Review of Industrial Organization 16: 229–246, 2000.
© 2000 Kluwer Academic Publishers. Printed in the Netherlands.
The Aggregate Relation between Proﬁts and
Concentration Is Consistent with Cournot Behavior
Department of Economics, University of New Mexico, Albuquerque, NM 87131, U.S.A.
RAYMOND D. SAUER
Department of Economics, Clemson University, Clemson, SC 29634, U.S.A.
Abstract. An important and controversial stylized fact in industrial organization is the positive
correlation between industry proﬁt and concentration. One interpretation of this ﬁnding is based on
the theories of Chamberlin and Stigler, which imply that concentrated industries facilitate collusion.
But non-cooperative proﬁt maximizing behavior can also generate a positive correlation. This paper
presents an equilibrium model of oligopoly which nests the behavioral assumptions of Bertrand,
Cournot, and Chamberlin. Simulations of the model under the Cournot assumption yield regression
coefﬁcients for the proﬁts-concentration relation that are very close to the estimated coefﬁcients in
Key words: Concentration, Cournot, oligopoly model, proﬁts.
Despite decades of research, the cross-sectional variation in the rate of proﬁt
across industries remains poorly understood. Although most agree that a “handful
of results have become conventional truths” in industrial organization, the ﬁeld
does not know what to make of them (Peltzman, 1991, p. 213). Instead, econo-
mists have generally abandoned inter-industry research to focus on what Bresnahan
(1989) calls “important idiosyncracies” of individual industries. A healthy reason
for this change in focus is the disorientation that follows from aiming a plethora of
models at a small set of inter-industry facts.
We take the variety of models (many emphasizing the interaction between strate-
gic elements and industry-speciﬁc conditions) as given and do not dispute the value
of detailed analyses of individual industries. But it is our view that the critique of
generalized models in industrial organization is overstated. There are important
stylized facts that characterize industry performance (see Schmalensee, 1989, for
a lengthy list), facts that have periodically motivated policy initiatives with serious
consequences. Systematic modeling of inter-industry differences is certainly useful
We thank Bill Dougan, Mike Maloney, Harold Mulherin, Geoffrey Shepherd, Robert Tamura,
and seminar participants at Clemson and the University of New Mexico for their helpful comments.