I have read many criticisms of the War on Drugs, but I never see it pointed out that it is internally inconsistent. It truly requires inconsistent assumptions to work.
Are drug users rational or irrational? If they are rational, then they are already accounting for the built in deterrents of drug use and the current rates of use are appropriate. If they are irrational, then legal punishments will fail to deter them.
One can think of this as asking two yes/no questions and filling out a grid with the four possible combinations of answers. The first question is: Is drug use harmful? The second question is: Are the users rational? The combination of answers can be used to set the optimal drug policy. If the answer to the first question is “No,” then your job is done. Legalize the drug for use. It doesn’t matter if people are rational or irrational if they’re doing something that isn’t harmful.
So what happens if the answer to the first question is “Yes”? Consider the “drug use is harmful/users are rational” combination. There isn’t a problem here, because the users are already accounting for the risks of drug use, taking appropriate precautions, and using appropriate amounts. “But what about costs to third parties?” you say? Those are also already accounted for. If taking “Drug X” might turn me into a murderer (or burglar or rapist), then as a rational user I would already be accounting for that probability. Not liking the prospect of a long prison term, I’ll have moderated my drug use or decided against use entirely. And in any conceivable regime DUI would remain a crime, regardless of the legal status of the drug. In the “rational user” scenario, drug use has some costs *to me*, but all the “external” costs are internalized, because all the ways I might harm somebody under the influence are already criminalized (and rightly so!). In this scenario, drug users will respond to criminal sanctions against drug use by reducing use, but that’s overkill because they are already accounting for the costs and have already moderated their behavior appropriately. You’ll achieve some additional deterrence, but it can hardly be worth the cost of needlessly punishing the users who are caught. Keep in mind that “drug prohibition” means “occasionally ruining the lives of a few random drug users.” This is a cost that shouldn’t be taken lightly.
Now consider the “drug use is harmful/users are irrational” scenario. In this case, users are not deterred by the pharmacological risks of drug use. However, they are also not deterred by the legal risks of drug use. So you can “ban” drugs and punish a few random drug users, but the deterrent effect is too small to matter.
In none of these scenarios can any serious person recommend drug prohibition as a policy. You either fail to achieve any significant deterrence, or you succeed in deterring something that’s not a problem. You don’t get to assert that drug users are too irrational to respond appropriately to one set of risks (the pharmacological risks of drug use) and yet rational enough to respond appropriately to another set of risks (the legal risks of drug use). Such an assertion, implicit in all advocacy for drug prohibition, is logically incoherent.
Of course this is all a little too glib. These things aren’t binary. Drugs vary in the degree of harm they might cause. People vary in their degree of rationality. The severity of the legal sanctions can vary. And not all the external costs of drug use are truly internalized (think of the spread of AIDS from IV drug users to their non-using sexual partners, which isn’t effectively criminalized). Still, it’s hard to tell a story where the punishment achieves any significant deterrence.
Assume for a moment that people are rational enough to be deterred by legal sanctions. To figure out how much deterrence you are achieving, you have to now compare the costs intrinsic to drug use to the legal penalty for drug use. The intrinsic costs include things like the possibility of death, serious health problems, alienation of friends, loss of social status, loss of a job, and the time and money spent acquiring and using the drugs. (You would have to discount the “severe but unlikely” items, like death, by considering the small probability of realizing those costs. Also, note how some of these “intrinsic” costs can be exacerbated by the black market status of the drug.) Seen this way, drug prohibition doesn’t raise the cost of drug use from “zero” to “something positive”. It raises the cost from “something substantial” to “something somewhat more substantial.” You can achieve a great deal of deterrence by threatening drug users with life in prison or the death penalty, but with severe penalties you risk doing more damage than the drugs themselves are doing. In this case the cure is worse than the disease, so what’s the point? You can expect significant deterrence ONLY in a regime where the expected harm from legal sanctions is comparable to the expected harm from drug use. If a government DOES achieve any kind of significant deterrence, its lawmakers and technocrats should suspect that the penalty has been set too high.
Let’s consider another set of two binary questions: Are the intrinsic costs of drug use high or low? Are the legal costs of drug use high or low? If it helps, think of these costs as something concrete. A “low” cost might mean a 0.1% chance of ruining your life, or perhaps the loss of a week’s income. A “high” cost might mean a 5% chance of ruining your life, or losing a full year’s income. Use whatever measure you like, and use whatever differential between “low” and “high” you like (I’m using a factor of ~50). In the low/low and high/high scenarios, you can probably assume there’s some degree of deterrence. The intrinsic costs of drug use (those costs that would be present in the absence of legal sanctions) are comparable to the legal costs. If you’re roughly doubling the price of something, you should expect to see a decline in the quantity demanded. But in both cases the added (legal) harm is comparable to the (intrinsic) harm deterred, so it’s hard to declare “victory.” In the “low intrinsic cost/high legal cost” regime, you can say with some certainty that there is a great deal of deterrence, but the harm done to the few random users who are sanctioned is difficult to justify. Naturally if you take an activity that is mildly dangerous and make it 50 times as dangerous, you’ll surely achieve some deterrence. But surely it’s not worth it. In the “high intrinsic cost/low legal cost” scenario, you don’t do much additional harm, but you don’t really achieve much deterrence either. (If you use the cost differentials I give at the opening of this paragraph, the increase in cost is about 2%.) So you achieve a completely insignificant amount of deterrence, and to do so you still have to harass a few drug users. In none of these cases can you confidently claim that the deterrence is worth the costs imposed by legal sanctions. Ideally you could achieve a great deal of deterrence in return for doing a small amount of harm to existing users, but it’s hard to imagine a scenario were this actually happens. If you’re an economist drawing someone’s demand curves for drugs, you’d have to specify that there’s an inexplicable discontinuity that the legal penalty exploits perfectly.
To properly do this analysis someone would have to specify everyone’s demand curves for drugs and show what happens when you add the costs imposed by legal sanctions on top of the intrinsic costs. That said, it’s worth exploring this exercise in plain English. I’d be very surprised if anyone came to a substantially different conclusion by drawing explicit demand curves and doing a formal cost/benefit analysis of imposing legal sanctions.
In this entire discussion, it may seem that I’m ignoring the legal sanctions on the supply side. For the most part, drug users get a slap on the wrist and drug dealers, smugglers, and producers get heavy penalties. These supply-side penalties simply serve to drive up the market price of illegal drugs. In my above discussion of the legal costs of drug prohibition, the cost includes both the legal penalty applied to the user and the increased market price. As a side note, this entire regime of extreme legal penalties on the supply side is terribly inefficient. If a government wishes to increase the market price of drugs, it should do so with a tax that it can collect. You have far less collateral damage in terms of drug dealers killing each other, tainted product, inconvenienced motorists being subjected to unnecessary searches, and terrified residents being subjected to no-knock raids. And in the case of a tax, the costs imposed on producers actually *go to* someone. They become a “transfer” instead of a “dead-weight loss.” The “tax” regime is infinitely preferable to the “criminalize producers” regime. But that could be another post all on its own.
Note that all of the above discussion is broadly consistent with the empirical record of drug prohibition. We have over a century’s worth of evidence covering very different anti-drug regimes (mild to severe punishment) and we see very little change in overall drug use. The War on Drugs is busted. It's time to move past it.