Responding to my post below, Radley makes the familiar (and valid) libertarian point that private certification services could take the place of state assessment of a restaurant’s sanitary status. He also makes an inapt comparison to the argument for smoking bans, which it’s worth getting out of the way first. As opponents of the smoking ban have been arguing all along, these are quite different from either building codes or sanitation requirements, because it’s perfectly obvious upon cursory examination whether or not there’s smoking in a particular bar. Since anyone who walks into such a bar has effectively agreed to whatever risk their exposure to smoke might entail, such that there’s no question of a patron being defrauded into thinking he was in a smoke-free room, preventing people from choosing such environments does, as Radley suggests, implicate “the sanctity of private property.” I don’t think food inspections are in quite the same boat. I’m not typically in a position to make a judgement about the sanitary conditions in a particular restaurant, at least not by direct inspection. And, presumably, if I take legal action against a restaurant claiming that it’s fed me poisoned food, it’s going to be OK for some agent of the state to go onto the restaurant property and take a look at the kitchen to discover whether the claim has any merit.

Still, it’s certainly the case that, as with, say, Underwriters Laboratories or Consumer Reports, the market would very likely produce some kind of private certifiers that, in tandem with tort and contract law and the market liability of being known as a rat-infested bistro, would serve much the same function as public inspection.

So I think we’ve got a pretty good set up for three categories of state action we might imagine. There’s a large category of intrinsically illegitimate government actions: Arresting people for doing recreational drugs, having consensual sex with a prostitute, running a high-stakes poker game, and so on. Assuming the consent of the parties involved, nobody’s rights are violated by those activities, so the state ought not normally to be involved at all. (It might be involved in punishing those who run rigged poker games or sell baking soda labeled as cocaine, but that’s another story altogether, better assimilated to the third category below.)

Then there’s a handful of “easy cases” where (as it seems to those of us who aren’t anarchists, anyway) only the state can adequately serve a certain function. Libertarians typically include here national defense, the provision of the legal framework within which markets operate, and (a bit more contentiously) some level of environmental protection, at least in cases (like air pollution) where it would be impractical to have courts trace ex post the marginal contribution of particular environmental harms to particular pollutors.

Things like food inspections or building codes seem to fall into a fuzzy third category. On the one hand, serving someone tainted food, selling a toaster that explodes after two months, or renting out a structurally-unsound building would all at least appear to fall under the rubric of libertarian rights violations, assuming the seller hasn’t explicitly informed the buyer that he’s getting an exploding toaster or ordering a salmonella sandwich. Most libertarians would allow that the state could legitimately be involved in those cases ex post in determining whether the seller had observed due diligence in each case and whether the buyer had a reasonable expectation that the product wouldn’t be “faulty” in that particular way, as well as in determining the appropriate compensation and/or punitive payment.

Now, in these cases, we face a question Robert Nozick grapples with in Anarchy, State, and Utopia: to wit, when do we attempt to prohibit certain risky actions ex ante, and when do we instead have a system of ex post compensation via the legal system, combined with the market deterrent of being known as a risky transactor? Obviously, we don’t always go with the ex post approach: Most of us think that police should, when they can, stop robbers or assailants in the act, and that the courts should punish attempted robbery and assault, rather than merely letting the action go on and having the victim seek redress after the fact. Equally obviously, we don’t always prohibit or set up intrusive inspection systems before the fact any time an activity has a risk of harm to third parties through negligence.

So the fuzzy third category of government actions I’m thinking of is those where we could deal with certain harms via a combination of private, voluntary certification or inspection ex ante and tort law ex post, but instead we’ve gone with a system of mandatory inspection that attempts to prevent harms rather than deterring and compensating. I place these in a third category, because it seems that here we have a case where state involvement is not (as with bans on strictly consensual activity) per se unjustifiable, but rather contingently inferior to the voluntary/ex post approach. When it is inferior, of course, we should certianly be interested in urging a transition to that latter approach, but it does seem at least not to have the moral urgency of, say, ending the drug war.

A few factors seem especially salient in determining which approach is preferable. One is the extent of the potential harm. In cases where I stand to lose money or property, it’s easy enough to say: “Well, sue after the fact for restitution.” In cases where life and health are threatened, the fact that the person or firm responsible is put out of business and the victim’s family recieves a settlement is cold comfort to the dead person, so there’s a better case for ex-ante action. These are also harms where the very magnitude of the injury makes both deterrence and compensation trickier because of judgement proofness: For the average falafel-stand operator, the difference (and the difference in deterrent effect) between being hit with a $1 million judgement and a $10 million judgement is effectively nil. Another factor is how easy it is to prove one’s case in court after the fact. Did I get sick because the kitchen in which my food was made was unsanitary? Did my apartment catch fire because of irresponsibly bad wiring? It won’t necessarily be clear—but, of course, that’s a problem that can be addressed by voluntary private certifiers.

On net, I tend to think that Radley’s right to suggest a private cert system would be more efficient and flexible than the government inspection model. But that’s a variable function of the specific informational and transaction costs of the two alternatives. These are cases where the libertarian’s attitude, I think, should be reformist but not condemnatory in quite the same way, analogous to the case where we think the police department would do better to outsource its forensics department, or that legislators are issuing an inefficient number of emissions vouchers each year. If the transaction costs of the private solution fell out differently, prior inspection wouldn’t be intrinsically objectionable, so the error involved in going that route is of a different order than the kind involved in the first category.