Organizational Learning: An Experimental Investigation

We propose to experimentally study decentralized organization learning. Our objective is to understand how learning members of an organization cope with the confounding effects of the simultaneous learning of other agents. An important distinction of our approach is that we test predictions from a simple stylized model of organizational learning with fully rational agents, developed in Blume and Franco, 2004. Decentralization is captured through explicit constraints on the joint strategies of the agents in the organizations. Rather than exogenously specifying individual learning rules, the model predicts learning behavior and ties its predictions to parameters about individual preference and about properties of organizations. This model yields sharp testable predictions about behavior in the organization and about how this behavior varies with the fundamental variables that characterize the organization. A side benefit from this research is that it sheds light on the roles of symmetry and randomization in games. The games we consider have numerous pure strategy equilibria. The efficient behavior is complex and asymmetric. In contrast, there is a unique symmetric equilibrium which is inefficient. Our preliminary results show that with repeated random pairwise matching the inefficient symmetric equilibrium provides a better description of behavior