October 30, 1999 Eric Rasmusen

Journal of Economic Literature review of Economic Games and Strategic Behavior, by Frank Stahler, Cheltanham: Edward Elgar Publishing, 1998.

Body plus references: 802 words.

 This book is about self-enforcing agreements between two people in the context of infinitely repeated  interactions and complete information. The well-known "Folk Theorem" of game theory says that  there exists an infinite number of   subgame perfect equilibria in such games, so for analysis to be interesting it must add extra structure. This extra structure can be supplied by changing the game, making the number of repetitions finite or information incomplete, or by changing the equilibrium concept, excluding some of the subgame perfect equilibria as being implausible.

        Frank Stahler's approach is to refine the equilibrium concept.  Since he is interested in games in which the two players can talk and  make an explicit agreement,  he requires that their agreement be a "weakly renegotiation-proof" equilibrium,   meaning, as he  puts it on page 27 that "none of its continuation payoffs is Pareto dominated by another continuation payoff."  The effect of this is to rule out  agreements that use the  "grim strategy," of complete noncooperation forever after if one player deviates from the agreed-upon behavior; the players cannot agree to give up future  gains from trade  as a punishment for opportunism. It allows agreements in which a deviating player is punished by being required to return to good behavior for a certain number of periods during which the other player need not reciprocate, after which  normal trade benefitting both sides will  resume.

The aims of this book are nicely laid out at the start: to answer a series of questions  on  page 4 with which any reader should begin, questions ranging from  "What role do contracts play in a non-cooperative environment in which the economic game does not guarantee compliance?" and "Which self-enforcing contract will be chosen?" to  "Might a party prefer a self-enforcing contract even if enforceable contracts were possible?" It will be apparent that the word "contract" is used here as a synonym for "agreement,"  just as it is by  many economists but contrary to the more precise usage in law, where many an exam question consists of "Is this agreement a contract?"-- that is,  "Is this agreement be enforceable in court?" Using the word in the legal sense, this book is about situations where agreements are  possible, but contracts are not.

      The questions to be answered  are good, but they become less interesting the more the context is narrowed. Without courts, what use are agreements? A number of things come to mind.   Agreements make the expectations of each party clear,  permitting better coordination and avoiding future disputes. They allow a wronged party to credibly disparage the  wronging party's reputation. They bind the parties morally, so that a principled party will keep the agreement even if no court binds him.  At the same time, explicit agreements do have   costs: the transactions cost of clarifying  what each party is to do, and costs arising from incorrect enforcement.

 	Most of these points are hard to address in a context of infinitely repeated games of complete information with no uncertainty and no transaction costs.  Transaction costs can be studied in such games, by preceding them with negotiation games  which explicitly model the process by which the agreement is reached. The "cheap talk" literature  (e.g., Farrell [1987])  does  this, and it would have been a nice way to put renegotiation- proofness into the structure of the game instead  of in  the equilibrium concept-- to introduce it structurally rather than as reduced-form.   Lacking that, the focus is on coordination of expectations, the agreement simply making clear what each party is to do and what happens if one of them deviates.   The analysis comes down to the study of  renegotiation-proof punishment paths and the effects  they have on the division of the gains from trade.

   	 Renegotiation-proof punishment paths are interesting in themselves, however, and perhaps  the most interesting conclusion of the book is its answer to the question, "Might a party prefer a self-enforcing contract even if enforceable contracts were possible?"   The transaction  costs of enforceable contracts and the possibility of court error are obvious  reasons, but not the only ones. In addition, a party with weak bargaining power might prefer that enforceable contracts be impossible because in the absence of court enforcement the two parties must make the agreement beneficial enough to each of them to prevent later deviation as well as initial consent. The stronger party cannot take too much surplus away from the weaker party, or the weaker party will have too little incentive to adhere to the agreement. Thus, we arrive at the surprising conclusion that the unbiased rule of law can sometimes hurt the weak and benefit the strong.  That is a notion  we may hope Dr. Stahler pursues   in his future  work.

 Farrell, Joseph (1987)  ``Cheap Talk, Coordination, and Entry,''    RAND Journal of Economics, 18:  34-39 (Spring 1987).