’t take X with Y. It will kill you!” Don’t tell people “Don’t take X at all. It will kill you!” The moment they survive an encounter with X, they’ll stop believing everything else you say. People should be warned about the dangers of drug use, but the warnings should be accurate and useful. The warnings should be crafted with the level-headed realization that some people will imbibe anyway. We might as well warn them so they do it safely. You can’t look at the above tables and say, “The solution is to prohibit drug use, beat up and imprison dealers, and force users into counseling.” Prohibition is an extremely ham-fisted (and ineffective) solution to the problem that actually exists.  Most of these deaths are naïve users who don’t understand the interactions between the substances they are using. Society should accommodate this kind of recreational use with proper guidance on potentially dangerous dosage and combinations.

If you want to know what kinds of drugs tend to found in combination, I’ve built a cross-table with the  14 most common drug poisoning codes, plus alcohol. In the top table, you can see that the first row and the first column are identical, as they should be. 4,791 deaths, for example, involved “Other Opioids” and “Other and unspecified drugs, medicaments and biological substances”, as you can see in row 1 column 2, and row2 column 1. In the bottom table, I am dividing by the largest number in the row. So read this as saying “25.5% of ‘Other and unspecified drugs, medicaments and biological substances’ deaths involve ‘other opioids’” and “47.7% of ‘other opioid’ deaths also involved ‘Other and unspecified drugs, medicaments and biological substances’. Clearly there’s 100% across the diagonal; 100% of deaths involving “cocaine” also involve “cocaine”, get it? 34.2% of “other opioid” deaths also involved Benzodiazepines, which confirms my suspicion that this is a big deal. 52.3% of benzodiazepine deaths also involve “other opioids”. (The 34.3% and 52.3% figures have the same numerator, 3431, but different denominators.)

Acute vs Chronic Causes of Death
Often these overdose death records include a cause of death that is associated with chronic alcohol or drug use, rather than acute poisoning (K700, “Alcoholic fatty liver”, accompanies several of the above mentioned alcohol poisoning codes). This suggests there is some confusion and uncertainty about the actual cause of death. The causes of death related to chronic drug use are F11-F16 and F18-F19; they are separate from the codes for acute poisoning and are not included in the “47,000” figure so often cited. A death that includes both chronic and acute cause of death codes is cast into doubt. (See page 74 of this document; the city of New York had the “chronic” and “acute” drug deaths confused for many years in one of its official documents.) I don’t have the chronic drug use cause of death totals handy, but they are quite a bit smaller than the number of acute poisonings (they are roughly in the ~2,000 range rather than the ~40,000 range). It seems like a meaningful measure of the drug problem would add chronic and acute accidental drug poisonings, not accidental and intentional drug poisonings (including murder). That number would be right around 41,000 in 2014.

Absent Autopsies and Catch-all Categories

There are some patterns in the data that are cause for concern, and I’ll spell out two of them in this section. The CDC data file has an autopsy indicator, with “Y” for “Yes, an autopsy was done”, “N” for “No” and “U” for “Unknown.” Autopsies are much more common for drug overdoses than for other categories of deaths (75 – 80% of drug overdoses involve an autopsy; I don’t have the same figure for total deaths but it’s definitely smaller). Of course that still calls into question 20-25% of poisoning deaths. There is just no way to identify a drug poisoning death without an autopsy. In his textbook “Pathology of Drug Abuse,” Dr. Steven Karch discusses all the various ways that, even *with* an autopsy, an unwary medical examiner might be misled about the cause of death. He repeatedly cautions that a toxicology screening alone can be extremely misleading. Methadone patients, for example, often walk around with levels of the drug in their bodies that would be toxic to most people. If such a person drops dead of a sudden heart arrhythmia, it’s unlikely that the true cause of death will be discovered (autopsy or not). Of the 38,841 accidental drug poisonings in 2014, 8,077 (20.8%) had no autopsy and 8,847 (22.8%) were marked with either an “N” or a “U”. This seems like a huge caveat that should be added to every CDC report that simply adds up the totals by cause of death code. Such a sum overlooks these underlying problems with data quality.

Another data quality issue I noticed was the liberal use of the cause of death code T50.9, “Other and unspecified drugs, medicaments and biological substances.” This is a very generic-sounding category that suggests the medical examiner didn’t really know what was going on. A total of 18,782 of the accidental overdose deaths were marked with this cause of death code; 7,389 were marked with this cause of death code *alone* and no specific category of drug (see above tables; the number falls to 7,024 if I include alcohol as a substance). I suspect that if the medical examiner really knew what killed the person, they’d mark down one of the other drug categories. Some of those other categories are pretty broad, so it shouldn’t be hard to find at least *one* that’s appropriate. The number of asterisks marking the drug poisoning deaths figure is growing here; 19% (=7,389 / 38,841) of these deaths can’t even be attributed to any specific class of drugs.

I don’t want to double-count in my last two paragraphs; there are some deaths for which there wasn’t an autopsy *and* it was marked with code T50.9 alone (with no other substances). When I take out the double-counting, I get 36.1% of drug poisoning deaths fall into one or the other category. So just over a third of the 38,841 accidental drug poisonings are in serious question. (Details: 5160 deaths had an autopsy and had T50.9 as the only drug-related cause of death, 8847 deaths had “N” or “U” under autopsy, (5160 + 8847)/38841 = 36.1%. Please, please, please, download the data file and check my work.)

There is another issue that I’ll mention without quantifying. Many of these deaths include cause of death codes for various types of organ failure. I don’t know if it matters what order the causes of death are listed in (remember there are up to 20, and most deaths include several), but sometimes the organ failure code is listed prior to the drug poisoning code on the death record. Here are some of the organ failure causes of death listed on some records that I’m essentially pulling at random: “Hypertensive heart disease without (congestive) heart failure”, “Atherosclerotic cardiovascular disease, so described”, “Cardiomegaly”, “Influenza due to identified avian influenza virus”, “Sleep apnoea”, “Chronic obstructive pulmonary disease, unspecified”. Sometimes multiple organ-related causes like those listed above were pulled from the same record, and sometimes the only category of drug poisoning on the record was the catch-all “Other and unspecified drugs, medicaments and biological substances” mentioned above. I’m really not sure what’s happening here. Did the medical examiner infer with certainty that a drug caused the death, then pedantically list the effect the drug had on the decedent’s organs? Or did the medical examiner write down everything that was wrong with the body and someone at the CDC (speculatively) coded it as a drug poisoning?

I’m sure that some of the increase in drug poisonings since 1999 is real. I don’t want to imply that the entire trend is spurious. The sheer tonnage of opioids prescribed has increased dramatically, and these drugs are bumping into other drugs that people are already taking (alcohol, benzodiazepines, antidepressants, etc.) But this is not the moral panic it’s being sold as. As I’ve said in previous posts, the illicit use of painkillers, sedatives, tranquilizers, and other prescription drugs is flat, according to the SAMSHA and Monitoring the Future surveys. It’s simply not the case that we’ve made recreational drugs more available, and drug-crazed addicts are popping them until they drop dead. This is not an addiction crisis. Until we properly understand the mechanism behind the increase in drug poisonings, we can't really claim an informed opinion about the solution.