I was initially going to post this as a comment to Greg Conko's recent post arguing against the Court's recent decision in Wyeth v. Levine, but the comment system didn't work correctly for me. I appreciate the force of Greg's argument (and I certainly agree that this particular case should have been decided much earlier on different grounds), but I think there is room for reasonable disagreement within the libertarian community about whether FDA preempting state tort law is good or bad. This is one of these questions about what to do in the real world, where first-best solutions just aren't politically possible. Most libertarians would likely agree that there should be only one system to deal with injuries caused by products, including drugs: the tort system. The tort system only operates once there has been an actual injury, not just some scared politician's prediction of a harm. And it doesn't ban anything; it just forces manufacturers to internalize the external costs of the injuries their products cause. If the benefits of the product outweigh the total (social) costs, the product will keep being made. This is the optimal outcome. And, further, the injured consumers will not simply be ignored; they get compensated for their harms, restored to a position as good as they'd be in had they not been harmed. Unfortunately, however, we don't just have the tort system. We also have the FDA and other prospective regulatory agencies that pass judgment on products before they hit the market, banning some and restricting how others can be marketed and sold. If the FDA isn't going away, what should we do? Greg's answer is essentially the one I gave in the context of internet regulations: "multiple levels of regulation [are] always worse than... only one." There is a difference between multiple levels of regulation and a tort system plus a regulatory system, though. If the FDA gets lots of stuff wrong, but the tort system functions ideally (a big, and admittedly untrue, assumption), then the FDA should not be able to preempt the tort system. If the FDA allows a drug that nonetheless causes injuries, the tort system has not failed or "overregulated" if it correctly assesses and assigns damages. It has internalized costs that would otherwise be external. If the Court had found preemption, however, then there would be calls for the FDA to regulate even more heavily, banning every drug that might cause any problem. The tort system currently functions as a safety net. Remove it and politicans and voters will demand more stringent protection from the system left - the regulatory one. However, finding no preemption emphasizes the point that the FDA is not infalible and that the tort system does a better job, at least sometimes. Unfortunately, it only does this in one direction, and does nothing to expose the much more common and deadly - but largely invisible - type I error. Still, I think though the tort system may get some stuff wrong and incorrectly over-compensate, this risk is more acceptable than the risk of further entrenched and onerous FDA regulation.