How does managers' pursuit of their own intraorganizational interests affect decisions about what work to outsource and how to contract with vendors? I study this question using a qualitative study of outsourcing in the information technology department of a large financial services firm.Traditional transaction cost-based theories argue that decisions about which transactions to outsource should reflect the characteristics of those transactions, yet I find only a weak link between transaction characteristics and outsourcing decisions. Qualitative evidence suggests that managers' pursuit of their own intraorganizational interests helps to explain why outsourcing decisions were often divorced from transaction characteristics. I found that the consequences of outsourcing projects were consistent with the assumptions of transaction cost and capabilities-based theories: managers had less authority over outsourced projects than internal ones, those projects were subject to weaker administrative controls, and outsourced vendors provided different capabilities than internal suppliers. However, the way that those consequences were evaluated often reflected managers' own interests rather than those of the organization.I highlight three aspects of organizational structure that affected how managers evaluated outsourcing: the nature of differentiated goals and responsibilities, the administrative controls that managers faced, and the pressures caused by interdependent workflows within the organization. I also show how the distribution of authority and other resources shaped which projects were outsourced. The analysis highlights the value of understanding make-or-buy decisions as an endogenous consequence of the structure in which those decisions take place, rather than as isolated decisions that are maximized regardless of their context. Keywords: Outsourcing; Firm boundaries; make or buy; transaction cost economics; politics; resource dependence; organizational decision-making; offshoring; information technology.Acknowledgements: I am very grateful to Heski Bar-Isaac, Henrik Bresman, Forrest Briscoe, Diane Burton, Isabel Fernandez-Mateo, Martin Gargiulo, Bob Gibbons, Javier Gimeno, Adam Grant, Quy Huy, Michael Jacobides, Sarah Kaplan, Jackson Nickerson, Joanne Oxley, Sean Safford, Felipe Santos, Brian Silverman, Andrew von Nordenflycht, Christoph Zott, seminar audiences at London Business School and the Rotman School, Nick Argyres and three anonymous reviewers for their comments and suggestions on various drafts of this paper. I am particularly grateful to all of the employees of the Bank who spent so much time talking with me and sharing the data that led to this paper.