If you took Economics 101, you can probably dredge up cases of market failure in which government intervention is justified. For example, governments tax the emission of pollutants (or regulate them directly) because the cost of pollution would otherwise be borne by third parties and thus not taken into account in the balance of supply and demand. By the same token, governments regulate pharmaceuticals because buyers would otherwise lack enough information about their safety and efficacy to judge their value.
"By the same token?" It's not the same token at all. The first part of the paragraph lays out the standard Pigovian case for government intervention to deal with externalities. The last sentence doesn't. It's a very different case. At most, it's a public goods case for intervention.
But it's hard, even granting the "information as a public good" case, to get from that to the particular intervention--regulating pharmaceuticals--that Eichengreen thinks is justified. If the problem is that people lack information, the best solution is for them to get information, not to have the government substitute its preferences for theirs. Even people with full information will make different choices than other people with full information. Imagine, for example, that you have a disease and there's a drug available, but not yet legal, that could help cure the disease. What would you do? Your answer might well differ from that of someone else who has the same information.
As I wrote in The Joy of Freedom: An Economist's Odyssey, "The FDA may have some expertise when it comes to drug safety and efficacy, but on the only issue that matters--your tradeoffs between various risks--you are the expert, and the FDA's scientists are rank amateurs."
Also, in that book, I quoted a 1995 letter from my friend, co-author, and former student, Charley Hooper. I still think it's one of the best statements of the issue. Charley wrote:
The choice a patient makes between therapies (with the help of his agent, the doctor) is based on many variables: efficacy, tolerability, side effects, riskiness, monetary cost, nonmonetary cost (e.g., hassle), speed of action. These drug costs and benefits must be judged within the context of many personal values and tradeoffs: the fear of death, the fear of surgery, the fear of the hospital, potential pain, and the individual's health profile, financial status, value of time, value of health, and risk tolerance. For the FDA to decide what compounds pass this complex tradeoff is preposterous, given that the FDA can never frame the problem from the individual patient's perspective. One individual's best alternative could be another's worst. We have seen this with AIDS patients: "I don't care if I develop cancer and this costs me $20,000 a year because without it I'm dead in 6 months." If, instead of medical therapies, they were telling us what kind of washing machines to buy or where to go on vacation, we would consider it laughable. This is what von Mises [Ludwig von Mises, the noted Austrian economist who showed that information problems would prevent socialism from working] said. Centralized bureaucrats cannot make the proper decisions for individuals because they lack the requisite information.