Flexibility, Rigidity, and Competitive Experimentation

Abstract

We study a framework in which agents can generate signals to increase their expected productivity. Such signals can be generated in heterogeneous environments: a flexible system in which the agent can freely allocate effort across different tasks, and a rigid system in which the agent must devote effort to all tasks. We provide sufficient and necessary conditions for optimal experimentation in each system. Experimentation is less likely if the agent has high bargaining power. Competition within the Flexible system makes specialization more likely. When agents from different systems compete, there is a unique equilibrium where both agents experiment if the Rigid System is restrictive enough. We discuss competition within and between different education systems as well as applications in organizational economics in light of our model.