Becker and Murphy have been making this argument for a while. Indeed, they first broached the subject back in the 1980s when they argued that many drug addicts are perfectly rational consumers. They, like the savviest grocery shopper, use the resources they have to get what they want in a cost-effective manner. It's just that what they want happens to be drugs.
But the most relevant paper for understanding the Becker/Murphy critique is one they coauthored with CUNY's Michael Grossman, who was last seen in these parts arguing for a tax on junk food. The paper attempts to describe the economics of the drug market in the broadest sense. In particular, the authors want to combat the idea that legalizing drugs would lead to more addictions.
That argument goes roughly as follows. Drug addiction is a major social malady, and one way to reduce its prevalence is to make drugs more expensive. You can do that via taxes, à la taxes on alcohol or cigarettes, or you can do it by making drugs illegal. Due to the costs associated with evading law enforcement, the production costs of illegal goods are higher than those for legal goods, and this shows up in the price consumers pay.
So you could, as drug legalization advocates propose, legalize all drugs and tax them at a high enough rate that use doesn't increase. But that would fuel a new black market. Comparatively, criminalization might even be cheaper. Indeed, Harvard economists Andrei Schleifer and Edward Glaeser argued in 2001 that "quantity reduction" through methods like criminalization is in certain cases more cost-effective than taxing undesirable activities.
Becker, Murphy and Grossman think this view is dramatically mistaken and they build a model to show why. They put together an equation that outputs the benefits to society of a drug prohibition regime, taking into account the social cost of drugs but also the cost of enforcement, and the enjoyment that drugs provide to their consumers. They then used that equation to determine the regime that maximizes social welfare and determined what the cost of drugs would be under that regime.
Here's the kicker: if drugs sold for that price after taxes in an environment where drugs are legalized, they'd still be cheaper than drugs sold on the black market. So the legal market would drive illegal producers out of business, there wouldn't be any of the enforcement costs — including huge social costs like mass incarceration — that come with drug prohibition, the government would gain considerable new tax revenue, and because the price is the same, consumption of drugs wouldn't be any different than under prohibition. In short, the best form of prohibition is still worse than legalize-and-tax.
Obviously, models like this can only do so much without empirical evidence and other experts have endorsed markedly different approaches. UCLA's Mark Kleiman, for example, has argued that a program of "coerced rehab" would easily pay for itself by reducing addictions rates and thus the attendant costs of drugs and their prohibition. But the Becker/Murphy/Grossman plan has a strong argument behind it, and one that suggests the current course of U.S. drug policy is dramatically wrongheaded.