BEYOND MARKETS AND COMMUNITIES: A COMPARATIVE APPROACH TO KNOWLEDGE EXCHANGE IN ORGANIZATIONS 1 S. S. LEVINE * , Singapore Management University, Singapore M. J. PRIETULA, Emory University, Atlanta, Georgia While knowledge transfer has been shown to affect organizational performance, little is known about the processes of knowledge exchange between organizational agents. We propose that combination of various modes of exchange and degree of tie strength produce at least five different configurations: neo-classical exchange, local search, embeddedness, community exchange, and performative ties. Using an agent-based simulation of problem solving in an organizational setting that involves knowledge exchange, we find that embeddedness and community exchange provide results that are superior to neo-classical exchange. Performative ties, however, outperform both, even if just a minority of the organizational agents is able to extend such ties. In addition, we find that the marginal returns on performative ties are greatest when such ties are relatively rare, suggesting that the cost associated with encouraging them can be minimized with little in adverse effects. We conclude by discussing managerial implications for team setup and facilitation of knowledge transfer. Keywords: Knowledge, Social Network, Exchange, Reciprocity, Performance Knowledge has grown to occupy a major role in the discussion on firm performance and survival (Winter, 1987). In the management literature, it has been recognized as a valuable resource (Kogut and Zander, 1995), a source of lasting competitive advantage (Teece, Pisano and Shuen, 1997), and even the very foundation for the existence of the firm (Grant, 1996). But unlike other economic resources, such as capital or land, knowledge is a social entity. In the last two decades, much ink has been spilled to argue and show that organizational knowledge, the kind that is necessary to create a competitive advantage, does not reside in manuals or training books, but in individuals and the regularities by which they cooperate, i.e. tacit knowledge and routines (Cohen & Bacdayan, 1994; Foss, 2003; Kogut & Zander, 1992; Nelson & Winter, 1982). Consequently, the organizational literature has been devoting much attention to the transfer of knowledge between agents, either individuals within the same organization (intra organizational knowledge transfer) or between organizations (inter organizational). Much has been achieved. We now have better understanding of the flows of knowledge between agents and the consequences for various phenomena of managerial interest, such as organizational structure (Reagans & McEvily, 2003; Sorenson, 2003), innovation (e.g. Hargadon & Sutton, 1997; Obstfeld, 2005) and performance (e.g. Haas & Hansen, forthcoming; Hansen, 1999). Some 1 We acknowledge helpful comments from participants at Organization Science Winter Conference 2005 at Steamboat Springs, Colorado, North American Association for Computational Social and Organizational Science 2005 at Notre Dame, Indiana, Agent 2005 at Argonne National Laboratory, the University of Chicago, and Inter- organizational Networks at Emory University. Projjal Ghatak and Esther Yansi Ye provided able research support. ∗ Corresponding author address: Sheen S. Levine, Singapore Management University, 50 Stamford Road, Singapore 178899; e-mail: sslevine@sslevine.com