A very nice application of the economic way of thinking that utilizes elasticity to consider public policy.
The U.S. government's policy of drug prohibition, like alcohol prohibition before it, is a failure—and not one that can be corrected by a mere tweaking of current policy. The economic analysis of fighting a supply-side drug war predicts that the war will enhance drug suppliers' revenues, enabling them to continuously ratchet up their efforts to supply drugs in response to greater enforcement. The result is a drug war that escalates in cost and violence.
Furthermore, the secondary consequences of prohibition are perverse. The drug war causes drugs to be more potent and their quality less predictable than if drugs were legal, leaving the remaining users at greater risk and, in the face of higher prices, more likely to commit crimes to support their habit.
In short, the means—drug prohibition—is incompatible with the ends of improving health and decreasing violence. There are two paths forward: a demand-side drug war or legalization.
A demand-side drug war that places draconian penalties on usage or possession does not necessarily fail a means-ends test. If the goal is simply to end drug usage, implementing a swift death penalty for anyone convicted of possession or use would do the trick. Of course, it would not improve those people's health. Most people, including myself, would consider such a policy even more unjust than the current drug policy.
The alternative, legalization, is a better path forward. It passes a means-ends test while better respecting individual liberty. Consumption may increase but the drugs that are consumed would be safer; and violence would decrease. To the extent that decreasing drug consumption is desirable, moral suasion and education would be more effective—without the nasty side effects—than the current policy of prohibition.