Adaptive Incrementalism and Complexity: Experiments with Two-Person Cooperative Signaling Games

Abstract

The literature on incrementalism is concerned with decentralized adaptation by decision makers with limited information to the complexity of the decision environment However, the specific aspects of limited information and complexity are often left unexamined, leaving conclusions ambiguous. In order to understand better the effect of limited information and complexity on decentralized decision making, we design an experimental game in which subjects play in the basic Cournot duopoly setting. This setting is similar to the incremental decision example given by Lindblom (1965). The game provides the players with no information about what each player's decisions are other than the history of their mutual interaction, thus guaranteeing imperfect information. We also manipulate two factors of complexity: the risk of making a bad decision and the strategic uncertainty of a particular course of action. Through these manipulations, we seek to induce and better understand incrementally adaptive behavior by the participants.