This dissertation consists of three independent research projects aimed at gaining a better understanding of the economics of organizations. The unifying theme of the dissertation is the protection of knowledge in a hierarchy structure. This runs through all three chapters. In the first chapter, I develop a theoretical model that incorporates different communication costs into a knowledge hierarchy structure. The model shows that, for a knowledge hierarchy to be optimal while containing middle managers, it must be that middle managers are better at communicating with the owner than are the workers beneath them in the hierarchy. The middle managers must also be relatively adept at communicating with the workers, otherwise a two-tier hierarchy is optimal where they are the workers in the hierarchy. The second chapter, co-authored with Michael Sanders, examines the formation of knowledge hierarchies in a laboratory experiment. We use a simplified version of the model in Garicano and Rossi-Hansberg (2006a) to see if participants respond to changes in communication costs. We find that participants do respond to incentives but there is a strong preference for togetherness. Moreover, participants are strongly discouraged from trying to join groups if they have been (exogenously) rejected in the past, suggesting that disappointment is a factor in their deciding whether or not to join a hierarchy. In the final chapter, I develop another theoretical model which incorporates on-the-job-effort into a knowledge hierarchy structure. This model shows that workers exert less effort in a hierarchy than they would if they were self-employed and, the more layers there are in the hierarchy, the less effort they exert. However, middle managers exert more effort than they would if they were self-employed or if they were the workers in a two-tier hierarchy.