Rethinking Antitrust in the Presence of Transaction Costs: Coasian Implications
Rethinking Antitrust in the Presence of Transaction Costs: Coasian Implications
Dennis W. Carlton 0 1 2
Bryan Keating 0 1 2
0 B. Keating (&) Compass Lexecon , 1101 K St., NW, Washington, DC 20005 , USA
1 D. W. Carlton Booth School of Business, University of Chicago , 5807 Woodlawn Ave, Chicago, IL 60637 , USA
2 This essay is based on Carlton's Keynote Address at the 12th Annual International Industrial Organization Conference , Chicago, IL, April 12, 2014
This article analyzes how transaction costs influence the ability to charge nonlinear prices and how market structure and industry behavior affect those transaction costs. The failure to recognize that nonlinear pricing produces a different equilibrium than linear pricing together with a recognition that the pricing mechanism can be altered by conduct under antitrust review explains why the usual antitrust analysis can be misleading. The paper illustrates its points using merger simulations with nonlinear pricing. Finally, the paper analyzes how to identify situations where market power might arise and applies the analysis to exclusive dealing, credit cards and FRAND royalties.
1 Introduction Coase (1960) altered fundamentally how most economists think about externalities and government intervention. Coase made the point that in a world with welldefined property rights and no transaction costs, parties would always wind up at an
efficient point, thereby eliminating externalities and the need for government
intervention.
But Coases main point was that we do not live in such a world and that by the
assignment of property rights, a government can influence transaction costs and
thereby the ability of the economy to reach an efficient point. Therefore, a
government should assign property rights in order to enable the economy to reach
an efficient solution.1
We explore in this article the implications of Coases insights for antitrust and
show how Coases insights mean that we should refocus much or at least some of
our economic analysis of the antitrust issues that are related to mergers and market
power. This article together with a companion paper (Carlton and Keating
forthcoming) makes several points that are related to transaction costs and antitrust.
In the absence of transaction costs, output would be at the efficient level, and
there would be no deadweight loss, and hence no need for antitrust if one uses a total
surplus criterion (Demsetz 1968). Since there are transaction costs, it is important
for an antitrust analysis to examine whether they are sufficiently low to enable
nonlinear pricing; and, if not, whether the conduct that is under scrutiny lowers
transaction costs so as to allow the use of nonlinear pricing. Failure to account for
the use of nonlinear pricing can lead to a mistaken antitrust analysisespecially
when efficiencies are involved.
Finally, the cost of creating a coalition of economic agents is related to
transaction costs. Cooperative game theory tells us that one can view all of antitrust
in a unified way as the creation of one coalition that exploits the non-coalition
members. By studying when coalition formation is low and high, one can identify
situations where the creation of market power is possible.
This paper is organized as follows. Section 2 examines the foundational model of
antitrust in which a firm (or group of firms) maintains a uniform price that is above
marginal cost, restricts output (below the level that would be dictated by the
criterion of price equals marginal cost), and harms consumer and total welfare. For
example, the famous Williamson diagram (Williamson 1968) is still the way
mergers are often thought ofwith the deadweight loss from the mergers
consequence of raising the price offset in part or in total by the efficiency gain from
the lowered costs that are also the consequence of the merger.
But Coases insight forces us to focus on transaction costs in our analysis. What
prevents any firm with market power from eliminating the deadweight loss that is
typically associated with that market power? How does competition affect that
transaction cost, and how will that transaction cost be altered by a change in market
structure or by certain conduct such as the imposition of vertical restrictions?
Section 2 explores these questions and shows that only by answering them can we
1 Or, more precisely, the efficient solution is one that takes account of transaction costs and depends on
the efficient assignment of property rights. Stigler, only half-jokingly, described Coases article and in
particular his first resultthe one involving no transaction costsas 60 pages of Pareto-optimality said
very slowly. Stiglers point is that since a bigger pie is always preferred to a smaller one, Coases point is
obvious if one defines no transaction costs to mean that one obtains efficiency. [Stigler would be the
first to admit that Coases point was not, at least initially, obvious to anyone; see Stigler (2003).]
hope (...truncated)