Strategic Experimentation in Networks

This paper examines experimentation and innovation when experimental results diffuse along social links. This public-goods nature of experimentation is a feature of many areas of economics, such as consumer choice, research and development, and agriculture. The paper asks: How do different patterns of social and geographic links affect experimentation? Who experiments and who free rides? Do more links enhance or diminish social welfare? The analysis finds, first, that social networks can foster specialization. In every social network there is an equilibrium where some individuals experiment and others completely free-ride. In many networks this extreme is the only equilibrium outcome. Second, specialization can benefit society as a whole. This outcome arises when specialists are linked, collectively, to many agents. Finally, new links can reduce overall welfare. A new link increases access to new results, but also reduces an individual's incentives to conduct his own experiments. Hence, overall welfare can be higher when there are holes in a network.