The Review of Austrian Economics, 18:1, 109–115, 2005.
2005 Springer Science + Business Media, Inc. Manufactured in The Netherlands.
Conﬂict, Cooperation and Competition in Anarchy
TYLER COWEN email@example.com
Department of Economics, George Mason University, Fairfax, VA 22030-4444, USA
DANIEL SUTTER firstname.lastname@example.org
Department of Economics, University of Oklahoma, Norman, OK 73019-2103, USA
Abstract. Caplan and Stringham (2002) attempt to rebut the “paradox of cooperation” (Cowen and Sutter 1999)
as it applies to libertarian anarchy. The paradox in the context of anarchy implies that if private defense agencies can
cooperate to avoid conﬂict they can also collude to reestablish coercion. Caplan and Stringham argue that arbitration
is self-enforcing while collusion requires solution of a prisoner’s dilemma. We agree that collusion requires more
cooperative efﬁcacy than arbitration, but maintain that arbitration requires considerably more organization than a
simple coordination game. If a network of protection agencies can organize sufﬁciently to arbitrate disputes, they
can also create a barrier to entry by refusing to arbitrate with entrants.
KeyWords: anarchy, cooperation, evolution of government
JEL classiﬁcation: D74, H11, D72.
Cowen and Sutter (1999) outlined a “paradox of cooperation.” If civil society can use norms
to enforce cooperative solutions, that same society will be prone to certain kinds of cartels.
In other words, cooperation-enhancing social features will bring bad outcomes as well as
good outcomes. To provide a simple example, the Nazis relied on cooperation in addition
to their obvious coercive elements in perpetrating their crimes. The ability to organize
therefore is a mixed blessing. In the context of libertarian anarchy, this argument implies
that private defense agencies are likely to collude and reestablish coercion. We refer the
reader to our original paper for the details of the argument (see also Cowen 1992, 1994).
Caplan and Stringham (2002) attempt to rebut the argument as it applies to libertarian
anarchy. They maintain that inter-agency collusion requires solution of a prisoner’s dilemma
problem, while private defense agencies face only a coordination problem in resolving
disputes peacefully through arbitration. They view membership in an arbitration network
of agencies as self-enforcing and believe that such a network will not evolve into a cartel.
Self-interested individuals will defect from a cartel in the absence of a sufﬁciently vigorous
punishment mechanism. Private defense agencies supposedly have enough cooperative
efﬁcacy to overcome the coordination problem but cannot collude. Thus the arbitration
network will not devolve into government.
Caplan and Stringham have advanced the debate on cooperation and anarchy. We accept
their contention that collusion requires greater organization among network members than
establishing a system of arbitration. We remain skeptical though about the likelihood of
benevolent noncollusive anarchy. Establishing an arbitration mechanism we contend brings