Networks in the Laboratory
There is a large body of empirical work which studies the influence of network structure on economic outcomes, but these observational studies are subject to identification problems. In the laboratory, by contrast, we can control subjects’ position in the network and their information. This provides an opportunity to test the predictions of network theories and study the effects of variables about which our models have little to say. This chapter surveys network research in experimental economics. The first part discusses experiments of games played on networks, focusing on how different architectures lead to different outcomes. The second part discusses trading in networks where connections represent potential trading relations. The aim of these experiments is to investigate the role of network architecture in determining the efficiency of markets. Overall, the results show that the clarity that is achieved by putting behavior “under the microscope” is well worth the necessary simplification.