Anarchy and Violence

   

Matt Zwolinski offers three arguments in support of Molinari’s pessimism, late in his life, about going all the way to market anarchy. The first is that violence is sometimes a pleasurable consumption activity, the second that individuals are particularly irrational with regard to violence, the third that violence imposes external costs.

My response to the first is that the argument for why rights-enforcement firms would be unlikely to use violence against each other does not depend on there being no goons available for hire who enjoy shooting people, only on there being few goons who enjoy being shot. Unless one agency has a large advantage over another, each should be able to make conflict costly for its opponent. That corresponds to my standard example of private property in the animal kingdom—territorial behavior. The reason why a trespassing bird or fish usually backs off when confronted by the “owner” of the territory is that, unless the inequality of strength is large, a fight to the death is a loss for both participants.

It is possible that individuals are less rational about violence than about other things, although what looks like irrational behavior may be a result of the sort of hardwired commitment strategies that, in my previous example, allow the claimant to retain his property, usually without fighting for it—irrational ex post, rational ex ante. But the violence at issue here is between firms, not individuals. If I am correct in believing that interagency violence is an unprofitable business strategy, we would expect over time that firms that failed to control such irrationality by their employees would lose out to those that succeeded.

Violence imposes external costs. That implies that individual rationality will not automatically produce the optimal level of violence—and, under current institutions, it doesn’t. But, as Ronald Coase pointed out quite a long time ago, the existence of externalities does not lead to inefficient outcomes if transaction costs are sufficiently low. The violence relevant to Matt’s argument is violence between rights-enforcement agencies, pairs of firms engaged in long-term repeat dealings with each other. That is a context in which we would expect transaction costs to be low, making it possible for the parties to bargain to something close to the optimal outcome, which in this case means little or no violence.

Finally, Matt writes:

That might be true of the version of anarchism encapsulated in Rothbard’s line about ending the state by pushing a button, but it is not true of either my version or Molinari’s, since neither of us is proposing to instantly instantiate our vision. My view is, and I think Molinari’s pretty clearly was, that the way to get to a stateless society is by a process of gradual evolution within the structure of existing institutions. Ideally, as in Stephenson’s Snow Crash, when the state finally ceases to be relevant nobody notices.