Home Address:
  123 York Street, Apt. 11B
  New Haven, CT 06510

Telephone: (203) 285-5607 (cell)
                   (203) 785-8321 (home)

Economic Theory
Finance

Finance
Microeconomics
Macroeconomics

2003 (Oral): Mathematical Economics (with distinction), Financial Institutions and Growth
2002 (Written): Macroeconomic Theory (with distinction), Microeconomic Theory

Perfect Conjectural Equilibrium and Endogenous Market Power

Professor John Geanakoplos (chair)
Professor Truman Bewley
Professor Donald J. Brown

May 2006

Ph.D., Economics, Yale University (expected May 2006)
M.A. Economics, Yale University, 2002
M.A. Economics, Universitat Autonoma de Barcelona, 2001
B.A. (magister) Quantitative Methods, Warsaw School of Economics, 1999

Carl Avid Anderson Fellowship, Cowles Foundation, Yale University, 2005–2006
John F. Enders Award, Yale University, 2005
Dissertation Fellowship, Yale University, 2005
Annual Cowles Foundation Prize, Yale University, 2002, 2003, and 2004
University Fellowship, Yale University, 2001–2006
ACE Phare Scholarship, 2001–2002

NSF/CEME/NBER Mathematical Economics Conference, University of California at Berkeley, October 2005
14th European Workshop on General Equilibrium Theory, University of Zurich, May 2005
Invited talk: Theory Lunch, Brown University, April 2005
10th Spring Meeting of Young Economists, Geneva, April 2005
1st Annual CARESS–Cowles Conference on General Equilibrium and its Applications, Yale University, April 2005

Co-organizer of the Inter University Conference, Yale University 2004

Professor John Geanakoplos
Department of Economics
Yale University
P.O. Box 208281
New Haven, CT 06520-8281
Phone: (203) 432-3397
Fax: (203) 432-6167
Email: john.geanakoplos@yale.edu

Professor Donald Brown
Department of Economics
Yale University
P.O. Box 208283
New Haven, CT 06520-8283
Phone: (203) 432-6934
Fax: (203) 432-7316
Email: donald.brown@yale.edu

Economic thinking about market power has long been shaped by a game theoretical tradition that originated in the work of Cournot, and by monopolistic competition. Both these approaches divide traders into strategic players and non-strategic consumers, and therefore a priori assume away the market power of some of the market participants. Worse than that, the non-strategic traders determine (exogenously) the price impacts of the strategic players. In my dissertation, I propose a framework that does not impose any a priori restriction on traders’ price impacts, but derives it. The model does not discriminate between consumers and producers. In addition, all traders react optimally in and out of equilibrium, a feature that Cournot-Nash equilibrium lacks.

Theoretical contributions of the dissertation include a choice theory for consumers and producers with market power and a new concept of equilibrium, that I call perfect conjectural equilibrium. The novelty of this concept lies in the treatment of non–equilibrium behavior, i.e., the behavior of the economy after unilateral deviations. A trader's deviation from equilibrium is followed by a price change sufficient to encourage all other traders to absorb the deviation, so that markets clear. The response of price to a deviation defines a downward sloping demand curve facing each trader, and the ability to affect prices is taken into account in the optimal decisions of all traders in and out of equilibrium.  The proposed concept of equilibrium is conceptually similar to a consistent conjectural variation equilibrium introduced by Bresnahan [1981] and Boyer and Moreaux [1983], or to a rational conjectural equilibrium in Hart [1985].  In equilibrium from this paper, however, traders have conjectures about the price impacts - price changes as a function of off-equilibrium deviations rather then absolute levels of trade.  In addition, unlike in Bresnahan [1981] traders make consistent conjectures not about the response of one rational opponent who maximizes preferences, but about the outcome of the interactions in an economic system with many heterogenous traders.  In the existing literature on conjectural equilibrium, the equilibria may not exist or be locally unique.  Perfect conjectural equilibrium does always exist in oligopolistic settings with differentiable utility and convex cost functions, and it is generally determinate.

I use the model to study the following phenomena in a general equilibrium framework:

• the determinants of market power,
• the effects of non-competitive trading on the equilibrium allocation, prices and welfare,
• the impact of ownership structure on the equilibrium outcome,
• cross-market effects of strategic pricing behavior,
• endogenous market structures.

An advantage over the standard competitive approach is that the framework defines equilibrium prices and quantities in the case of producers with increasing returns to scale and bilateral monopoly.

The dissertation has four chapters.

Chapter 1: Perfect Conjectural Equilibrium and Endogenous Market Power

In the first paper, I study interactions among traders with market power in a smooth and separable exchange economy. In such an economy, consumers' utility functions are additively separable in non-numeraire goods, and producers use only the numeraire as an input.

In the first part, I define a concept of equilibrium and also establish the following results characterizing it: the existence of equilibrium, its (generic) local uniqueness, its (generic) Pareto inefficiency, convergence of equilibria to the Walrasian equilibrium in a k-replica economies, and testability.

In the second part, I study the effects of non-competitive interactions on allocations, prices and welfare. I find that, in general, market power reduces the volume of trade. The sign of a price bias relative to a competitive price and the distribution of welfare loss depend on the third derivatives of utility and cost functions.

The model suggests that a trader’s price impact, captured by the slope of a demand faced by him or her, is directly related to the convexity of the preferences or cost functions of the trading partners. The price impacts mutually reinforce each other and they negatively depend on the number of traders.

Compared to Cournot, my model will typically predict more competitive outcomes. Even with very high Herfindahl-Hirschman Index (HHI), some industries can be competitive, when the convexity of the technology is low.

Chapter 2: Perfect Conjectural Equilibrium and Endogenous Market Power: a Non-Separable Case

In this paper, I relax two assumptions made in Chapter 1: separability of the utility function and numeraire as the only input. This allows me to study phenomena associated with strategic pricing across markets. In particular, I investigate cross-subsidization, and substitution of inputs traded in noncompetitive markets by competitive inputs. I also discuss the problem of existence of equilibrium, determinacy and testability of equilibrium in a non-separable economy.

Chapter 3: The Institutional Traders and Non-competitive Financial Markets

The existing equilibrium asset pricing models, for example CAPM or C-CAPM, assume that investors are "return takers." That is, when making their decisions they assume that they cannot affect the returns of asset. A vast body of empirical literature suggests that the trades of large institutional investors do impact asset prices and hence returns, and that investors take this impact into account when trading. In this paper, I use the abstract framework from the first chapter to explain the effect of institutional trading on asset prices and allocations. I discuss the effects of non-competitive trading in a model with mean-variance optimizing traders, in a strategic version of CAPM (S-CAPM). I also show that market power may account for the equity premium and the risk-free interest rate puzzles, when the utility functions are characterized by prudence.

Chapter 4: Perfect Conjectural Equilibrium in Games (work in progress)

In this paper I formulate the concept of Perfect Conjectural Equilibrium in an abstract non-cooperative game, possibly with two or more players. Intuitively, in such game any marginal off-equilibrium deviation of player i triggers an infra-game among all remaining players, and the equilibrium outcome of the infra-game is "predicted correctly" by player i. When playing the infra-game other players are assumed to take into account their impacts on the strategies of other players. In this paper I also study the properties of such equilibrium.