In light of this Dan Drezner post, a better way to phrase one of my points in the “Do Not Call” post below occurs to me. If you apply a stripped down version of the Coase Theorem to this case, and (just work with me for a second) assume away transaction costs, it actually shouldn’t matter whether we assign the contested right here to the consumer or the telemarketer. If the initial owner of the right values it less than the other party, it’ll be purchased, and the efficient outcome will be reached. (Though, of course, who gets the surplus may change.)

Except, in the case of initial allocation to the telemarketer, that doesn’t hold after all, because you get the “ear flicking problem” I mentioned. In other words, it may become profitable to set up a telemarketing firm exclusively to extract payments from people who don’t want to be bothered. And having paid one firm to leave you alone just shifts business to every other telemarketer you haven’t yet paid off, so you don’t get any net reduction in harassment. The upshot is that under an initial allocation to the telemarketer, either you create a deadweight loss by encouraging resources to be expended on threat production, or, as appears to actually happen, people realize there’s no benefit to engaging in such transactions, and so the potentially efficient transfer of the right doesn’t occur.

As far as I can tell, there’s no comparable problem for an initial allocation to the phone owner. If that allocation is inefficient, in other words, each individual can opt in and transfer the right to the marketer… and in a more nuanced or carefully constructed form (only call at such and such times with respect to this kind of product), which further expands the horizon of mutually benificial trades.

Incidentally, this is probably a generalizable principle: I’m actually lifting heavily here from Robert Nozick’s argument in Anarchy, State, and Utopia for prohibiting blackmail. I suspect there’s a range of cases like this where you start with the Coasean idea that the efficient allocation of rights can be reached if transaction costs are low enough, however you set up the initial allocation… but then you look to see whether one initial allocation or another actually either facilitates or impedes those efficiency-increasing transactions.