Jeff's piece is comprehensive and persuasive. I suggest that the national
defense issue is hardly insurmountable. If we apply Coase's transaction cost
analysis, people at large will never be able to afford the combinatorial
transactions necessary to mount national defense. The usual argument is
then: So, obviously we need government to do it, the sovereign is our only
hope for defense against evil empires and barbarian hordes. But that gives
away too much (as I think Robert Nozick did in Anarchy, State and Utopia).
The real question is: Are there low-transaction cost alternatives to the
sovereign? The answer is yes (and is Coase's own analysis of the firm in his
famous article in Oeconomica (1937)). Large corporations exist because they
minimise transaction costs. Large corporations can cooperate (the premise of
our antitrust law is that they will unless prevented from doing so).
Aggregations of capital such as cooperating large companies can defend
themselves and their markets. Everyone else is a "free-rider" of that
defense, just as we are of state defense. But the state taxes us for the
privilege. The defending companies would simply work the cost into their
price, which we voluntarily pay or do not. Big companies already privide
themselves both domestic and overseas security (Wackenhut is a private army).
A counter argument is that these companies will form monopolies. So what? A
perfectly price discriminating monopolist produces an efficient output on an
efficient allocation of inputs. The transer fo wealth to the monopolies
(consumer surplus is gone) is a transfer to the shareholders who are the
public. So, even at some deeper level, the need for a state for defense
seems to me to be an over-determination. We can do with less, at least in
theory.