Game-theoretic Context Research Articles

Decision Rules, Decision Styles and Policy Choices

The article emphasizes the importance of theory construction for comparative policy research. Attempts to identify the impact of institutional arrangements on policy choices are complicated by interaction effects between institutionalized boundary and decision rules on the one hand, and `decision styles' on the other hand - which are defined as cognitive and normative patterns that characterize the way in which interests are defined and issues framed and resolved under the applicable rules. A typology of such styles is developed with reference to recent findings of experimental social psychology, and the empirical implications of different combinations of decision rules and decision styles are examined in a game-theoretical context.

To pay or not to pay: A model of international defaults

AbstractMost lenders in international financial markets accede to reschedule overdue payments when sovereign borrowers fail to meet repayment requirements rather than foreclose and force a default. This practice provides borrowers with strong incentives to avoid prompt payment and to sak rescheduling instead. Because lenders cannot distinguish between borrowers who are unable, and those who are merely unwilling, to repay the loan, they will agree to reschedule loans of borrowers who would choose to pay rather than face outright default. This paper augments the traditional lender‐borrower model by explicitly considering rescheduling, in addition to default and repayment, and places it in a game theoretic context. It develops a theory of unwillingness to pay, as opposed to inability to pay, and derives strategies by which lenders can at least partially protect themselves against rescheduling requests by borrowers unwilling, but not unable, to fulfill their obligation.

Differential Game Theory Applied to a Model of the Arms Race

The arms race model of L. F. Richardson is considered within a differential game theoretic context. Five different approaches to the game are considered and the results compared. The particular model considered is a three nation arms race model which becomes a three player differential game. The five approaches to the game are the minimax, Nash, and Pareto approaches as well as two coalition approaches: C.P.O. (Coalitive Pareto Optimality) and N.C.O. (Nash Coalitive Optimality). Some suggestions are made for future research and some possible economic, social, and political implications of this work are considered.

Comment on “The Confusion of Is and Ought in Game Theoretic Contexts”

Comment about Kadane, J. B., P. D. Larkey. 1983. The confusion of is and ought in game theoretic contexts. Management Sci. 29(12) 1365–1379.

The Confusion of Is and Ought in Game Theoretic Contexts

This paper explores the distinction between normative and positive theoretical statements in a game theoretic context from a Bayesian perspective. Normative and positive theoretical statements are often confused in decision making research. The confusion results from some unique epistemological and methodological problems associated with cognitive behavior as an object for scientific inquiry. The confusion persists because of poor model validation procedures which are themselves further complicated by the confusion. The confusion greatly impedes the development of more useful prescriptions for and predictions of human decision behavior. From a Bayesian perspective which acknowledges the importance of incomplete information and imperfect theories of behavior, the confusion is unnecessary.

Some reward–penalty rules for the multi-armed bandit problem which are asymptotically optimal

In the mathematical learning literature, reward–penalty rules have been studied in various decision-theoretic and game-theoretic contexts, the multi-armed bandit problem included. Here we propose an elaboration of Bather's randomised allocation indices which yields rules for the multi-armed bandit which are both reward-penalty and asymptotically optimal.

Some reward–penalty rules for the multi-armed bandit problem which are asymptotically optimal

In the mathematical learning literature, reward–penalty rules have been studied in various decision-theoretic and game-theoretic contexts, the multi-armed bandit problem included. Here we propose an elaboration of Bather's randomised allocation indices which yields rules for the multi-armed bandit which are both reward-penalty and asymptotically optimal.

On Nonterminating Stochastic Games

A stochastic game is played in a sequence of steps; at each step the play is said to be in some state i, chosen from a finite collection of states. If the play is in state i, the first player chooses move k and the second player chooses move l, then the first player receives a reward akli, and, with probability pklij, the next state is j. The concept of stochastic games was introduced by Shapley with the proviso that, with probability 1, play terminates. The authors consider the case when play never terminates, and show properties of such games and offer a convergent algorithm for their solution. In the special case when one of the players is a dummy, the nonterminating stochastic game reduces to a Markovian decision process, and the present work can be regarded as the extension to a game theoretic context of known results on Markovian decision processes.