Division of labor and the organization of knowledge in production : a laboratory experiment

Following Garicano (2000), we consider groups whose members decide what knowledge to acquire and how to use this knowledge in production. If efficient production requires common knowledge, all group members should become workers and acquire common knowledge. But if efficient production requires diverse knowledge, one group member should become manager, acquire rare knowledge, and stand ready to help the other workers. In our laboratory experiment, we find that most groups eventually manage to coordinate on an efficient division of labor. Still, we find substantial adoption frictions. Coordination takes time, and some groups coordinate on an inefficient division of labor, probably because they do not understand what organization of knowledge is most efficient.