Systematic empirical evidence on the consequences of incomplete contracts is meagre, because contracts are often classified as business secrets, and are accordingly difficult to obtain for research purposes. In this study we employ the extensive information from the life of a Finnish research joint venture, Micronas, during 1980–1992 to cast light on this issue. Of particular interest is the principal‐agent relationship between the owners and the management of the research joint venture. Our investigation confirms the strong message of the incomplete contract theory. Hold‐up problems are pervasive in R&D projects, and the allocation of power and ownership is important to the success of the projects.